These are the year-on-year figures.

They answer a narrower question than they appear to, because the half being compared against, the six months to June 2025, is the half immediately before this company's revenue base was rebuilt.

Set the same figures against the half that actually preceded the period, the six months to December 2025, and the direction reverses. Revenue fell 11.0%. Gross profit fell 12.9%. Profit attributable to owners fell 32.1%.

That comparison is available because the prospectus carries the audited full-year accounts on the same basis as the interim statements [2], so the second half of FY2025 can be derived by subtraction.

There is nothing wrong with the company leading on the year-on-year figures; the reporting rules ask for them. But the year is a strange stretch to measure across, because the customers changed almost entirely within it. US revenue fell from 70.5% of the total to 8.8%, Switzerland rose from 10.4% to 37.6%, and Hong Kong from nothing to 21.7%. Comparing the two ends of that is comparing two different customer bases. By the second half of FY2025 the new one had settled, so the December and June halves are the first that measure the same business twice.

The sequential comparison is not perfectly clean either. If the second half is genuinely the stronger season, then measuring a first half against a second half overstates the decline, and part of the 11.0% would be calendar rather than performance.

The filing also does not help to settle it. Its outlook says the Group "expects a stronger performance for the second half ended 31 December 2026 due to seasonality, that is higher shipment volumes during key festive periods in the fourth quarter" [1, page 20]. A note to the same statements, eleven pages earlier, says the Group's "businesses are not affected significantly by seasonal or cyclical factors during the financial periods" [1, page 9]. One document, filed on one day, points both ways, so how much of the 11.0% is seasonal it does not let anyone say.

Where the thesis stood

Our first look, published 30 July at the S$0.53 offer price, made several arguments. This half-year’s results let us test two of them directly. For the rest, we will need to wait for later periods.

The first was that the fifteen-fold revenue growth concealed falling profitability: gross margin dropped from 40.4% to 12.0% and profit to owners fell 25% from its FY2024 peak.

The second was that the revenue was not quite the company’s own, because roughly 90% of it arrived through customers referred by All-Link PRC, a China-based forwarder controlled by the husband of the controlling shareholder and expressly outside the listed group.

We put the offer at about 9.9 times trailing earnings, in line with size-matched regional forwarders rather than cheap, and at about 2.8 times book against a peer median just under 1.0. The closing line was that at close to ten times a falling earnings line, the market was being asked to pay for a transition that had only just begun.

One thing in that piece needs correcting before going further. We treated the shift away from American revenue as something that might or might not happen after the listing. It had already happened, and the prospectus disclosed it. We read the operating-geography table, which reports by the entity booking the revenue, and did not read the customer-location table in the audited accounts, which reports on the same basis as these interim statements. Had we read it, we would have seen Switzerland at US$24.9 million, 33.6% of FY2025 revenue, matching the 33.6% the prospectus attributes to Customer A. The transition was largely complete at the offer.

The numbers, on both comparisons

Gross margin is where the two comparisons disagree least: down 0.92 points on the year, 0.25 on the half, and now 11.52%, the lowest on record.

Below the gross line, profit fell far faster than revenue, because costs rose rather than fell: administrative expenses nearly doubled from the half before, from US$1.3 million to US$2.5 million, US$1.0 million of it one-off listing costs. That is why profit before tax fell 35.6% and profit to owners 32.1%, against a revenue fall of 11.0%.

The company’s adjustment strips the US$1.0 million of listing costs to reach adjusted profit of US$3.4 million, up 11.6% year on year. It leaves in a US$0.4 million gain on the MF Logistics earn-out, though, which its own results release calls a “non-operating gain” [7].

Strip both and profit is about US$3.0 million, roughly flat on the June 2025 half and about 16% below the December one. That last figure is rough rather than exact: the gain is disclosed pre-tax in note 6 [1], so netting it against a post-tax number mixes bases.

The segment note localises the damage, though only year-on-year: the three-segment split is new this period, so there is no December half to measure against, and these figures compare with 1H FY2025. Singapore segment profit before tax fell from US$3,165k a year earlier to US$1,756k, down 44.5%, while its revenue rose 27.5%. The Philippines segment grew profit from US$397k to US$562k. Malaysia contributed US$487k against no comparable figure, since that segment did not exist a year earlier.

The filing does not disclose how the US$1.0 million of listing expenses is allocated between segments, so some or all of it may sit in Singapore. Even attributing the entire amount there leaves about US$0.4 million of the Singapore decline unexplained.

What the guidance requires, and what the filing says about it

Management expects FY2026 revenue to exceed FY2025, attributing the expected stronger second half to heavier fourth-quarter festive volumes [1].

The revenue bar is low, though not automatic. FY2025 revenue was US$74.1 million and the half just reported was US$40.1 million, so the second half needs only about US$34.0 million to match it, roughly 15% below the half just delivered. Even a repeat of this half’s 11% fall would produce about US$35.7 million and still clear FY2025. A full-year increase is therefore better read as the expected outcome than as evidence of momentum.

It also revives the seasonality contradiction flagged earlier: the outlook leans on a stronger festive-quarter half, while the “Seasonal operations” note says the business is not significantly affected by seasonal factors [1, page 9]. The two need not strictly conflict, since the note concerns whether seasonality distorts the period presented while the outlook concerns whether the fourth quarter carries more volume than the rest of the year. Still, they are an odd pair, and a reader leaning on the guidance is entitled to ask which one management means.

Thesis check

Read more

The pattern was familiar. Again and again, the headline number said one thing, while the operating business said another. That gap — between what companies disclose and what investors actually understand — is exactly why we started The SEA Analyst.

August in Review

Not every update warranted a standalone email, so this roundup brings the reporting cycle together. Each header links to the full article.

HL Global (AVX), 7 Aug The board said recurring profits have not been sufficient to support dividends, and the market is pricing exactly that.

JustCo (JCO), 10 Aug The SEA organic business is improving, but the reported profit was an accounting gain and the revenue growth came from a consolidation rather than new business.

Jollibee (JFC), 11 Aug Record headline profit, but overseas earnings went backwards, net debt rose ₱12 billion, and the Wall Street re-rating thesis is quieter, not louder. Since publication: Jollibee has confirmed Hong Kong over the US as the venue for its international unit listing.

Foundation Healthcare (FHH), 17 Aug The reported profit was distorted by a pre-IPO swap structure that ended on 1 July, the day after the first-half closed; the number that matters, core margin, moved down.

Sheng Siong (OV8), 20 Aug The 31.8% gross margin confirms the quality thesis, but the S$520 million Sungei Kadut build now governs the stock.

UltraGreen.ai (ULG), 20 Aug Two US approvals have appeared — Zydus in August, Provepharm in July; the moat is real, and how far competition erodes it is the open question.

Haw Par (H02), 21 Aug Strip out last year's anniversary special dividends and the income stream is intact; at S$15.06, the market caps the company below its UOB and UOL stakes alone.

Q&M Dental (QC7), 22 Aug Core dental profit grew 10%, but Q&M is committing close to a third of its market value to a pan-Asian rollup across three markets at once.

Centurion (OU8), 23 Aug Revenue grew 31% and core profit grew 34%; reported EPS fell 64% because of CAREIT minority interests and non-core valuation effects, not the operating business.

HRnetGroup (CHZ), 24 Aug Singapore posted its first gross-profit rise in three years, and management for the first time framed the company around unlocking value for minority shareholders — the language has moved; the action has not.

PropNex (OYY), 26 Aug The thesis held through the launch drought with market share rising in every segment, but profit fell 3.1% on flat revenue and that margin compression is worth watching into the second half.

Hafary Holdings (5VS), 30 Aug Singapore held, but the Kluang plant lost more money on less revenue — the thesis has one more half to show that the miss was timing, not structure.

Tiong Woon (BQM), 1 Sep Revenue and profit both hit records for the full year; the stock still trades at half its book value.

A Coverage Status System

Every company we cover carries a thesis.

But theses age at different rates — some are live and moving, others are valid but waiting, and a few have run their course. Treating all of them the same way would mean updating when there is nothing new to say, and going quiet when something is actually happening.

Active Coverage means the thesis is live and we are tracking it. Expect updates when material developments occur.

Catalyst Watch means the thesis is valid but the trigger is indefinite. The company is not a value trap — there is no structural reason the discount never closes — but there is no near-term catalyst and no management action expected. We will not send regular updates on Catalyst Watch names, but we will revisit when the facts move.

Stop means the thesis has closed — either because the discount is structural with no credible path to close, or because the investment case now depends on a factor that is unlikely to materialise.

Going forward, every company in our coverage carries one of these statuses. The full list, all nineteen names with status and rationale, is in our coverage map.

What We Are Working On

Over the next three to six months, three areas are where we will be watching most closely and where most of our new initiations will come from.

Alongside these, we will continue reviewing IPOs as they come to market across the region. And where we find a company or angle that the market has significantly overlooked, outside of any planned sector, we will write about it.

Subscribe now

A Note on Feedback

Several readers have written in with corrections, privately and in the comments, and we value every one of them.

If the direction we are taking — whether it is the depth of analysis, the companies we cover, or the cadence — is not what you were looking for, write to us at research@theseaanalyst.com.

We have also heard kind words. Our OSAT sector survey was reposted by Unisem, one of the companies the survey covered. Prof. Mak of NUS wrote that "the market needs deep analysis like this more than ever as stock exchanges chase more listings" on our All-Link IPO Review. And privately, readers have written in with notes like: "Keep doing good work. Always looking for high quality SEA stock research."

We are grateful for all of it. We also post on LinkedIn. Find us here if you prefer catching the research that way.

Thank you for reading.

The SEA Analyst

The SEA Analyst is an independent research and education publication. It is not licensed or regulated as a financial adviser in any jurisdiction, and nothing we publish is financial advice or a recommendation to buy or sell any security. We do not publish price targets or forecasts of how a security will trade, and we disclose our own position in every piece. Do your own research, and consult a licensed professional before making any investment decision.

Thesis valid, no near-term catalyst. Updates only when something moves.

Thesis closed. No further updates.

Our 17 April 2026 initiation flagged three risks to watch:

1. A gross margin decline that might be a trend rather than a blip;

2. Revenue concentration in one business segment and one country; and

3. A capex-and-borrowings ramp that needed to translate into fleet use rather than idle equipment.

Two of the three moved the reassuring way this period: margin recovered almost to its prior peak, and capital expenditure moderated while gearing improved. The third, concentration, is where the result gets interesting, because it runs straight into a target management had set for itself seven weeks before these numbers landed.

In early July, The Business Times reported management’s goal of more than S$200 million in annual revenue by FY2030, alongside plans to grow beyond crane rental and deepen the company’s regional footprint.[8] FY2026 revenue came in at S$187.7 million, already 94% of the way to that five-year target, in year one. On the revenue-scale half of the ambition, the company is essentially there, and the target looks conservative rather than stretching.

The diversification half went the other way. Of the group’s S$24.2 million revenue increase, Singapore alone contributed S$25.6 million (more than the entire gain), which means that, taken together, everywhere else net-subtracted from revenue. The non-Singapore book shrank from S$44.4 million to S$43.0 million even as the group grew 15%. India was the one genuine bright spot, up 30%; Malaysia roughly halved and the Middle East fell 15%.

One weak year is not proof that the strategy has failed. India, the Middle East and Malaysia are small, project-driven markets, where a single heavy-lift contract won or lost can swing revenue by millions, so a soft year in them proves little on its own.

What it does show is a real gap between what management says it is building and what it just reported. That gap is also the most likely reason a result this strong barely moved the share price: investors may be weighting the Singapore concentration more heavily than the margin recovery. The rest of this update works through what changed, what didn't, and where the thesis now stands.

What Happened

Group revenue for FY2026 came in at S$187.7 million, up S$24.2 million or 15% from S$163.5 million in FY2025, driven mainly by the core Heavy Lift and Haulage segment, which grew 12% to S$179.6 million on higher activity in Singapore, India and Brunei. Gross profit rose faster than revenue, up 24% to S$76.4 million, because gross margin recovered to 40.7% from 37.6%, a 3.1-percentage-point improvement that management attributed to stronger margins within the Heavy Lift and Haulage book. Profit before tax rose 21% to S$29.5 million, and net profit attributable to equity holders rose 24% to S$23.9 million, taking earnings per share from 8.29 cents to 10.31 cents.

The balance sheet strengthened alongside the income statement. Cash and bank deposits rose to S$86.6 million from S$64.5 million a year earlier, while net gearing fell to 9.5% from 14.7%, as the pace of fleet capital expenditure moderated: gross additions to property, plant and equipment were S$53.6 million in FY2026, down from S$65.5 million in FY2025, and net cash capex fell more sharply, to S$23.8 million from S$45.2 million, with the cash outlay running well below gross additions, reflecting financing and timing effects. Net asset value per share rose to S$1.47 from S$1.39.

The proposed final dividend of 2.50 cents lifts the payout ratio to 24.2% of net profit, from 21.1% in FY2025, still a conservative payout given the company’s now-lower gearing and larger cash balance.

The FY2026 results were released after the market close on Friday 28 August[1], when the shares had closed at S$0.955.[9] In the first full session after the release, on Monday 31 August, they closed at S$0.965[9], up about 1%, a muted move for a result that lifted both revenue and profit growth above the prior year’s pace.

Prior Thesis Recap

Our initiation argued that Tiong Woon was a well-run, family-controlled industrial company trading at a significant discount to tangible asset value (roughly 0.5 times book, 9.2 times earnings and 3.6 times EV/EBITDA at the time), with cumulative ten-year free cash flow (~S$204 million) exceeding the market capitalisation (~S$183 million).[2]

The central risks flagged were the FY2025 gross margin decline from a 41.2% peak, heavy revenue concentration in the Heavy Lift and Haulage segment (98% of FY2025 revenue) and in Singapore (73%), and a capex-and-borrowings ramp that, if it did not translate into utilisation, would represent a capital-allocation misstep.

New Financial Data

The segment and geographic breakdowns behind the headline numbers matter more than the headline itself.

By segment, Heavy Lift and Haulage's profit before tax rose 23% to S$28.4 million on the higher revenue and improved margin; Marine Transportation's profit before tax fell 35% to S$0.8 million on lower inter-segment revenue and a smaller share of associate profit; and Trading swung to a S$0.3 million profit from a marginal loss, on a low base (external revenue up 282% to S$5.5 million, still under 3% of group revenue).

As a share of total revenue, Heavy Lift and Haulage eased slightly to 95.7% from 97.8%, but that improvement came almost entirely from the small Trading segment's low-base growth, not from a structural shift in the core business.

Geographically, the detail behind the headline is the single most important finding in this result, and it is a dollar story before it is a percentage one. Singapore revenue rose S$25.6 million, from S$119.1 million to S$144.7 million, larger than the group's entire S$24.2 million increase.

The rest of the portfolio, taken together, therefore shrank: non-Singapore revenue fell from S$44.4 million to S$43.0 million, a 3% decline in a year the group grew 15%. India was the exception, and a real one, up 30% to S$15.4 million. But Malaysia roughly halved, to S$3.7 million from S$8.0 million; the Middle East fell 15% to S$6.4 million; Indonesia and Thailand were flat-to-lower; and Brunei contributed a new S$1.3 million. On a share basis the same movement reads as Singapore rising to 77.1% of revenue from 72.9%, and non-Singapore falling to 22.9% from 27.1%, the opposite direction from the regional-diversification story the company's own outlook commentary continues to tell.[3][4]

Thesis Check

Read more

The board declared a first-half dividend of 1.00 cent per share (an interim of 0.75 cent plus a special interim of 0.25 cent), against 1.25 cents in the same period last year, where the special component was 0.50 cent.

The decline was broad, but it was not even. The two distribution segments held up reasonably: General (retail and showroom) revenue fell 6.3% to S$59.5 million and Project (developer and contract) revenue fell 3.1% to S$39.0 million.

The damage was concentrated in Manufacturing, where revenue fell 27.7% to S$24.2 million. More important than the top line, the Manufacturing segment’s recurring EBITDA loss widened, from S$1.2 million in H1 FY2025 to S$2.7 million in H1 FY2026, and at the operating level (the group’s “ORBIT” measure) the segment lost S$4.1 million, against S$2.6 million a year earlier.

That single fact reframes the half. The bull case set out in April rested on the Kluang manufacturing operation moving toward breakeven and then into a 10–15% operating margin at scale. In its first reported period since the initiation, the segment moved in the opposite direction.

What the initiation said

The 19 April 2026 initiation described Hafary as a company that had “evolved from a single-product tile trader into a S$287 million revenue platform spanning distribution, project supply, and manufacturing across multiple geographies,” anchored on a 50–60% share of Singapore’s general consumer tile market built over four decades.

The central, explicitly stated question was whether the manufacturing segment, still approaching breakeven, could deliver on its promise of margin expansion and export-driven growth. If it did, Hafary’s earnings power would step up meaningfully. If it did not, the company would remain a well-run distributor with elevated leverage from its aggressive expansion phase.

The flip triggers attached to that thesis were specific. On the bull side, the case would break if manufacturing failed to reach breakeven by FY2026 (the breakeven point the initiation cited), if US tariff policy turned against Malaysian tile exports, or if Singapore construction demand fell materially below the BCA forecast. On the bear side, the case would strengthen if manufacturing reached a 10–15% operating margin at scale, if net debt to equity fell below 1.5x, or if a second or third overseas market matured past S$15 million.

The initiation carried an implied timeline, that manufacturing would be finding its footing around FY2026, and this half moved against it: the loss widened on falling revenue rather than converging on breakeven. That is what changed since April, and it is what this update has to make sense of.

The numbers, and the disclosure that reframes them

The single most clarifying disclosure in the filing is the geographic revenue split, because it relocates the entire decline.

Singapore revenue was essentially flat (S$79.5 million against S$79.7 million a year earlier, a 0.3% dip), so the domestic distribution core, roughly two-thirds of the group, did not weaken. The S$14.5 million fall came almost entirely from abroad, and it was concentrated: the United States dropped 42% to S$11.4 million (from S$19.7 million), Malaysia fell 28% to S$19.7 million, Indonesia collapsed to S$0.1 million from S$2.4 million, and China slipped to S$3.3 million from S$4.4 million, partly offset by a jump in Vietnam.

Those declining markets are the ones the Kluang plants predominantly sell into (the US export leg and Malaysian domestic demand), so the manufacturing-segment slump and this export-market slump are, on reasonable inference, the same event seen from two angles.

One caveat sits under that inference: the filing reports revenue by customer location, not by plant origin, so the link between the geographic decline and the manufacturing segment is a reasonable read, not a reconciliation the company discloses. Either way, the deceleration is not a Singapore construction story: the problem sits in overseas manufactured-tile demand, not the domestic franchise.

That also explains a number that would otherwise look strange: group gross margin rose, to 43.4% from 40.0% on the filing’s stated basis.

As the lower-contribution manufactured-and-export-linked volume shrank, the mix tilted back toward higher-margin Singapore distribution, lifting the blended gross margin even as revenue and profit fell. The profit decline, then, was not mainly a pricing or gross-margin failure: it was volume, higher inventory impairment, and operating deleverage on the manufacturing base.

That cuts two ways. It is reassuring, because a volume dip is easier to recover from than a collapse in pricing, and Hafary's margins in fact held up. But it is also the more concerning kind of finding, because the weakness is not spread across the business; it sits squarely in manufacturing and exports, the exact part the bull case depends on.

Beyond the snapshot, three further numbers matter more than the headline profit decline.

The first is operating cash flow, which fell 60% to S$9.3 million from S$23.4 million. Operating cash flow before working-capital movements was still healthy at S$26.2 million; the shortfall came almost entirely from working capital, which consumed S$12.7 million. Inventory alone absorbed S$8.2 million of cash, and a S$9.2 million reduction in trade and other payables took more. This is the mirror image of the FY2025 story, where operating cash flow had more than doubled to S$57.9 million on improved working-capital discipline. One soft half does not undo that, but it does show how quickly the cash conversion can swing when demand softens and inventory keeps arriving.

The second is inventory itself, which rose to S$132.4 million at 30 June from S$124.1 million at year-end, building into a market that was buying less. The oversupply signal sits in finished goods, which rose to S$120.2 million; work-in-progress also climbed, to S$7.5 million from S$0.8 million, indicating the Kluang lines kept producing even as sell-through slowed. The group booked a further S$2.3 million allowance for inventory impairment during the half, up from S$0.5 million a year earlier, lifting the cumulative allowance balance to S$23.2 million, about 15% of gross inventory (S$23.2 million against S$155.6 million gross, before the allowance). The initiation flagged the S$124 million inventory book (266 days of turnover) as a specific impairment risk. This half is the first evidence of that risk crystallising, in small but rising increments.

The third is the dividend. The interim payout was cut to 1.00 cent from 1.25 cents, entirely through halving the special interim component from 0.50 to 0.25 cent. The ordinary interim of 0.75 cent was held. Management framed nothing around this, and the results carried no prospect statement at all, but the signal is legible on its own: the board chose to conserve about S$1.1 million of cash rather than defend a headline first-half figure. For a stock whose appeal to income buyers rested partly on an unbroken, growing dividend record, it is a meaningful gesture even if the ordinary portion is intact.

Not everything moved the wrong way. Finance costs fell 17.8% to S$4.6 million, with interest on borrowings down 19.1% to S$4.2 million, as lower rates and the FY2025 deleveraging fed through; annualised, finance costs are now tracking below the S$10.7 million FY2025 full-year figure. The overseas equity-accounted businesses improved: the Vietnam associate (Viet Ceramics) contributed S$0.3 million of profit share, up from S$0.1 million, and paid the group an S$0.9 million cash dividend during the half; the Myanmar joint venture contributed S$0.7 million, up from S$0.3 million. A translation gain of S$1.5 million (against a S$2.0 million loss last year) meant total comprehensive income actually edged up 2.4% to S$12.1 million despite the fall in reported profit.

Thesis check

Read more

Underneath a nearly flat top line sat the exact rotation we described when we initiated on this name in May [3].

Commission income from agency services (resale, HDB, leasing) rose 6.9% to S$360.5 million, while commission income from project marketing, the new-launch business, fell 7.8% to S$238.4 million against a thinner launch calendar [1]. One engine slowed; the others sped up enough to keep the group level. Market share by transaction volume rose to 64.3% for the half, from 60.6% for the 2025 financial year, and it rose in every single segment PropNex reports [1].

Where the thesis stood

We published our initiation on 20 May 2026, using the 19 May close of S$1.83, then roughly 30% below the October 2025 high.

Our central call was that the market was pattern-matching “new launches down, so PropNex down” without doing the segment arithmetic. Project marketing was only about 39% of revenue; the other 61% was driven by volumes management guided as flat to growing. We flagged a 60/30/10 split between our base, bear and upside cases, called it a positive-expected-value position rather than a high-conviction one, and named the single most dangerous risk explicitly: a fresh government cooling measure aimed at the HDB upgrader chain [3].

The half in numbers

Two lines carry the story.

The first is the split within revenue. Project marketing commission fell S$20.1 million year on year; agency commission rose S$23.3 million [1]. The offset was not merely present, it was slightly larger than the drag, which is why revenue edged up rather than down. This is the segment-mix resilience our initiation rested on, showing up in the accounts on schedule.

The second line is gross profit, down 3.5% to S$63.9 million, which fell faster than revenue [1]. Gross margin slipped to 10.6% from 11.0% as the cost of services rose faster than revenue, and segment profit margins compressed [2]. We would not read this as a mix effect: project marketing in fact carries a lower segment profit margin than agency services, so this half’s shift toward agency should, if anything, have helped. It is a modest compression, not a rupture, but it is real and it is the one number in this release that does not flatter the bull case.

On the balance sheet, cash and cash equivalents stood at S$130.2 million at 30 June 2026, down from S$149.1 million at the start of the year, with a further S$31.0 million held in other investments and no bank borrowings [2]. The S$18.9 million cash decline over the half came after a S$33.2 million operating cash inflow, S$33.3 million of dividends to owners and S$1.4 million to minorities, and about S$18.0 million redeployed into long-term deposits and other investments [2]. This is a company returning nearly all of its profit and still sitting on net cash, carrying only about S$3.7 million of lease liabilities and no bank debt.

The salesforce, the asset the whole franchise rests on, grew to 14,574 licensed salespersons as at 3 August 2026, up from 13,945 at the start of the year [1]. PropNex added roughly 630 net agents through a period in which a competitor was publicly recruiting against it.

Below for paid subscribers: our judgment and the valuation - the point-by-point thesis check, the PropNex-vs-ERA scorecard, the policy read, the risk section, and the price.

Read more

Revenue was broadly flat at S$292.2 million, gross profit rose 1.5% to S$62.1 million, and operating profit before tax, which strips out the volatile treasury and grant income, rose 10.0% on a constant-currency basis to S$20.2 million on a 1.9% cut in operating costs. The entire drop in reported profit came from other income, which fell 65.5% to S$5.4 million on lower grants, interest income and an adverse fair-value swing. The board raised the interim dividend 10% to 2.2 cents.

The more consequential news was strategic. In the 14 August release, management set out three “public market initiatives,” the first time it has framed capital allocation around minority shareholders rather than balance-sheet preservation.

Prior Thesis Recap

Our 7 July 2026 initiation called HRnetGroup an income-and-execution holding, not a rerating candidate.

The central observation was that FY2025’s reported profit ran well ahead of the operating business: strip out treasury income and one-off grants and the group earned roughly S$38 million, not the S$51.2 million headline. We valued the operating business at about 12 times normalised earnings plus a haircut treasury, and concluded the market was pricing the stock close to its steady-state value. Two flip triggers framed the two-sided case.

The bull flip: Singapore gross profit stabilises, removing the structural-decline discount. The bear-to-bull catalyst we thought least likely: a family capital event, a special dividend, a real buyback or a widening float, that releases the trapped cash, since SIMCO’s 79.177% control meant a minority could not summon it. Our dated forecast was that FY2026 group gross profit would be roughly flat to modestly up while reported profit fell below FY2025.

New Financial Data

The composition inverts the story we told at initiation. In FY2025 reported profit rose while the operating business was flat, lifted by items below the gross-profit line. This half did the reverse. The entire fall in reported profit traces to a S$10.3 million drop in other income, a line that sits below gross profit and reflects nothing about how the recruitment business traded.

Three items account for S$9.8 million of that drop.

The largest, S$4.4 million, was an adverse swing in fair-value marks, which turned from a S$2.9 million gain a year ago into a S$1.5 million loss.

The second, S$4.1 million, was lower Progressive Wage Credit Scheme support, mostly the non-recurrence of a S$3.0 million deferred grant received in 2025.

The third, S$1.3 million, was lower interest income.

Underneath, the operating business strengthened. Flexible staffing grew gross profit 3.7% to S$33.0 million, with average monthly contractors up 6.9% to 17,253 and segment margin widening to 12.6% from 12.0%. Professional recruitment held gross profit stable at S$27.6 million on 2,311 placements. Management calls the contractor rise flexible staffing’s “first post-COVID volume U-turn.”

The geographic mix is the surprise. The four-year Singapore decline that anchored our bear case paused:

The rest of this update is for Paid Subscribers: our thesis check against each initiation call, the Singapore turn and whether it flips our structural-decline bear case, the capital-return reframing we treat as the centre of gravity of this update, the risks revisited, what to watch into the February 2027 result, and the bottom line.

Read more

Then the same statement reported that profit from core business operations attributable to equity holders fell 16%, from S$57.8 million to S$48.8 million. Reported earnings per share fell 64%, from 8.79 cents to 3.15 cents. Net asset value per share fell from S$1.47 to S$1.40.

The market read the headline. OU8 closed at S$1.70 on 12 August and S$1.60 on 13 August, a 5.9% fall on the print, recovering to S$1.62 by 14 August and closing at S$1.60 on 21 August [14].

There are two separate declines here and they have different causes, which is worth establishing before anything else.

1. The fall in reported profit and earnings per share is explained by the fair-value loss on investment properties and the swing in associate accounting, both of which the company’s core measure deliberately strips out.

2. The fall in core attributable profit is explained by something else entirely, and by only one thing: the non-controlling interest share of core profit.

In the same fortnight, two other things happened that the earnings headline buried:

1. Centurion Accommodation REIT reported its first half-year as a listed trust and beat its IPO forecast distribution per unit by 9.6%.

2. On 4 August, Centurion won a Building and Construction Authority tender for a 7,000-bed dormitory site at Kranji Close for S$343.0 million, a commitment larger than any single land purchase disclosed in the filings we hold.

Where the thesis stood

Our initiation of 25 May 2026 [15] argued that the September 2025 CAREIT listing had converted Centurion from a dormitory operator into a three-pillar platform, that the consolidated accounts obscured rather than revealed this, and that the market was still applying operator multiples to a business that had become an operator, an asset manager, and a REIT unitholder.

We warned specifically that consolidation would make the reported headline diverge from the underlying economics. We described the FY2026 revenue line as set to jump for accounting reasons that would not translate cleanly into attributable earnings. What we did not do was size the effect on the attributable line. That omission is the reason this print landed harder than our framing prepared readers for, and we address it directly below.

The numbers

Revenue growth came from three places.

1. Westlite Mandai, consolidated since 25 September 2025, contributed S$22.9 million of the S$30.6 million increase in Singapore worker accommodation revenue.

2. New beds at Westlite Toh Guan added S$4.5 million.

3. EPIISOD Macquarie Park, operational from January 2026, added S$7.2 million and doubled Australian student accommodation revenue to S$16.5 million.

Occupancy moved the other way. Singapore worker accommodation financial occupancy fell from 99% to 94% on the ramp-up of newly completed beds at Toh Guan and Mandai; committed occupancy at Westlite Toh Guan and Westlite Mandai as at 31 July was disclosed at 99% and 87% respectively [1], which supports the ramp explanation rather than a demand explanation. Malaysian occupancy fell from 83% to 73% on foreign worker quota caps.

Thesis check

The accounting-versus-economics point: validated, and larger than we said

We wrote in May that the consolidation “does not translate cleanly to attributable earnings.” That was correct in direction and inadequate in magnitude. It is worth separating the two declines cleanly, because they are often run together and they carry different meanings.

The entire core decline is the minority share. Core profit is the company’s own measure, and it already strips out fair-value movements. So the gap between core profit for the whole group and core profit for shareholders is, by definition, the slice that belongs to outside investors, chiefly the other CAREIT unitholders. That slice went from S$7.6 million in 1H 2025 to S$38.9 million in 1H 2026.

It grew fivefold for two reasons: timing and ownership.

CAREIT only listed in late September 2025, so it was absent from the 1H 2025 comparison as a listed REIT. By the first half of 2026 it was a full six months in the accounts, and Centurion owns only about 38% of it, so roughly 62% of CAREIT’s profit belongs to other unitholders. Because CAREIT holds a large share of the group’s best, stabilised assets, a big part of group profit now flows to those outsiders before it reaches Centurion shareholders. That is why group core profit could rise 34% while the shareholders’ share fell 16%: the company earned more, but a bigger portion of it was owned by someone else.

The reported decline is fair value and an accounting reclassification. The reported number, which unlike core still includes revaluations, fell from S$73.9 million to S$26.5 million, and earnings per share from 8.79 cents to 3.15 cents. Two things caused it, and neither is worse trading. The first is a fair-value loss on investment properties, which widened to S$32.8 million this half from S$3.5 million a year ago, S$19.4 million of it in student accommodation; we return below to what actually drove that loss.

The second is the reclassification of Westlite Mandai. An associate is a company Centurion owns part of but does not control, and its share of that company's profit sits on a single line, "share of associates." A year ago Mandai was an associate, and that line carried S$28.4 million from it, of which S$22.7 million was a one-off gain from revaluing Mandai's property upward rather than recurring earnings. Centurion has since taken full control, so Mandai is now a subsidiary, and its revenue and profit are consolidated into the main lines instead. The associate line therefore lost the large Mandai figure it held last year and swung from a S$27.8 million profit to a S$4.2 million loss, not because anything deteriorated but because Mandai moved elsewhere in the accounts and last year's figure was flattered by a one-off gain.

Neither decline is an operating deterioration, but the core decline is real in the sense that a shareholder's claim on this year's core profit is genuinely smaller than last year's. The question is whether that reduction is permanent.

Is the shareholder’s share leaking? No, and it runs the other way

This was the question we could not answer from the results statement alone, so we went to the two filings that settle it.

On 2 June 2026, Centurion completed the Dividend in Specie it had proposed in February. A dividend in specie is one paid to shareholders in assets rather than cash, "in specie" being the term for payment in kind: here, Centurion handed out CAREIT units it held rather than selling them and paying cash. It distributed 84,077,200 CAREIT units to shareholders on a one-unit-per-ten-shares basis, a cash-equivalent S$0.11 per Centurion share at the S$1.10 unit price. Its holding fell from 42.9% to 657,873,900 units, or 38.1% of the trust [4].

That is a reduction in Centurion's stake in CAREIT. It is not a reduction in value for Centurion's shareholders, because they received the CAREIT units directly into their own accounts. The economic interest moved from one side of the shareholder's portfolio to the other.

The second filing is the more interesting one, and it runs the other way. As manager of CAREIT, Centurion Asset Management, a wholly-owned Centurion subsidiary, earns a base fee for running the trust, and it can take that fee in cash or in new CAREIT units. For the first quarter it elected to take 100% in units: on 20 May 2026 CAREIT issued 2,529,442 new units at S$1.1396 to settle the fee [5]. The manager then assigned the entitlement to Centurion Capital Investments Ltd, so the units were issued there rather than to the manager itself. Centurion Capital Investments is a wholly-owned subsidiary of Centurion Overseas Investments, which is wholly owned by Centurion Corporation, so the units stay inside the Centurion group.

Subscribe now

This is the sponsor being paid its management fee in more of the REIT, and the effect compounds. Each issuance accrues entirely to Centurion, which held only about 38% beforehand, so its percentage stake rises every time the fee is settled this way. And as CAREIT grows its assets and distributable income, the base fee itself grows, so each subsequent issuance is larger. It is the opposite of the dilution a sponsor’s stake usually suffers:

The Dividend in Specie was a single step down. The fee-in-units mechanism is a continuous step up: every unit issued or to be issued for the half-year was for manager fees, and on the election the manager has made so far, those units land with Centurion. Manager fees are now running at roughly S$14.8 million a year on the first-half figures, and while the manager elects to take them in units, that fee stream converts directly into a larger stake.

So the answer to the question is that the non-controlling interest drag is a one-time reset to a lower base, not a widening leak.

The REIT leg of our bull case has fired. The operating leg has not

Our initiation named a specific activation trigger for the bull scenario. We reproduce it here because it can now be scored:

“CAREIT 1H 2026 results print (likely August 2026) showing DPU run-rate of ≥3.4 cents (annualised ≥6.8 cents, above the 6.67 prospectus projection); plus Centurion 1H 2026 results print showing PBWA Singapore occupancy stabilising above 95% and Malaysian occupancy recovering to >75%.”

CAREIT reported a first-half distribution per unit of 3.499 cents against a prospectus forecast of 3.192 cents, a 9.6% beat, annualising to roughly 7.0 cents. Gross revenue beat forecast by 5.1% and net property income by 4.3%. Aggregate leverage is 29.9%, interest coverage 5.91 times, weighted-average interest cost 3.60% [3].

That leg fired.

Singapore occupancy came in at 94% against the 95% we specified. Malaysian occupancy came in at 73% against 75%. Both missed, narrowly. The Singapore miss looks like timing given the committed-occupancy disclosure. The Malaysian miss is a policy constraint that management expects to ease as quota applications reopen for selected industries.

The honest score is a half-fired trigger. The REIT and fee side is running ahead of plan; the operating side, the dormitories themselves, is behind on a lag.

The moat: validated, and more contested than we implied

On 4 August, the Building and Construction Authority awarded Centurion the Kranji Close site for S$343.0 million [6]. It is 22,079 square metres at a plot ratio of 3.0, on a 30-year lease from November 2026, for 7,000 New Dormitory Standards-compliant beds, with completion targeted for the second quarter of 2028 and operational readiness in the third [7]. Centurion will develop it through a 90:10 joint venture with an unnamed minority partner.

Our initiation argued that compliance is expensive, that roughly 900 dormitories housing 200,000 workers must retrofit by 2030, and that the operators who can fund compliance would take share. Winning a 7,000-bed compliant site is that argument made concrete.

What the tender documents also show is how contested that advantage is. The Kranji tender drew ten bids. Centurion is separately the top bidder for a second site at Lok Yang Way, a 5,000-bed plot [9] where the tender closed on 23 June and which the Jurong Town Corporation publishes in full [8]:

Centurion’s Kranji bid works out at S$49,000 per approved bed. Its own Lok Yang Way bid is S$44,333 per bed, and the Kranji site is 80% denser in beds per square metre of land, which accounts for much of the difference.

The useful comparison is not against the carrying value of the existing portfolio, which reflects assets bought years ago on leases now part-expired. It is against the other nine bidders. On that basis Centurion sits at the top of a tightly clustered band, winning Lok Yang Way by 1.22% over Banyan Capital and by 16% over the median.

That tells us something the initiation did not say. The regulatory advantage is real, but it is not proprietary. Ten well-capitalised parties, including established contractors, can all see the same opportunity and price it within a narrow range. Centurion’s edge is the willingness and the balance sheet to clear the top of that range repeatedly, not an insight the market lacks.

Management frames the supply backdrop as tightening rather than loosening. According to The Edge’s report of the results briefing, the chief operating officer of the accommodation business said the roughly 40,000 beds due to enter the market this year and early next are largely replacements for temporary dormitories whose land leases will not be renewed, and that the Dormitory Transition Scheme’s interim space standards will cut current operating supply by about 15% [16]. This is the operator’s own characterisation, reported second-hand and self-serving, but it is consistent with the compliance-driven attrition our initiation described.

A correction: the Australian pipeline is not Centurion’s in the way we described

Our initiation listed the Australian development pipeline under Growth Driver 4 as “Australian PBSA developments (Centurion-direct, separate from CAREIT-held)” and counted “approximately 1,791 beds” of additions. We described the Stirling Highway project in Perth as part of “Centurion’s directly-held Australian pipeline.”

That was wrong, and the company’s own disclosure is clearer than our reading of it. The correction matters enough to state precisely:

Of the 1,791 beds we presented as directly held, 472 are held through a vehicle in which Centurion owns a quarter. On an economic basis the pipeline is 1,437 beds. The Stirling Highway subscription was announced on 1 December 2025 and headed as an interested person transaction. We published on 25 May 2026. The filing was available to us for nearly six months and we did not read it against the growth-driver claim we were making.

Centurion holds 25% of the Perth developments. The other 75% is held by Centurion Properties Australia Investments, a wholly-owned subsidiary of Centurion Properties, which is wholly owned by Centurion Global Ltd. Centurion Global is owned in equal shares by Han Seng Juan and Loh Kim Kang David, the joint chairmen and controlling shareholders. Each subscription has been announced as an interested person transaction under Chapter 9 of the Listing Manual [10][11][12].

The terms are proportionate and on the record. Shareholder loans are advanced in proportion to shareholding, the corporate guarantee on the Fairway facility is exactly 25% of the loan, and the Audit Committee has reviewed each transaction as on normal commercial terms and not prejudicial to minority shareholders. Each is an announced interested person transaction, and the results presentation labels both Perth assets as "25% stake."

What matters analytically is what the structure does to a growth claim. Centurion supplies the brand, the operating platform and proportionate credit support.

When a development stabilises, the exit is a sale into CAREIT. That is exactly what happened with EPIISOD Macquarie Park: Lachlan Avenue Development owned the 732-bed Macquarie Park site, and CAREIT completed the acquisition in January 2026.

Centurion’s own first-half accounts confirm the route, attributing part of the associate loss to “the share of finance costs of the associated company from the prepayment of loan due to the sale of Macquarie Park to CAREIT.”

So the Sydney asset was owned three-quarters by the controlling shareholders and one-quarter by Centurion, and Centurion's REIT, in which Centurion now holds 38.25%, provided the exit. We are not able to say what development margin was realised or how it was split: the forward purchase price was agreed at the time of the CAREIT listing, the vehicle's cost base is not disclosed, and Centurion's associate line for the half shows a loss on the prepayment of the related loan. What we can say is who owned the asset on the way in. The transaction was disclosed, the CAREIT prospectus flagged Macquarie Park among properties "sold either wholly or in part by third party vendors," and two independent valuers appraised it.

We are correcting our own framing, and the 25/75 split matters less than it first sounds: it applies only to Sydney and Perth, the smaller projects, while Melbourne, which dominates the pipeline, is wholly owned. On a bed-weighted basis Centurion's economic interest in the current Australian pipeline is therefore roughly three-quarters, not a quarter:

So the correct statement is narrower than “the pipeline belongs to the family.” It is that two of the four current projects, and the Sydney project before it was sold, are three-quarters owned by the controlling shareholders, and that our May article described one of them as directly held when it is not. A reader building a bed-count model of Australian growth should discount Perth by 75% and leave Melbourne alone. According to The Edge, the REIT manager gave the same split from the trust’s side at its 6 August briefing: the Macquarie Park master tenant, Herring Road Management Pty Ltd, is 25% owned by Centurion Corporation and 75% by Centurion Properties [16], which matches the ownership we set out in Figure 4.

Two details from the Fairway announcement are worth recording, and they cut in opposite directions.

The first is that Centurion’s A$250,000 was a subscription for new shares, not a purchase from the family vehicle: issued capital rises to A$772,011 on completion, so the cash went into the development company rather than to CPAI. Nobody cashed out.

The second is that the announcement states “no independent valuation was conducted on FDA by COI(II).” Centurion relied on a valuation FDA itself commissioned in April 2026, which put the site at A$6.7 million against a book value of A$7.13 million. At a 25% interest the difference is worth about A$108,000, which is immaterial to a group with S$2.2 billion of net assets. We note it because our published Risk 6 said to watch how related-party terms are set, not because the amount matters.

The new question: what it costs to keep winning

The rest of this update, what the land actually costs Centurion to keep winning, the rebuilt look-through valuation, the scorecard against every trigger we published in May, and where the thesis stands now, is for paid subscribers.

Read more

The profit decline came mostly from lower dividend income from strategic investments, down 27.7% to S$75.2 million from S$104.1 million in H1 2025, with a smaller drag from lower interest income. The dividend cause sits one level up the ownership chain, and it is more a special-dividend story than an ordinary-payout cut.

The UOB dividend Haw Par banked in the first half fell to S$0.71 per share (UOB’s FY2025 final) from S$1.17 in the year-ago period. That S$1.17 was not an ordinary payout: it comprised UOB’s FY2024 final of S$0.92 plus a S$0.25 tranche of UOB’s 90th-anniversary special dividend. [2] Stripped of the one-off, UOB’s ordinary final eased only from S$0.92 to S$0.71, and the rest of the year-on-year drop was simply the non-recurrence of the special.

With Haw Par holding roughly 74.8 million UOB shares, that S$0.46 per-share reduction cut UOB dividend income by about S$34 million, most of the S$36 million fall in net profit. Group dividend income fell a smaller S$28.8 million, as higher dividends from UOL and other holdings offset part of the UOB drop.

Healthcare segment revenue retreated 9.0% to S$106.3 million (from S$116.8 million), which management attributed to softer consumer sentiment affected by inflationary pressures from higher energy prices, ongoing trade policy uncertainties, and a higher comparative base from advanced orders in the first half of 2025. [1]

Here is the part worth pausing on, and the thread this update pulls: at S$15.06, the market values the entire company below the quoted worth of its UOB and UOL shares alone, before any credit for its cash, Tiger Balm, or its properties.

Healthcare segment revenue retreated 9.0% to S$106.3 million (from S$116.8 million), which management attributed to softer consumer sentiment affected by inflationary pressures from higher energy prices, ongoing trade policy uncertainties, and a higher comparative base from advanced orders in the first half of 2025. [1]

Here is the part worth pausing on, and the thread this update pulls: at S$15.06, the market values the entire company below the quoted worth of its UOB and UOL shares alone, before any credit for its cash, Tiger Balm, or its properties.

Prior Thesis Recap

Our initiation piece, published 4 April 2026, framed Haw Par as a company whose real value had nothing to do with muscle rubs. [4]

The central observation was structural: Tiger Balm is the brand people know, but the S$3.4 billion investment portfolio (primarily 74.8 million UOB shares and 72 million UOL shares) was the dominant value driver, contributing three times the healthcare business’s profit through dividend income alone.

The thesis rested on three legs:

• the healthcare business as a cash-generative consumer franchise with pricing power,

• the investment portfolio as a growing dividend machine anchored by UOB’s rising payouts, and

• a persistent discount to net asset value that the FY2024 special dividend (S$1.00 per share, paid alongside FY2024’s 20-cent final in H1 2025, bringing that year’s total payout to S$1.40) suggested management was beginning to address.

The implicit flip trigger on the income side was a reversal in UOB’s dividend trajectory. The prior article noted that net profit had compounded at ~8.7% annually, “driven almost entirely by rising dividend payouts from UOB.”

Where the initiation's calls stand. The dividend picture to anchor on: the S$1.00 special was declared with FY2024's results and paid in H1 2025 alongside the FY2024 20-cent final (Note 8); the regular annual payout is 40 cents (20c interim + 20c final).

Against that baseline, the initiation's calls have played out unevenly.

The NAV-accretion thesis was correct: the portfolio and NAV both rose double digits.

The income thesis, that profit would keep compounding on UOB's rising payouts, was wrong, or at best too early, since UOB's payout reset lower within a single period and drove a 25% profit decline.

And the downside scenario we explicitly named, a 40-cent payout with the yield near ~2.4%, is now the live case, with H02 yielding roughly 2.7% on 40 cents at S$15.06.

Financial Snapshot Update

The half’s income statement carries the full weight of the dividend story.

Revenue slipped 8.2% and gross profit 9.6%, but the real mover was other income, down 29.9% as dividend and interest income fell together. That is what pulled pre-tax profit down 24.2% and net profit down 25.2%.

The balance sheet moved the other way.

Fair-value gains on the strategic portfolio lifted net assets over the six months, taking NAV per share up 10.6% to S$21.43 even as reported profit fell. The cash position stayed a fortress: the group rotated S$169 million out of bank deposits into Singapore Treasury bills, yet the combined cash-and-securities pile still grew to S$881 million against just S$53 million of borrowings.

Underneath the group totals, the two engines diverged. Healthcare absorbed the revenue decline while holding most of its profit, and the investment segment, though lower year-on-year on the softer dividend, still earned more than twice the healthcare business.

Thesis Check

Bull thesis (UOB dividend income as a passive, growing stream): Income intact, growth premise overstated.

This was the core value-driver claim in the initiation: Haw Par’s investment segment generated S$205 million in FY2025 profit, “three times the healthcare business,” primarily through S$189 million in dividend income. The thesis implicitly assumed that UOB’s dividend trajectory would continue upward; the prior article traced UOB dividends to Haw Par growing from an estimated ~S$67 million (2016) to ~S$135 million (FY2025).

H1 2026 dents that trajectory, but less than the headline suggests. Dividend income dropped to S$75.2 million from S$104.1 million, and most of that is a special-dividend effect, not an ordinary cut. UOB’s FY2025 ordinary dividend was S$1.56 per share (S$0.85 interim + S$0.71 final); Haw Par’s H1 receipt reflects that lower S$0.71 final against a year-ago S$1.17 that had bundled UOB’s FY2024 final with a S$0.25 tranche of its 90th-anniversary special. [2] The ordinary line is intact: UOB has since declared an FY2026 interim of S$0.88 per share (payable 28 August 2026), up from S$0.85 a year earlier, alongside 1H2026 net profit growth of about 3%. [5]

So the income “reset” is really the fading of UOB’s anniversary specials plus a modestly lower final, not evidence that the underlying stream is shrinking. Haw Par still banked S$75.2 million in six months from an effectively passive holding. The ordinary interim improved from S$0.85 to S$0.88, but total H2 2026 UOB cash received will still land below H2 2025, because last year’s S$1.10 half-year receipt included a S$0.25 special tranche on top of the S$0.85 ordinary. The point is the shape of the H2 shortfall: it is entirely the vanished special, while the ordinary interim is edging up. The honest revision to the initiation is narrower than “the machine is breaking.” The ordinary dividend stream is stable-to-rising, but the elevated 2025 cash, inflated by anniversary specials, was never the run-rate, so full-year 2026 dividend income will sit below 2025’s special-boosted level even as the ordinary trend holds.

Bull thesis (Tiger Balm’s consumer franchise with pricing power): Pricing power intact, Asian demand weakening.

Healthcare segment revenue fell 9.0%, but the group gross margin held at 55.7% (vs 56.5%). The margin dip was marginal and attributed to product sales mix shifts, not pricing erosion. [1] Segment profit declined to S$32.2 million from S$34.4 million, a 6.6% drop on a 9.0% revenue decline, which means the cost base flexed. At the group level, distribution and marketing expenses fell 22.7%, suggesting management pulled back on promotional spend rather than chasing volume at deteriorating returns. [1]

Pricing power holds at the gross-margin line, but the demand trend is weaker than a single soft half suggests, and we are less sanguine than at initiation.

Healthcare revenue fell about 7% in FY2025, and because H1 2025 was inflated by advanced orders, the underlying second half was materially softer; H1 2026's further 9% decline extends that run. [6] The weakness is regional, not universal: FY2025 sales in the Americas, Europe and the Middle East grew about 6%, while ASEAN fell 9% and the rest of Asia 15%. [6]

Tiger Balm is really two stories: it is growing in the West but losing ground in Asia, where competitive undercutting and soft sentiment are eroding volume. That is a real demand problem, not simply a brand riding out a one-off comparative.

Bull thesis (NAV accretion / asset discount): Validated; capital-return catalyst still unproven.

NAV per share rose 10.6% to S$21.43, driven by S$388.6 million in fair value gains on strategic investments flowing through other comprehensive income. [1] At S$15.06, Haw Par trades at about 70% of NAV, a 29.7% discount.

But the sharper way to frame the mispricing is this: at S$15.06 the whole company is capitalised at about S$3.33 billion, while its two listed stakes alone, UOB (S$2,975.3 million) and UOL (S$685.9 million), are worth S$3.66 billion.

The market is paying less for all of Haw Par than the quoted value of its bank and property-developer holdings, implying little to no value for the S$829 million of net cash and treasury bills, the Tiger Balm franchise, and the investment properties.

This is not costless free money, and the discount can persist indefinitely: the Wee-family control structure removes any takeover pressure, the stock is thinly traded, there is no buyback, and management has signalled no path to monetising the stakes. What the SOTP establishes is a floor, not a catalyst; the price is anchored to assets that are themselves liquid and marked to market daily.

On the payout, the ordinary dividend is steadier than the initiation implied. Haw Par has paid a 20-cent interim since FY2023, and it maintained that 20 cents here despite the earnings drop, so the 40-cent ordinary annual payout held from FY2023 through FY2025, and the FY2026 interim keeps Haw Par on that path. The genuine outlier was the FY2024 S$1.00 special (paid in H1 2025); FY2025 reverted to 40 cents, and this interim gives no signal on whether a special will accompany the FY2026 full-year results.

Bear thesis (passive conglomerate structure, no catalyst for discount closure): Unchanged.

Nothing in this filing addresses the structural discount. There is no share buyback activity, no corporate restructuring signal, and no change to the conglomerate holding structure.

The investment portfolio continues to appreciate (UOB +13.4%, UOL +9.3% in six months), widening the absolute NAV gap even as the proportional discount fluctuates. The fortress balance sheet, S$881 million in cash and short-term debt securities against S$52.6 million in borrowings, remains under-leveraged by any measure.

Flip-trigger status. Three events would force us to revise the call, and none tripped this period.

1. A UOB dividend cut in two consecutive years or a credit-quality deterioration (a sharp rise in non-performing loans): not triggered — UOB’s ordinary interim rose to S$0.88, 1H2026 net profit grew about 3%, and no asset-quality stress was flagged, so the lower FY2025 final is not yet a sustained ordinary cut. [5]

2. A Haw Par sale, demerger, or monetisation of Tiger Balm or the investment portfolio, which would be a positive catalyst: no such signal, and no buyback.

3. A regulatory change to banking-holding-company rules that forced a reduction in the UOB stake: no development this period.

The thesis stands unchanged against all three breakers.

Subscribe now

Answering the Questions We Opened in April

Our initiation left four questions explicitly open, and named four downside scenarios. Here is where H1 2026 leaves each.

Will the S$1.00 special dividend recur? On the evidence, no. It was a one-time FY2024 event; the regular payout is 40 cents (20c interim + 20c final). Specials, if declared at all, come with full-year results, so this interim offers no fresh signal. But the burden of proof has flipped: a repeat special is now the exception that would need explaining, not an emerging norm.

How would dividend policy evolve? The floor held: the 20-cent interim was maintained despite a 25% earnings drop, confirming the step-up from the old 15-cent interim is durable. But the growth our initiation extrapolated did not appear. The income line is now visibly hostage to UOB’s board, not Haw Par’s.

Is Tiger Balm gaining Western-market traction? Unanswerable from this filing; the interim carries no geographic split. Healthcare segment revenue fell 9.0%, which management ties to demand softness and a high comparative base rather than a retreat in the international push. It stays genuinely open, to be revisited at the FY2026 full-year, when the geographic breakdown returns.

Any restructuring signal from the Wee-family / UOI structure? None in this filing. No buyback, no structural change, no distribution of the Haw Par stake. The discount-closure catalyst the initiation was watching for remains absent.

Of the four downside scenarios we flagged, the income-volatility risk has moved from theoretical to demonstrated, though H1 showed it as special-dividend normalisation rather than a cut to the ordinary stream. The concentration risk, UOB as the dominant swing factor in both value and income, is no longer abstract. It produced this period’s headline.

Risks Revisited

The primary risk flagged in the initiation, dependence on UOB’s dividend policy, became visible this period, but H1 flattered the bear case. Most of the income drop was the non-recurrence of UOB’s anniversary specials plus a lower ordinary final, not a structural cut; UOB’s higher FY2026 interim confirms the ordinary stream is intact. [5] The real, unchanged risk is the shape of the dependence: a genuine shift in UOB’s capital-allocation priorities (higher CET1 buffers, buybacks over cash dividends) would compress Haw Par’s reported earnings with no offsetting operational lever. That has not happened, but this period is a reminder of how directly it would flow through.

A second risk worth noting: the healthcare revenue decline coincides with management’s commentary on “persistent inflation from prolonged geopolitical conflicts” and “trade policy uncertainties.” [1] The “both engines soft at once” scenario has eased on the income side, since UOB’s ordinary FY2026 interim came in above the prior year, leaving the healthcare trajectory as the swing factor for H2 2026.

Interest income also halved (S$6.0m vs S$12.2m), reflecting the declining rate environment. The shift of S$169 million from cash into Singapore Government Treasury Bills (debt securities rose from S$43.3m to S$212.3m) is a partial offset, with management actively locking in yields. [1]

What to Watch

Three forward indicators will determine whether this is a cyclical soft patch or a structural downshift:

First, and now answered, UOB’s FY2026 interim dividend. UOB has declared S$0.88 per share, payable 28 August 2026, up from the S$0.85 ordinary interim a year earlier, with 1H2026 net profit up about 3%. [5] That payment lands in Haw Par’s second half and confirms the H1 income drag was a timing-and-specials effect rather than a structural cut. What remains open is UOB’s FY2026 final (declared February 2027): the FY2025 final of S$0.71 was the figure that hurt Haw Par’s first half, so whether the next final rebuilds or holds near S$0.71 will set the direction of Haw Par’s 2027 income.

Second, Haw Par’s H2 2026 healthcare revenue trajectory. Management flagged the H1 2025 comparative base as elevated by advanced orders. If H2 2026 revenue recovers toward S$115–120 million (closer to FY2025’s S$210 million annualized run rate), the healthcare softness was base-effect, not demand erosion.

Third, the FY2026 full-year dividend decision (expected February 2027). FY2024’s total payout was S$1.40 (20c interim + 20c final + S$1.00 special); FY2025 reverted to 40 cents (20c + 20c, no special). If FY2026 delivers at least 40 cents total (20c interim already declared + 20c final), the 40-cent floor is confirmed. Any special dividend on top would signal the elevated payout is more than a one-off.

Bottom Line

Haw Par’s H1 2026 is a split-screen result. The income statement shows profit 25% below a year ago, but the cause is narrower than it looks.

The year-ago half was inflated by UOB’s 90th-anniversary special dividends; strip those out and the ordinary dividend stream is intact and, on UOB’s just-declared S$0.88 FY2026 interim, still edging up. The balance sheet tells the louder story: net assets grew S$452 million in six months to S$4.74 billion, the strategic portfolio’s carrying value rose S$409 million (S$388.6 million of it fair-value gains), and the company still sits on S$881 million of cash and treasury bills against just S$53 million of borrowings.

The thesis from April stands, with one refinement. The investment income is not a straight-line “rising machine,” since its year-to-year cash can swing on UOB’s special dividends, but the ordinary stream is stable and slowly rising, and the H1 drop overstated the underlying trend.

What has not changed is the structural mispricing: at S$15.06 the market values all of Haw Par below the quoted worth of its UOB and UOL stakes alone (S$3.66 billion), implying little to no value for the S$829 million of net cash and treasury bills, the Tiger Balm franchise, and the properties. The portfolio’s market value has continued to appreciate (UOB up 13.4% in six months), healthcare still earns 55%-plus gross margins, and the discount to NAV, 29.7% at S$15.06 against NAV of S$21.43, remains wide.

The interim dividend of 20 cents, maintained despite the earnings decline, suggests management intends to hold the 40-cent ordinary floor that has been in place since FY2023. Whether a special dividend is repeated for FY2026 depends on how UOB’s own capital return evolves.

For now, Haw Par remains what it was at initiation: a deep-value holding where the passage of time, UOB dividends, and portfolio appreciation do the work.

Data Integrity Notes

• Share price of S$15.06 is based on Google Finance market data as of the 21 August 2026 close.

• UOB dividend per share history (2024–2026) sourced from Stock Analysis (stockanalysis.com), cross-checked against the dividend income decline reported in Haw Par’s filing. The April 2026 UOB DPS of S$0.71 is consistent with Haw Par receiving ~S$53 million in UOB dividends, which aligns with the S$75.2 million total dividend income (with UOL and other sources contributing the remainder).

• The filing has not been audited or reviewed (per Note 15 of the filing). [1]

• Haw Par’s UOB shareholding (~74.8 million shares) is carried forward from the initiation article based on FY2025 annual report data. The H1 filing does not disclose the exact number of UOB shares held, only their fair value (S$2,975.3 million). At UOB’s 30 June 2026 closing price, this is consistent with the ~74.8 million figure.

• UOL shareholding (~72 million shares) is similarly carried forward. The UOL stake at S$685.9 million fair value is consistent with ~72 million shares at UOL’s 30 June price.

• No forecast or prospect statement was previously disclosed for H1 2026. [1]

• Dividend-year attribution: our April initiation labelled the S$1.00 special as an FY2025 dividend. This update follows the company’s own Note 8 attribution: the special was declared with FY2024 results and paid in H1 2025; FY2025’s attributable payout was 40 cents. The S$1.40 “FY2025” figure in the initiation reflects cash paid during calendar 2025 rather than FY2025’s financial-year distribution. [1]

• UOB dividend basis: UOB per-share dividends here separate ordinary from special components. UOB’s FY2025 ordinary dividend was S$1.56 (S$0.85 interim + S$0.71 final); the S$0.25 tranches paid during 2025 were part of a S$0.50 90th-anniversary special. The FY2026 interim of S$0.88 (payable 28 August 2026) and 1H2026 profit growth are from UOB’s own results release. [5] Calendar-year “cash received” totals (e.g. ~S$2.27 across 2025) mix fiscal years and specials and are not used here as the ordinary run-rate.

• Haw Par share price of S$15.06 and derived market cap / SOTP discounts are as of the 21 August 2026 close; the H1 FY2026 financial figures are as of the 30 June 2026 balance-sheet date.

References

[1] Haw Par Corporation Limited, “Condensed Interim Consolidated Financial Statements for the Half Year Ended 30 June 2026,” SGX filing, 14 August 2026. Announcement ID: RCSKDYV9X0P0K4OA.

[2] UOB dividend history: Stock Analysis (stockanalysis.com/quote/sgx/U11/dividend/) and UOB “Shares & Dividends” (uob.com.sg/investor-relations/shares-and-dividends/dividends.html), accessed August 2026. UOB FY2025 ordinary dividend S$1.56 (S$0.85 interim + S$0.71 final); the S$0.50 90th-anniversary special was paid in S$0.25 tranches across 2025; the April 2025 receipt of S$1.17 = S$0.92 FY2024 final + S$0.25 special tranche.

[3] Haw Par Corporation Limited, “Cash Dividend/Distribution — Mandatory: First & Interim Dividend of 20 cents per share,” SGX filing, 14 August 2026. Announcement ID: 5F8JVXONDOBYHLLC.

[4] “Haw Par Corporation: The Tiger Balm Conglomerate Hiding a S$3.4 Billion Portfolio,” The SEA Analyst, published 4 April 2026 (prior initiation article).

[5] United Overseas Bank Limited, “1H2026 financial results,” UOB newsroom (uobgroup.com/uobgroup/newsroom/news-releases/2026/uobgroup-1h26-financial-results.page), accessed 18 August 2026. FY2026 interim dividend of S$0.88 per share, payable 28 August 2026; 1H2026 net profit up ~3% year-on-year; interim represents ~50% payout ratio.

[6] Haw Par Corporation Limited, FY2025 full-year results and Annual Report 2025 (year ended 31 December 2025), SGX filings. Healthcare revenue down about 7% to ~S$240 million; by region, Americas/Europe/Middle East +6%, ASEAN −9%, rest of Asia −15%.

IMPORTANT DISCLAIMERS

This article is published for informational and educational purposes only. It does not constitute financial or investment advice, a recommendation, or a solicitation to buy, sell, or hold any securities. The author is not a licensed financial adviser under the Financial Advisers Act 2001 of Singapore. This content is exempt from the requirements of the Singapore Financial Advisers Act pursuant to Regulation 34 of the Financial Advisers Regulations, as it is published in a generally available publication. This publication does not take into account the specific investment objectives, financial situation, or particular needs of any individual. Before making any investment decision, readers should consult a licensed financial or investment adviser in their relevant jurisdiction. Past performance is not indicative of future results.

No Warranty: While the data and analysis have been prepared in good faith from public sources believed to be reliable, no representation or warranty, express or implied, is made as to the accuracy, completeness, or timeliness of the information. The author accepts no liability for any loss arising from the use of this material.

Disclosure: The author holds no position in the securities discussed. No compensation has been received from any company mentioned in this article.

The story is what management is doing with a cash-generative Singapore core: committing roughly US$113.2m to buy dental chains in Australia and Thailand [8], funding the deals with a mix of debt and freshly issued stock, and deferring its own secondary listing to press ahead [9]. Alongside that, its China-listed subsidiary Aoxin has doubled its share base to keep expanding, and the group has quietly agreed to sell most of its dental-AI arm for scrip.

The 27% headline is the least interesting number in the half.

The interesting one is that Q&M is committing close to a third of its market value to turn a Singapore compounder into a pan-Asian roll-up, and betting a lean team can integrate it.

Where the thesis stood

At initiation we called Q&M a coin flip at around S$0.60: a defensive, recurring Singapore dental platform wrapped in a multi-country acquisition story whose execution was unproven. The local press had already flagged that acquisition-led shift [7].

Three things carried weight:

1. the Singapore business as the durable core;

2. Aoxin as a China call option that could equally be a capital sink; and

3. the acquisition war chest as optionality the market was not paying for.

The core has held up. What this half does is turn the acquisition optionality into committed capital, which is a different risk.

The core that makes the bet possible

What makes the acquisition programme possible is the core, though not by paying for the deals directly. The core dental business earned S$15.0m of profit after tax this half, up 10%, on revenue of S$98.6m, and it throws off cash: the group closed the period with S$119.2m of it.

That cash generation and the balance-sheet capacity behind it are what let Q&M raise the S$130m of notes and issue stock to fund the offshore purchases; the acquisitions are paid in debt and shares, not out of half-year profit. The self-funding quality of the core is the whole basis of the thesis, and it is intact.

The group’s S$4.9m bottom line is what is left once the core carries everything else.

Start with the S$15.0m the core dental business earned: corporate head-office costs and unallocated group interest take S$8.9m of it, the other businesses (family medicine and aesthetics) lose S$1.1m, and minority interests, mainly the outside share of Aoxin, take another S$0.2m. Finance costs alone rose 57% to S$3.7m, the cost of the S$130m of notes issued last July.

Two one-off items also lift the 27%. This half’s S$5.4m of pre-tax profit includes a S$1.4m gain on the sale of the Jurong East property, about a quarter of the total; the year-ago S$4.8m base was the mirror image, held down by a S$4.3m non-cash loss booked when Aoxin and EM2AI were consolidated. The 10% rise in core-dental profit is partly bought rather than earned, too: this half carries six months of Aoxin against only one a year ago. Taken together, it is not the clean underlying growth it appears to be.

What does the real work is the S$15.0m core dental line, led by Singapore and Malaysia with the consolidated China clinics alongside. That is the annuity Q&M is now leveraging.

The full analysis below (the sizing of the bet, the deal structure, the Aoxin and EM2AI reads, the triggers, and the bottom line) is for paid subscribers.

Read more

The top line is clean. Revenue grew 11.9% to S$855.4 million, gross profit 15.6% to S$272.4 million, and earnings per share reached 5.38 cents against 4.81 a year earlier. Sixteen new and comparable-new stores supplied 9.7 of the 11.9 points of revenue growth, comparable same-store sales added 3.3% on the back of June’s CDC vouchers, and the network now stands at 90 supermarkets in Singapore plus six in Kunming [1].

That is the quality leg of the thesis, delivered in full.

The price leg is untouched, and the cash question is only now beginning.

What we said in April

We initiated on Sheng Siong on 30 April 2026.

The business was never the question. Eleven straight years of dividends, gross margin that had climbed six points over a decade to 31.3%, mid-single-digit revenue growth turning into roughly 10% earnings growth, a share count unchanged since the 2011 IPO, and S$435.5 million of net cash. The question was the price.

At roughly 31x trailing earnings and just under 8x book, the stock left, in our words, limited room for it to be wrong. We reduced the whole case to one variable: whether the gross-margin engine of sales mix, house brands and direct sourcing could keep grinding higher. At that price, the market was underwriting that it would, with little room for a slip. The flip triggers we have tracked on the name turn on two things:

1. a dividend cut or a capex step-up that thins dividend cover, and

2. a deterioration in the Group’s HDB-tender competitiveness.

   The SEA Analyst is a reader-supported publication. To receive new posts and support our work, consider becoming a subscriber.

The numbers

The first-half scoreboard validates the quality leg of the thesis and does nothing to relieve the valuation leg.

Three things sit underneath the headline.

First, the gross-margin gain is real and it is the number that matters most to this thesis. Management attributes the move to 31.8% to sales-mix improvement, and the mix effect held even as the cost base rose.

Second, the operating leverage below the gross line was partly absorbed. Selling and distribution expenses rose 15.7% to S$149.7 million and administrative expenses rose 14.0% to S$33.4 million, both driven by staff costs, which climbed to S$142.3 million from S$121.5 million on higher headcount for new stores, higher variable bonuses, and the September 2025 Progressive Wage Model increase. Operating profit still grew 15.7% to S$98.0 million, ahead of revenue, so the margin story survived the wage ratchet this half.

Third, cash and the cash flow statement need a second look. Operating cash flow for the half was S$54.8 million, down from S$86.4 million a year earlier. That decline is a working-capital swing, not an earnings-quality problem: trade and other payables fell S$68.7 million as the Group paid out accrued bonuses and settled vendor balances earlier than in the prior year, a timing item that sits inside the operating line.

Separately, the overall cash balance fell S$33.2 million over the half, to S$402.3 million at 30 June 2026 from S$435.5 million at 31 December 2025, as the operating inflow and modest S$8.4 million of capital expenditure were outweighed by S$57.1 million of dividends and S$25.7 million of lease payments in the financing line.

The Group still carries zero interest-bearing debt today, though a build of Sungei Kadut's scale may ultimately introduce some borrowing; the S$159.2 million of lease liabilities is store-lease accounting, matched by right-of-use assets on the other side. Finance income fell 44.5% to S$3.1 million as fixed-deposit rates came down, a small but growing headwind on a S$400 million cash pile as rates ease.

Step back from the half and the pattern is the real story. Gross margin has risen in every year of the past decade, from 24.7% in FY2015 to 31.3% in FY2025 [4] and 31.8% this half, six-plus points of expansion with no share issuance and no acquisition doing the work. That grind, sales mix, house brands and direct sourcing compounding quietly, is the engine the whole thesis rests on.

Testing the thesis

Read more

The share price went the other way. It had climbed into the result, from US$1.18 on 7 August to US$1.33. After the numbers came out it reversed, sliding to US$1.26 by 14 August and on to US$1.15 by 19 August, near a 52-week low and below where the run had begun [3].

The reason for the gap between a confident headline and a falling price is not in the announcement. Part of it is in the accounts, where most of the profit growth is a currency swing that will not repeat and the operating margin actually shrank. The larger part is not in the filing at all: it is the thing the company left out, and the bigger story here.

One point the company has genuinely proved is worth stating plainly, even though it is not what moved the stock. Since 2024 it has raised the US price of the dye by about a quarter and now sells about a fifth more of it, a large increase met with more volume rather than less. Half of that you can see only by rebuilding a chart the company has stopped publishing.

The rest of the numbers, quickly. Sales rose 24% to US$87.2 million, or 31% excluding the software business sold last year. Gross margin improved to 86.6%. Vials shipped rose 11% to 589,511: America 377,616 (up 4%); EMEA 196,165 (up 24%); APAC 15,730 (up 45%). Cash and short-term liquid investments finished at US$197.6 million with no borrowings excluding lease liabilities. Full-year sales guidance narrowed to US$175 million to US$185 million. And the board declared the first post-listing dividend, one US cent a share, payable 4 September.

What we said in June

We started covering UltraGreen on 15 June.

Our argument was that 2025 profit looked worse than the business really was, for two dull reasons: an US$8.5 million tax charge that might get refunded, and a full year of almost no interest earned on IPO cash that only arrived in December. Fix both and we got to roughly US$87 million of profit for 2026.

We also put a stake in the ground. If second-quarter vials came in below 248,000, our explanation for a weak first quarter fell apart.

And we flagged the one thing the whole case depended on. UltraGreen had raised American prices three times: 60% in August 2023, 30% in April 2024, and again in the third quarter of 2025. American hospitals do not buy drugs one at a time. They buy through large purchasing groups that negotiate on behalf of hundreds of hospitals at once. Whether UltraGreen’s increases survived those negotiations was, in our words, “the single most important undisclosed structural variable in the filings.”

We named a second risk too, and set it right behind the first. In June the field was empty: UltraGreen held both US approvals and the FDA’s own register showed no rival, so a hospital paying the higher price had nowhere else to go. We put a generic entrant on the short list of things that would break the thesis, calling it the event that would “change the pricing architecture permanently.” We thought it distant. It was not.

So we came into this filing with two questions left open from June: the buying groups, and the generic.

It answers both. One held. The other happened. Neither answer is in the part the company wrote for journalists.

The numbers behind the headline

Two things to clear up before we get to the interesting part.

First, that 45%. Almost all of it is a currency swing that will not repeat. The company lost US$7.02 million on currency moves in the first half of 2025, and almost nothing in the first half of 2026. The 45% sets this year’s clean profit against last year’s currency-hit profit, so the growth rate looks bigger than it is. Strip the currency noise out of both halves and profit grew about 16%, which is exactly what the company’s own adjusted EBITDA grew. The company all but concedes the point: its slide credits the 45% to “significantly lower foreign-exchange losses recorded in 1H2025.” The figure is real arithmetic, but it is about three times the growth the business actually delivered, and the cause is already fixed, with every subsidiary moved to dollar reporting in January.

Second, interest income. It was US$2.47 million for the half, up from just US$67,008 a year earlier. Double it for a full year and it reaches about US$5 million, the bottom of the US$5 million to US$7 million we forecast in June. We were right that the IPO cash would throw off real interest income. Our range was just a little too high.

Did the thesis hold?

Volumes: yes, comfortably

The half-year filing gives six-month totals only, but the company published a full quarterly chart back in May [1], so subtraction does the rest. Second-quarter vials: 308,611. Our line was 248,000. It cleared it by 24%. Against the same quarter last year it is up 38%. It is the biggest quarter the company has ever had.

So our explanation for the weak first quarter holds. But the number that actually matters is the American one, and that is the number the company has stopped giving.

That is one thing the company left out of this filing. There is a bigger one, the very risk we flagged in June, and it has nothing to do with volumes.

In the roughly five weeks before these results, the US Food and Drug Administration approved two competing indocyanine green products.

• On 10 July it cleared a ready-to-use version from Provepharm called ZYOGREEN, now approved and listed for the US market [6][9].

• On 3 August, nine days before UltraGreen reported, it approved a straight generic from Zydus, one of India's largest generic drugmakers, under a programme the agency runs for markets it judges short of competition; Zydus announced it the next day [7][8].

For a company whose whole worth rests on being the only place to buy this dye at these prices, that is the most important development of the year, and it is the one thing the announcement does not mention.

That is the week the story changed. What it does to everything else in this filing is what the rest of this update is for.

The rest of this update is for paid subscribers: what those two FDA approvals do to the thesis, the reconstructed US quarterly series the company stopped printing, the seam answer (where the price increases actually went), the operating-leverage teardown, the tax, dividend and governance record, and the valuation at US$1.15.

Read more

That number is the S$1.2 million of profit the company reported for shareholders, and the results-day headlines led with it [1]. It looks alarming, but it is misleading. It comes from an old ownership structure that ended on 1 July, the day after this period closed. It will not be there next time.

The number that matters is the margin, and it slipped. The business kept less profit from each dollar of sales than a year ago. Adjusted profit fell 17% to S$16.1 million, even though revenue rose 20%. The adjusted EBITDA margin dropped from 37.3% to 30.8%. That is what the market reacted to, and it is the number the whole valuation is built on.

The result came out on 13 August, after the market closed. The shares fell from S$0.805 to S$0.740 the next day, about 8%, then recovered over the following two sessions to S$0.775 by 17 August, back above the S$0.76 they listed at on 8 July. The half also carried S$5.4 million of one-off IPO and acquisition costs, against S$0.6 million a year earlier.

Prior Thesis Recap

We first wrote about Foundation Healthcare on 29 June, while the IPO was still open [2].

Our point then was simple. This is a good business, but it was coming to market through a structure that flatters its earnings just before listing. In FY2025 the group earned S$41.2 million, but only S$14.0 million reached shareholders. The doctors still owned 40% of most of their own practices, and took their share first. A pre-IPO deal called the Share Swap, done on 1 July, bought those doctors out and moved that profit to shareholders. We said the swap was struck cheaply and was good for new investors. We also said the real thing to watch was not the fast revenue growth, but what happened to margins once the practices sat inside the group. At the time we noted the stock listed at 21 to 25 times its pro forma earnings and 13 to 15 times pro forma EBITDA [2]; on the group's pro forma adjusted profit of S$51.4 million, the swap-neutral base we use below, that is about 20 to 24 times. This is the first result we can test that against.

New Financial Data

Here is the thing to understand about that S$1.2 million, because it explains this result and the next one.

Profit is counted before it is split between the group and the doctors who own minority stakes. The margins, the revenue growth, the adjusted profit, all of it is measured before that split. The Share Swap only changes the split. It takes the doctors’ share and gives it to shareholders. It does nothing to the profit the business makes in the first place.

So the S$16.1 million of adjusted profit the group earned this half is the same whether the doctors own 40% or nothing. This half ended on 30 June, one day before the swap. The old split still applied, so the doctors’ share, the non-controlling interests, took S$11.5 million of the S$12.7 million reported. And the S$5.4 million of acquisition and capital-raising costs, mostly the IPO, landed on shareholders too. That is how a business earning S$16 million of adjusted profit reports S$1.2 million to its owners.

Next half, the swap runs the same trick in reverse. The doctors’ share is gone, so shareholders keep almost all the profit, and the headline earnings will jump. That jump will look like a big recovery. It will not be one. It is just the swap.

The simplest way to hold this in your head: the swap changes the size of your slice, not the size of the pie. This half it made a decent pie look like a sliver. Next half it will make the pie look bigger than it grew. If you want to know how the business is really doing, ignore the slice and look at the pie. And the pie got thinner this half.

That is the real news. The business made less profit on each dollar of sales. Adjusted profit margin fell from 18.1% to 12.5%, so even with 20% more revenue, adjusted profit went down.

This matters because of what the market paid for.

Read more

Update on 2 September 2026

Since this update published, the catalyst we described as quiet has moved. On 1 September Jollibee disclosed that it is contemplating a proposed Hong Kong Main Board listing of its international business, to be held by the listing vehicle, Jollibee Foods Corporation International (JFCI).

Existing JFC holders would receive JFCI shares corresponding to their prevailing JFC interest, effectively an in-specie distribution. Richard Shin, JFC's Chief Financial and Risk Officer and CEO of JFC International, was named to lead JFCI. JFC itself stays PSE-listed with the Philippine business.

The company calls the plan preliminary, subject to restructuring, market conditions, and required corporate, regulatory, and HKEX Listing Committee approvals, with no assurance on final terms, timing, or completion; no size or valuation was given. JFC closed at ₱152.80 on 1 September, below the ₱156 close after the August results and above our ₱147.10 initiation anchor. [6]

This advances the thesis rather than upsetting it.

The venue-mirage argument was that a US listing cannot hold a US growth multiple; picking Hong Kong over a US listing makes a US-style re-rating far less likely, since Hong Kong-listed Asian restaurant and coffee names trade nearer Asian marks.

We read the choice as pointing toward crystallisation at roughly the arm's Asian value, not proving it, since venue alone does not settle valuation.

The pro-rata structure looks friendlier than feared for a Manila holder, who gets the shares directly rather than through an access wrapper, though tax, currency, and custody frictions remain. And the announcement is silent on the JWPL 2025 notes' change-of-control put, under which holders can require JWPL to repurchase their notes at 101 percent upon a Change of Control Event, including if JFC ceases to own, directly or indirectly, at least 50 percent of JWPL's outstanding voting stock, a real constraint if JWPL sits inside the separated group [4][7].

The muted response fits our frame: a listing may crystallise value, not multiply it. We will cover the mechanics when the circular or notice to bondholders appears.

Subscribe now

This is an update to our 22 July 2026 initiation. It is for informational and educational purposes only, is not investment advice, and its author is not a licensed investment adviser. The author holds no position in Jollibee Foods Corporation and has not traded in it in the 30 days before publication.

On 11 August 2026 Jollibee led with a record. Net income to the parent (NIAT) of ₱3.40bn, up 5.7 percent on the year and up 130.5 percent on the first quarter, the best quarter the group has ever printed [1]. Revenue rose 10.7 percent to ₱85.9bn. System-wide sales grew 14.2 percent. The gross margin clawed back to 18.5 percent from 16.5 percent in Q1 and left June at 19.0 percent [1]. Same day, the board approved redeeming the ₱9bn Series B preferred (JFCPB) and trimmed parts of the full-year guidance [1][2].

The market liked it. Shares jumped about 5 percent to ₱156.00, above the ₱147.10 we anchored three weeks earlier [5]. Fair enough on the quarter. The half is another story: first-half NIAT was ₱4.87bn, down 13.3 percent year on year, operating income down 7.1 percent, net income down 16.7 percent [1]. So the record Q2 is a bounce off a weak, cost-pressured Q1, not a higher earnings base. And look under the operating line. Q2 income before tax actually fell 1.7 percent, to ₱4.84bn. NIAT rose anyway, because the tax rate dropped from 30.6 to 27.3 percent and the minority share fell 39 percent [3]. That gap, between the headline and what sits beneath it, is what this update is about.

Prior thesis recap

Our 22 July initiation made three claims.

1. The profit is almost entirely Philippine. Two-thirds of the stores are overseas, yet the international arm loses money at the net line while the home business pays for the build-out.

2. This is not the net-cash compounder some assume. At FY2025 the group carried about ₱86.5bn of debt against ₱34.9bn of cash, net debt near ₱52bn, plus ₱52.8bn of leases.

3. The “venue mirage.” A US listing of the arm can crystallise value but cannot sustain a US growth multiple, so its value is already explainable on Asian marks and needs no re-rating.

No price target. The sum-of-the-parts ran ₱145 to ₱225 on limited debt, ₱114 to ₱194 once the arm carries its fuller net debt. This quarter tests all three.

New financial data

The margin recovery is real, and we said as much. Pricing from April, plus sourcing and cost discipline, took the gross margin from 16.5 percent in Q1 to 18.5 percent in Q2, and June exited at 19.0 percent gross, 9.1 percent operating and 6.2 percent NIAT, the best run-rate of the year [1]. Group EBITDA rose 7.3 percent. No argument there.

It just was not broad. The EBITDA growth came from the Philippines, where EBITDA rose 12.8 percent. International EBITDA fell 0.4 percent, weighed down by ₱239m of transition costs as Smashburger and Yonghe King close stores and move to franchising [1]. Take out the home engine and the overseas arm grew its earnings by nothing this quarter.

Thesis check

Bear: “the profit is almost entirely Philippine.” Validated. The quarter says it in the segment lines. Philippine EBITDA up 12.8 percent, international EBITDA down. And yet international system-wide sales grew 25.4 percent, led by Highlands Coffee (+46.7 percent), Compose (+39.7 percent), Tim Ho Wan (+23.0 percent), Jollibee North America (+21.6 percent) and Milksha (+12.4 percent) [1]. Fast sales, no earnings. That is the shape we drew: a coffee-led arm that is real and growing and does not yet pay the group’s bottom line. The bull points at the 25 percent. The income statement still says Manila.

Bear: “not net cash; leverage is real and rising.” Validated, and worse. The balance sheet went the wrong way.

• Gross interest-bearing debt rose to ₱93.2bn at 30 June from ₱86.5bn; cash fell to ₱29.8bn from ₱34.9bn.

• Net debt ex-leases climbed to about ₱63.4bn from ₱51.6bn, up roughly ₱12bn in six months, near 1.5 times annualised EBITDA against about 1.2 at year-end [1][3].

The composition is the tell we flagged: senior notes fell ₱16bn to ₱36.6bn as JWPL redeemed the shorter 2020 tranche, but bank debt rose ₱21.9bn. “Partly refinanced with bank borrowings,” on the page. Financing interest rose 9.8 percent [3]. The record NIAT swallowed that bigger interest bill; it was not handed a smaller one.

Bear: the ₱9bn preferred redemption is another call on that same cash. Redeeming all nine million Series B preferred at their ₱1,000 price, about ₱9bn, on 14 October helps common holders a little in isolation: out goes a claim and a dividend that sit ahead of them [2]. But the balance sheet is already tighter than six months ago, and this is the same cash the change-of-control bond put we analysed at initiation could one day demand [4]. Pay it from cash and group cash drops toward ₱21bn. Pay it from debt and net debt heads toward ₱72bn. Either way, the liquidity we called committed gets more committed.

Bull: the coffee-led engine is real. On sales, not yet on profit. Highlands and Compose are compounding system-wide sales north of 40 percent. Jollibee Vietnam is the standout, same-store sales up 17.9 percent, store payback under four years. The group folded in Shabu All Day, a Korean brand under Jolli-K (172 stores acquired, 156 in the network at 30 June), about 5 percent of international sales [1]. The asset-light turn shows up where it matters: China’s franchise ratio is 62 percent, from 14 percent in 2016; Yonghe King is at 65 percent and aiming for 70; Canada added franchise development deals. This is the honest, non-venue route to a higher multiple we named as a trigger. It has not landed. International EBITDA is still flat-to-down, because the very restructuring that makes the arm asset-light costs money now. The winners win. The arm, as a whole, does not yet pay.

Flip trigger: “Compose and the coffee platform inflect to clear, sustained group-level profitability.” Not crossed. The one thing that would hurt us most is the overseas arm turning profitable at the net line. It went the other way. International EBITDA fell. We are watching for the quarter when “the profit is almost entirely Philippine” stops being true, and this half moved away from that, not toward it.

Bear: the venue mirage. No real update; the catalyst stayed quiet. Fifteen pages on record earnings, margins, capital-light growth and awards, and not one on the separation or the US listing, the centre of the whole bull case [1]. The 17-Q is barely louder: a routine subsidiary note tied to the planned spin-off, with no venue, no timing, no strategic-review update [3]. In January this was the headline. This quarter it is a footnote. One release is not cancellation, and we will not pretend it is. But it fits the drift we mapped, firm January plan to hedged June review, and it is not how a company markets a re-rating. A mirage stays a mirage while nobody points at it.

Risks revisited

Two risks moved.

1. Leverage moved up. Net debt rose about ₱12bn in six months, the preferred adds another ₱9bn, and the flexibility is thinner, so the change-of-control put, if a separation ever fired it, would hit a tighter balance sheet.

2. Execution moved down, a little: the margin recovery and the franchise discipline in China and North America show management can react to cost and will eat near-term charges to get asset-light.

Against that, the guidance cut. Full-year same-store growth trimmed to 3-4 percent, gross openings to 1,000-1,100 stores, operating-income growth to 10-15 percent, a quieter outlook than a year ago [1].

What to watch

Three things.

1. International EBITDA first: the quarter to wait for is the one where the Smashburger and Yonghe King costs roll off and the coffee winners are big enough to lift overseas earnings, not just sales. Until then the trigger holds.

2. The separation second: the next annual meeting, or any circular or notice to bondholders, answers the venue question, and its absence from results talk is itself worth logging.

3. The balance sheet third: the 14 October redemption, capex of ₱13-15bn, and whether net debt keeps climbing into the second half will tell us whether “committed liquidity” is easing or worsening. The 17 August briefing is the first chance for management to speak to any of it.

Bottom line

Call it a beat. We read it as a confirmation. Jollibee posted a real, home-led margin recovery and a record headline NIAT, and did it while the half was down double digits, overseas earnings went backwards, net debt rose about ₱12bn, and the US listing behind the bull case drew nothing but a footnote. None of that moves the initiation. The profit is still Philippine. The balance sheet is still levered, and more so. The venue mirage is quieter, not louder.

We stay non-directional, no price target. At ₱156 after the pop, the market sits in the lower half of our limited-debt band and above the fuller-debt one; the framework holds. JWPL’s own note balance fell as it redeemed a tranche, which helps the arm’s standalone equity a touch, but group net debt rose and a ₱9bn preferred call is booked, so the capital-structure-dependent range, with its wide gap between the two debt cases, matters more now, not less.

What would change our mind has not: an arm that turns durably profitable at the net line, or a separation document that gets more value, more cleanly, to a Manila holder than we expect.

Neither showed up this quarter. A record was set. It was set at home.

Data Integrity Notes

Q2 and 1H 2026 figures are taken from Jollibee’s 11 August 2026 press release and the accompanying SEC Form 17-Q; balance-sheet figures are from the unaudited consolidated statement of financial position as at 30 June 2026 in the 17-Q [1][3]. System-wide sales and same-store sales growth are management metrics and are not part of the audited financial statements. Net debt is our calculation (short-term debt plus current and non-current long-term debt plus senior debt securities, less cash and cash equivalents) and excludes IFRS-16 lease liabilities of ₱52.8bn. The ₱9bn preferred-redemption figure is nine million JFCPB shares at a ₱1,000 redemption price plus accrued dividends, per the 17-C redemption disclosure [2]. The share price of ₱156.00, up 5.05 percent on the day, is the PSE quote at the time of writing [5]; our initiation anchored ₱147.10 on 20 July 2026, so the stock has risen into the results and should be refreshed at publication. The results press release did not mention the US listing; the 17-Q carries a routine subsidiary disclosure tied to the planned international spin-off but no venue, timing, or strategic-review update [3]. This is our reading of the filings, not a company statement that the plan has changed.

References

[1] “The Jollibee Group Reports Record Q2 2026 Results, with Margin Recovery from Controlled Pricing and Record-High Quarterly Net Income Attributable to Equity Holders of the Parent Company,” Jollibee Foods Corporation press release (PSE 17-C, Form 4-31, C06118-2026), 11 August 2026.

[2] “Redemption of Series B Preferred Shares,” Jollibee Foods Corporation (PSE 17-C, Form 4-21, C06121-2026), 11 August 2026: 9,000,000 JFCPB shares at a ₱1,000 redemption price plus accrued dividends, less customary transfer costs, redemption date 14 October 2026, record date 30 September 2026, kept as treasury.

[3] Jollibee Foods Corporation SEC Form 17-Q for the quarter ended 30 June 2026 (PSE CR06002-2026), 11 August 2026: unaudited consolidated statement of financial position and statement of comprehensive income. Gross interest-bearing debt ₱93.2bn, cash ₱29.8bn, senior debt securities ₱36.6bn, long-term debt ₱40.6bn (current plus non-current), lease liabilities ₱52.8bn.

[4] The SEA Analyst, “Jollibee’s Global Arm Is Real. Its Wall Street Re-Rating Is a Mirage.” (initiation), 22 July 2026. https://www.theseaanalyst.com/p/jollibee-jfc-philippines-pse-stock-us-spin-off

[5] Jollibee Foods Corporation (PSE: JFC) share price of ₱156.00, up 5.05 percent on the day, the Philippine Stock Exchange quote as at 11 August 2026. https://www.pse.com.ph/company-information-JFC/

[6] Jollibee Foods Corporation, SEC Form 17-C / PSE disclosure, “JFC Advances Planned Spinoff of International Business; Selects Hong Kong for Proposed Listing and Names JFCI CEO,” 1 September 2026 (signed by CFO and Chief Risk Officer Richard Chong Woo Shin). https://edge.pse.com.ph/downloadHtml.do?file_id=1962201

[7] Jollibee Worldwide Pte. Ltd., Final Offering Circular dated 25 March 2025 (SGX): change-of-control put at 101 percent triggered if JFC as guarantor ceases to own at least 50 percent of JWPL’s voting stock. https://links.sgx.com/FileOpen/JOLLIBEE%20WORLDWIDE%20PTE.%20LTD.%20-%20Final%20Offering%20Circular%20-%20dtd%2025%20March%202025.ashx?App=Prospectus&FileID=65389

On the figures themselves the documents agree.

• Revenue for the six months to 30 June rose 24% to US$80.8 million.

• Cash EBITDA, the company’s measure of what the business generates after rent is actually paid and after corporate costs, rose 147% to US$10.6 million, lifting the margin from 6.6% to 13.1%.

• Free cash flow was US$3.2 million, against US$0.9 million a year earlier.

• The group ended the half with US$169.4 million of cash and no bank debt.

• It also reported a net loss of US$0.8 million, or a profit of US$0.1 million once one-off listing expenses are stripped out [6].

Two things came with the numbers.

1. The board intends to pay out 50% of net profit after tax from FY2027, the first dividend policy in the company’s history.

2. And the network reached 57 operational centres and 37,350 workstations, from 50 and 35,067 at the end of December, with 21 more centres described as committed.

The shares closed at S$0.630 on 6 August, before the release [13]. In the first session after it, on 7 August, they closed at S$0.640, up 1.6%, having traded between S$0.615 and S$0.660 on 1,282,200 shares against 455,100 the day before [1]. That leaves the stock 32% below the S$0.94 offer price of eleven weeks earlier.

What we said in May, and what we got wrong

Our 15 May initiation read the offer document [8] and declined to give a verdict on the price. The position we published was that both cases were true at once: a working operator with city-level density, and a structure carrying US$402.2 million of lease liabilities against US$40.4 million of equity, priced at roughly 130 times a maiden profit that leaned on a one-off gain. We wrote that the offer was “not a value entry point” and that it “asks the buyer to underwrite the next two years of expansion going broadly to plan”.

That piece named the dependencies but published no scorecard. It said that “every one of the bull points depends on the occupancy cycle staying friendly”, that “the maiden profit is flattered by a one-off”, and that “the raise itself adds 28 centres of fresh lease liability and the associated drag before those centres mature”. We are putting thresholds on those dependencies here, and adding one on how the expansion is paid for, as four tests of the franchise rather than of the price. This result is the first that can score any of them.

1. The occupancy test fails if group occupancy drops materially below 80%, or if the renewal rate reverses from the 72.0% it reached in FY2025.

2. Payback fails if new centres stretch well beyond the historical 16.4-month weighted average.

3. Profitability fails on the first post-listing result that shows an underlying loss once one-off items are stripped.

4. Funding fails on any sign that the expansion is being paid for by re-leveraging instead of by the IPO proceeds and internal cash.

Two things in that piece were wrong, and the second one we only found on re-reading the prospectus for this update.

The first is the call itself. We wrote that the post-listing share price would "partly reflect scarcity rather than fundamentals, which flatters it now and creates an overhang later". Instead the stock opened at S$0.835, closed its first day at S$0.775, 17.6% below the offer price, on 10.9 million shares [2], and has since traded as low as S$0.495 [1][3]. Scarcity supported nothing. The cohort was weak too, with all six companies that listed on SGX in 2026 trading flat or below their offer price as at 22 July [5], though that report gives no magnitudes, so we cannot say where JustCo's 32% ranks among them.

The second is the arithmetic underneath it. We described the float as 6.6% and said the cornerstone tranche would be released by a six-month moratorium in the fourth quarter. The offering was indeed 32,092,000 shares, or 6.56%. But the prospectus states plainly that “the Cornerstone Investors are not subject to any lock-up restrictions in respect of their shareholdings” [8]. Their 74,291,000 shares, a further 15.19%, could be sold from day one. The sellable pool at listing was therefore about 21.7% of the company, not 6.6%, and there is no cornerstone unlock ahead to price. We got that wrong in May and we are correcting it here.

The filings also show how thin the real demand was. DBS, as stabilising manager, bought 5,319,000 shares between listing and the close on 2 June, when it told the market that the over-allotment had been fully covered and the option would not be exercised [9]. That figure is 16.6% of the entire offering [10]. Through those first two weeks of trading, DBS bought back the equivalent of one share in six of the 32,092,000 Offering Shares. The price fell further once that stopped.

Where the growth came from

The revenue increase decomposes cleanly. Average occupied workstations rose about 12%, and revenue per occupied workstation rose 11%. Multiply the two and you have the 24%.

The extra workstations came from an acquisition. JustCo bought the remaining 51% of its Japan operations on 1 July 2025, so the first half of 2025 contains no consolidated Japan revenue and the first half of 2026 contains six months of it. Japan was an associate before that, so a share of its operating EBITDA does sit in the 1H2025 Cash EBITDA reconciliation, US$0.3 million of it, recorded in North Asia. That is the acquisition the review of performance names and the press release does not.

North Asia revenue went from US$15.3 million to US$27.1 million. That single segment accounts for 75% of the group’s entire revenue increase. On Cash EBITDA, North Asia went from US$1.1 million to US$5.4 million, or 68% of the group’s increase on the reported basis, and 73% if the equity-method share is stripped out of the 1H2025 base. Strip North Asia out and group revenue grew 7.7%.

Two operating metrics that JustCo published at the IPO are absent from all three documents it filed on 6 August: the membership renewal rate and the payback period on new centres [6][8]. For a business whose memberships run about 15 months against leases of three to fifteen years, those are the two numbers that say whether the model is working. Neither has been published since the prospectus.

Below the paywall: the half of the business with no acquisition in it more than doubled its Cash EBITDA. Rent took about 95% of the cash the business generated, leaving US$2.7 million. We score the result against four tests of the franchise, including the metric the company published at the IPO, and restate every valuation anchor at the 7 August close.

Read more

HL Global Enterprises reported its half year results on 3 August. Cash and bank balances rose again, to S$64.07 million from S$63.48 million at the year end. The shares are S$0.385 [1], which puts the market capitalisation at about S$37.1 million across the 96,334,254 shares in issue. The cash is therefore worth roughly 173% of the entire company, or 177% excluding the 2.4 million shares held in trust for a dormant option scheme, as the company does when it reports net asset value per share.

That gap has widened, not closed. When the Securities Investors Association (Singapore) put its questions to the board ahead of the April annual general meeting, it cited a market capitalisation of about S$45 million against S$63.5 million of cash. Four months later the cash is higher and the market value is lower.

The operating half was fine and beside the point. Revenue rose 12.2% to S$3.19 million on a better first half at the Cameron Highlands hotel. Net profit fell 10.0% to S$406,000: interest income on the cash dropped 30.7%, or S$252,000, as deposit rates came down, which was more than the S$237,000 improvement in the hospitality segment result. No dividend was recommended, in the same sentence the company has used in each of the last three interim filings.

What is worth reading is the board’s answer to SIAS. Asked directly whether share buybacks, special dividends or a broader strategic review had been formally evaluated, and on what timeline, the board replied: “While the Group maintains a positive cash position, it has not generated sufficient recurring profits in recent years to enable the declaration of dividends.” It added that it had considered various capital management and strategic options from time to time, and that “to date, no options have been identified that meet the Group’s strategic and financial criteria.”

Asked separately what drives the discount, the board listed the group’s size, limited trading liquidity, earnings profile, the concentration of its operating assets in a single hospitality location, and “the absence of dividend distributions in recent years as the Group conserves cash for operations and investment.” It named the absence of dividends as a cause of the discount, and then gave the absence of profits as the reason for the absence of dividends.

Both statements are accurate. The listed parent does carry accumulated losses of S$55.98 million and has no retained earnings to distribute from. But a shortage of distributable profits is a solvable problem rather than a permanent condition, and the mechanism is a capital reduction, which does not require distributable profits at all [2]. Two SGX-listed small caps used it in 2024, one to clear accumulated losses and one to hand back cash while loss-making [3][4]. HL Global used it itself in 1990, and the S$12.47 million special reserve still sitting on its balance sheet is what remains of that exercise. It has not proposed another since the 2017 disposals that created the cash pile, and it did not put a buyback mandate to the April meeting either.

Our position is unchanged. The market is not mispricing the cash. It is pricing the chance that anyone ever acts on it.

References

[1] HL Global Enterprises (SGX: AVX) share price and market data, at the 6 August 2026 close of S$0.385. https://stockanalysis.com/quote/sgx/AVX/

[2] Companies Act 1967 (Singapore), Division 3A of Part 4, sections 78A to 78K, on the reduction of share capital, including the court-approved route and the solvency-statement route. https://sso.agc.gov.sg/Act/CoA1967

[3] mDR Limited, “Circular to Shareholders dated 25 September 2024” in relation to a proposed capital reduction cancelling S$27,397,446 of accumulated losses against share capital, the company’s third such exercise after 2019 and 2021. https://links.sgx.com/1.0.0/corporate-announcements/OO9M37GKC7L839S3/819699_MDR_Capital_Reduction_Circular_25September2024.pdf

[4] Global Testing Corporation Limited, “Circular to Shareholders dated 5 April 2024” in relation to a proposed capital reduction and cash distribution of S$0.05 per share. https://links.sgx.com/1.0.0/corporate-announcements/HQ8T8ZBUH7YGHUYT/794834_GTC%20-%20Circular%20to%20Shareholders%20dated%205%20April%202024.pdf

Primary filings (SGX disclosures, no inline citation): HL Global Enterprises Limited, unaudited half year financial statements for the six months ended 30 June 2026 (3 August 2026); full year results for FY2025 (13 February 2026); half year results for 1H2025 and 1H2024; Annual Report 2025; Minutes of the Sixty-Third Annual General Meeting held on 21 April 2026 (19 May 2026); and Responses to Questions Received from the Securities Investors Association (Singapore) (15 April 2026).

IMPORTANT DISCLAIMERS

General Disclaimer: This article is published for informational and educational purposes only. It does not constitute financial advice, a recommendation, or a solicitation to buy, sell, or hold any securities. The views expressed are based on publicly available data from HL Global Enterprises’s SGX filings and publicly accessible market data, and may not reflect the most current developments.

Not Licensed Financial Advice: The author is not a licensed financial adviser, and this publication is not issued by a holder of a Capital Markets Services Licence under the Securities and Futures Act 2001 of Singapore. This content does not fall within the definition of “financial advisory service” under the Financial Advisers Act 2001 of Singapore. Readers in Singapore should note that this content is exempt from the requirements of the Financial Advisers Act pursuant to Regulation 34 of the Financial Advisers Regulations, as it is published in a generally available publication.

MAS Compliance Notice: In accordance with the Monetary Authority of Singapore’s guidelines, this publication does not take into account the specific investment objectives, financial situation, or particular needs of any individual. Before making any investment decision, you should consult a licensed financial adviser who can provide advice tailored to your personal circumstances. Past performance of any security discussed herein is not indicative of future results.

No Warranty: While the data and analysis have been prepared in good faith from public sources believed to be reliable, no representation or warranty, express or implied, is made as to the accuracy, completeness, or timeliness of the information. The author accepts no liability for any loss arising from the use of this material.

Disclosure: The author may or may not hold positions in the securities discussed. No compensation has been received from any company mentioned in this article.

The spike is real, and it’s big too. Revenues increased from $4.8 million during the fiscal year to December 2023, to $71.5 million during FY2024, and $74.1 million during FY2025. This is over fifteen times in two years. Not many publicly listed firms have a chart looking like this. But this type of a chart should rather increase the readers’ caution because such an incredible rate of growth rarely happens due to a lot of clients. Instead, it happened due to only one.

From a Shanghai supply-chain desk to an SGX listing

The company that will trade on the SGX Mainboard was incorporated in Singapore only on 24 December 2021, and became a public company as recently as 19 June 2026. It was set up as a joint venture between AGX Singapore, a subsidiary of the Bursa-listed logistics group AGX Group Berhad, and Mr. Xu Hao, a Chinese logistics entrepreneur. Mr. Xu Hao does not sit on the board. His wife does. Mdm. Tang Ying, formerly a vice-president at a Shanghai supply-chain company, is an Executive Director and, before the offering, the 70% controlling shareholder. AGX Singapore holds the other 30%.

The name came with the relationship. As per the prospectus, the name chosen by the company on incorporation was meant to indicate its origin and commercial relationship with All-Link PRC, a different Chinese-owned freight forwarding firm which Mr. Xu Hao owns through his 64.8% ownership and which the document is very careful to point out does not belong to the group to be listed. There are also at least two other firms trading under related names. Since then, All-Link has established its own customer base using the trade name, its own logo, and even filed an application for trademark registration of the name in Malaysia – the application remains pending – in the name of its Malaysian subsidiary rather than the company that is to be listed. Bear the relationship in mind, because it is the entire thesis. The listed company was incorporated to provide logistical services to the customers of All-Link PRC, moving the cargo of these customers into Southeast Asia and earning its money from doing so.

The operating history is short. Singapore operations began in 2022. A Philippines subsidiary was incorporated in September 2022. In 2024 the group was appointed a logistics provider to the TikTok Group. In 2025 it incorporated a Malaysian arm and, in August, bought the freight-forwarding business of a local operator called MF Logistics, which is the source of the only goodwill on its balance sheet. Headcount went from ten to seventy in about a year, and had reached 112 by the time the final prospectus was registered. This is a company that grew up around a single opportunity, very fast, and is now asking for permanent public capital on the strength of it.

What the offer actually contains

All-Link is selling 37,924,500 new shares at S$0.53 each, split into a 35,824,500-share placement and a small 2,100,000-share public tranche. The public offer opened on 28 July and closes at noon on 3 August, with trading expected to start on the Mainboard at 9.00 a.m. on 5 August 2026. Every share on offer is newly issued, so the money raised goes to the company rather than to the sellers, and the raise is sponsored, underwritten and placed by CGS International Securities Singapore. The prospectus expressly grants no over-allotment option, and discloses no cornerstone investor or tranche, which for a deal this size is worth noting rather than assuming.

At S$0.53, the 151,037,900 shares in issue after the offering value the whole company at about S$80 million. The offering represents 25.1% of that, so roughly three-quarters of the company stays with the people who already own it: Mdm. Tang Ying with 51.7% and AGX Singapore, and behind it the Bursa-listed AGX Group Berhad, with 23.2%. The public float is 25.1%, but the portion actually offered to retail investors through the public tranche is only 2.1 million shares, about 1.4% of the company. This is a placement-led listing with a thin retail slice, and the controllers keep firm control.

The proceeds, roughly S$17.7 million after expenses, are earmarked for expansion, for “strategic acquisitions” (part of which is intended to buy out Mr. Xu Hao’s 30% stake in a related Vietnamese operation), for technology, and for working capital to pay airlines and carriers. None of it goes to repaying debt, because there is essentially none.

One item in the offer deserves to be read slowly. In March and June 2026, in the months before the listing, the company declared a US$8.0 million dividend in respect of FY2025, split as a US$2.0 million interim and a US$6.0 million final payment. Roughly US$5.6 million of that flows to Mdm. Tang Ying and US$2.4 million to AGX Singapore. That US$8.0 million is about S$10.3 million, or close to 13% of the entire post-listing market value, paid out to the controllers on the way in. On a pro forma basis it cuts the group's cash from US$23.6 million to US$15.6 million and its net assets from US$16.2 million to US$8.2 million. New shareholders do not share in it. They are buying the company the morning after the payout, then supplying fresh capital of their own on top.

The revenue that is not quite the company’s own

The single most important disclosure in the prospectus is not a number in the accounts. It is the sentence that says roughly 90% to 99.9% of the group’s revenue over the three years came from customers referred by All-Link PRC. In FY2023 and FY2024 the figure was above 99%. In FY2025 it was 90.4%.

This is important since All-Link PRC is not the customer that has an agreement and a purchase order. This is the family-run Chinese firm that provides the listed firm with the orders. The agreement between the two is now under the Non-Compete and Collaboration Deed of 30 June 2026 where All-Link PRC promises to give priority to the group when shipping to ASEAN countries. However, the risk factors of the group itself are rather blunt about the fragile nature of this protection. The deed is terminable and it automatically terminates if the firm is delisted or if Mrs. Tang Ying and Mr. Xu Hao lose control of the firm. Beyond this, as stated in the prospectus, All-Link PRC may reduce, redirect or stop the referrals “without liability to our Group”. In other words, the source of almost all the revenues lies with a firm that is run by the controlling shareholder’s husband, and continues to function due to the family retaining control and not due to any commercial obligation.

The dependency runs the other way too. To service that referred business, the group buys freight-forwarding and related services back from All-Link PRC, about US$20.5 million worth in FY2025, a sum equal to roughly 139% of the group’s latest net tangible assets, for origin-side handling in China where the listed company has no operating presence. So the referrer is also a major supplier, accounting for 31.5% of last year’s cost of sales. Money flows to the family entity on both sides of the trade.

The cost side concentrates too, and around the same name. The group’s largest supplier, at 40.5% of last year’s cost of sales and 79.7% the year before, is a company called All-Link Air and Sea Company Limited, which the prospectus says is not part of the group and is managed independently. The document does not define it, does not say what country it is registered in, and does not name its owners, though it does state that no director or substantial shareholder holds an interest in any major supplier apart from Mr. Xu Hao’s stake in All-Link PRC. Between them, two separately managed businesses carrying the All-Link name supplied 72% of last year’s costs.

Who this company answers to

Ownership answers part of the question the offer poses. After listing, Mdm. Tang Ying holds 51.7% and AGX Singapore 23.2%, so roughly three-quarters of the company stays with the insiders, and the 25.1% float, most of it placed rather than offered to the public, carries little voting weight. A minority buying in here is a genuine minority.

The more unusual feature is who sits behind AGX Singapore. It is wholly owned by AGX Group Berhad, which is itself a listed company, on the ACE Market of Bursa Malaysia. The prospectus is explicit about what that parent does:

“AGX Group Berhad is a company listed on the ACE Market of Bursa Malaysia and is principally in the business of providing sea and air freight forwarding, aerospace logistics, warehousing, road transport and distribution services globally. Save for aerospace logistics, the AGX Group operates in substantially similar business segments as our Group in overlapping geographical markets, including Singapore, Malaysia and the Philippines.”

In plain terms, one of All-Link’s controlling shareholders is also a listed competitor. That is the company’s own disclosure, not our characterisation. The alignment of the people running it points the same way. Mr. Peter Neo, the chief executive, owns no All-Link shares at all; his 18.90% economic stake is in AGX, the competitor he co-founded and whose board he left only on 1 January 2026. Mr. Chang Poh Sheng sits on All-Link’s board while serving as chief financial officer of AGX and holding 2.36% of it. The hands on the wheel still have meaningful ties next door.

That overlap is managed by agreement rather than by competition. On 30 June 2026 the two companies undertook not to solicit each other’s ten largest customers, and the All-Link names covered by that pact accounted for 100%, 100% and 94.7% of the group’s revenue across FY2023 to FY2025. Almost the entire customer book, in other words, is fenced off from the parent by a private undertaking, and like the referral deed it lasts only while AGX stays in control and the company stays listed. Half of the six-person board is independent, which is the regulatory minimum rather than a comfort, and the controlling shareholders’ lock-ups run just six months. What a buyer is being offered is a quarter of a company controlled by a listed rival, fed by a private family entity, with its customers allocated by contract rather than won in the open market.

The financial scoreboard

Look past the revenue line and the picture is not one of a business getting stronger. Gross margin fell from 40.4% in FY2023, when the company was tiny, to 14.7% in FY2024 and 12.0% in FY2025. Profit attributable to owners, the figure that matters to a shareholder, fell 25% in FY2025 even as revenue edged higher, from US$8.4 million to US$6.3 million. Administrative expenses nearly tripled to US$1.8 million as the company built out and prepared to list. And roughly US$0.8 million of the US$7.9 million pre-tax profit was simply interest earned on the cash pile, much of which is now being paid out as the pre-IPO dividend.

Cash flow tells the sharpest version of the story. Operating cash flow was positive US$29.9 million in FY2024, then turned to negative US$8.0 million in FY2025, as a large swing in trade payables unwound. For a business with almost no fixed assets, working capital is the whole game, and in its most recent year the working-capital tide went out.

The loophole that closed

To understand why the economics turned, you have to understand what the TikTok cargo was. It was low-value e-commerce parcels moving by air from China to the United States, the same flow that powered the rise of Shein, Temu and TikTok Shop. That flow existed on the scale it did because of the United States de minimis exemption, which let goods worth under US$800 enter the country free of duties and taxes. Cheap parcels, cheap entry, enormous volume.

On 29 August 2025, the United States removed the exemption. Every shipment now attracts duty regardless of value, which raises the landed cost of exactly the parcels All-Link was flying. The prospectus does not hide the consequence. It states that the removal “materially and adversely impacted” the group’s performance through a decline in volumes for the TikTok Group. The number tells the story: TikTok went from about 98% of revenue in FY2024 to about 45% in FY2025. The company is listing into the aftermath of the event that made it.

The pivot the story now depends on

Management is not blind to any of this, and the FY2025 numbers already show the beginning of a pivot. A new customer, described only as a US-listed multinational technology company with over US$2 billion of annual profit, arrived and contributed 33.6% of FY2025 revenue. The Malaysian and Philippine arms, which were rounding errors a year earlier, together reached about 9.6% of revenue. The stated plan is to push into Vietnam and Thailand, buy the Vietnamese affiliate, and diversify away from both TikTok and, over time, the All-Link PRC referral channel.

It is the obvious strategy. Whether it is a moat is a different question, and honesty requires stating where the strategy is weaker. Swapping a 98% dependence on one customer for a 34% dependence on another is real progress, but it is still concentration. The Malaysian growth was bought, not built, which is a faster route but a more expensive and less certain one. And the deepest dependency, the referral relationship with a family-controlled entity, is not something the diversification plan removes so much as leans on, since the new customers are still being won inside a network the family assembled. A forwarder’s genuine edge is density and relationships in specific trade lanes, and All-Link does have a credible one, handling an estimated 12% of Vietnam-origin air cargo to the United States in 2025. The trouble is that its largest lane by far was the China-to-US e-commerce lane, and that lane is precisely the one policy has turned against.

The cross-border parcel question

Every sector has a cautionary tale, and All-Link’s is not a single failed company but a whole category. The China-to-US low-value parcel boom created a generation of logistics and e-commerce businesses whose unit economics quietly assumed the de minimis exemption would last. When the exemption went, the volume went with it, and the businesses built on top of it discovered that a regulatory subsidy is not the same as a competitive advantage. All-Link’s own industry report, prepared by Frost & Sullivan, acknowledges the shift, noting a sharp contraction in low-value direct-to-consumer parcels, partly offset by a move toward consolidated bulk freight.

What is genuinely different in All-Link’s favour is that it is small, nimble, profitable and debt-free, and that it saw the shift early enough to start diversifying before listing rather than after. What is not different is the structural lesson: a business whose scale came from a policy window should be valued as if that window can close, because in this case it already has. The honest verdict is that the pivot is plausible and underway, but unproven, and the burden of proof sits with the next two years of results, not with the FY2024 revenue peak.

What the market is asking you to pay

Now that the price is fixed, the question stops being abstract. At S$0.53, All-Link is valued at about S$80 million. Against FY2025 profit attributable to owners of US$6.3 million, roughly S$8.1 million at the prospectus reference rate, that is a trailing price-to-earnings multiple of about 9.9 times. The post-offering earnings-per-share the company itself discloses, US 4.17 cents, lands on the same 9.9 times, so the figure is internally consistent. Measured instead against the FY2024 peak, when owner profit was US$8.4 million, the multiple is about 7.4 times. In other words, the market is paying roughly ten times the earnings of a year in which those earnings had already fallen a quarter.

There is a second lens that matters for an asset-light, cash-rich company, and it cuts the other way. After the pre-IPO dividend and the fresh money raised, All-Link should sit on something like US$29 million of net cash and near-zero debt, against a market value of about US$62 million. Strip the cash out and the operating business is being valued at roughly US$33 million, or a little over five times its FY2025 owner earnings. That is a much less demanding number, and it is the honest way to see what you are paying for the franchise itself rather than for the bank balance that comes attached. The tension between those two figures, ten times on the headline and about five times net of cash, is the whole valuation debate in miniature: the cash is real and large, but it is the earnings power, not the cash, whose durability is in question.

On dividends, there is little for a new buyer to price yet. The US$8.0 million already declared went to the existing owners before listing, so new shareholders do not receive it. Going forward the board has stated an intention, though not a binding policy, to pay out at least 30% of profit attributable to shareholders for FY2026 through FY2028. Applied to last year’s profit, that would have come to roughly 1.6 Singapore cents a share, or about 3% at the offer price. Whether it comes to that again depends on a profit line that fell 25% in the year just reported, and on cash reaching the listed holding company itself, since dividends must be paid out of its own distributable profits rather than the group’s.

How that compares with the neighbours

A multiple means little without a comparison of similar size, since small companies trade at structurally lower multiples than large ones. Nine freight forwarders listed in Malaysia, Thailand and Vietnam sit within a third to four times All-Link’s S$80 million. Their median trailing multiple is 9.9 times. All-Link is being offered at 9.9 times. Against the companies it most resembles, it is priced almost exactly in line.

One name in that set breaks the pattern, and it is the one that matters most. AGX Group Berhad, All-Link’s own controlling shareholder, running substantially the same business in the same markets, trades at about 15.8 times. The subsidiary is being sold to the public roughly 40% cheaper than the parent that controls it. That gap is not the sector’s verdict, because the sector median is 9.9 times. It is a verdict on this company in particular.

Singapore itself offers little to measure against. The exchange has no listed pure freight forwarder, and the three groups that do carry forwarding businesses, Vibrant Group, GKE Corporation and Chasen Holdings, trade on 5.4 to 7.7 times earnings. But each owns its warehouses and plant, so what they price is asset-heavy logistics rather than forwarding.

On assets there is no such ambiguity. Price-to-book measures what a buyer pays against the accounting value of what a company owns, and All-Link is the most expensive name in the comparison by a wide margin: about 2.8 times book, after the pre-IPO dividend and the new proceeds, against a forwarder median just under 1.0. No other company in the set, the parent included at 2.0, reaches two times.

The two readings meet in the middle. An asset-light forwarder owns almost nothing, so its book is thin and its return on that book looks extraordinary for as long as the work keeps arriving: about 29% on post-listing equity in FY2025, against a peer median near 8%. That ought to command a premium to book, and it does. What it is not getting is any premium on earnings, which is the market’s way of saying it doubts the 29% survives. If the return holds, three times book will look cheap. If it drifts back toward the sector’s 8%, the buyer holds the premium with no earnings discount to cushion it.

The bull, the bear, and the honest answer

The bull thesis is simple.

• It is an asset-light, profitable, and cash flow generative company that started from scratch and generated US$74 million of revenue in two years, has no debt, has management with extensive local logistics experience, is already diversifying its customer base and geographically, and is going public with new money to fuel an ASEAN expansion into some of the fastest-growing airfreight markets in the world.

• It is trading at five times net of cash last year’s profits, which is not an ambitious valuation for a logistics company with growth potential. If the strategy succeeds, the small surprise may just be a growing pain.

The bear case is equally clear, and sits mostly in the ownership and dependency structure rather than the trading.

• The revenue engine is a related party the minority cannot control.

• The single biggest customer’s volumes have already halved on a policy change with no obvious reversal.

• Profitability is falling, not rising. Margins are thin and getting thinner.

• The controllers extracted US$8 million in dividends immediately before listing, kept three-quarters of the company, and floated only a sliver to the public, and

• the whole edifice of referrals rests on the family remaining in control, which is exactly what an eventual sell-down would erode.

• On the one measure where the peer comparison is unambiguous, the shares are being floated at roughly three times what comparable forwarders fetch for their book value, with no discount on earnings to offset it.

The honest answer is that both cases are true at once, and that the deciding question is not really about logistics. It is about whether you believe a business assembled around one family’s China relationships and one customer’s tariff-advantaged parcels can become a diversified, self-standing ASEAN forwarder now that the advantage is gone. The prospectus gives you the evidence for the attempt and, at last, the price of the ticket. What it cannot give you is the result. At close to ten times a falling earnings line, the market is asking investors to pay for a transition that has only just begun.

Data integrity notes

Items materially relevant to readers, including estimates, definitions and known limitations.

1. All company figures come from the offer documents. Financial data is drawn from the prospectus dated 28 July 2026 and the accompanying Product Highlights Sheet, covering audited results for the financial years ended 31 December 2023, 2024 and 2025 [1][2].

2. Currency conversions use the prospectus rate. The group reports in United States dollars while the offer is priced in Singapore dollars. Conversions use the prospectus reference rate of US$1.00 to S$1.2881 as at the Latest Practicable Date [1].

3. The valuation measures are our own calculations. Market capitalisation, the trailing price-to-earnings and price-to-book multiples, the post-listing net cash estimate and the illustrative dividend are calculated from disclosed figures at the S$0.53 offer price and the 151,037,900 shares in issue after the offering. Book value is struck after the pre-IPO dividend and the net proceeds [1][2].

4. Two concentration measures should not be read as one. The Major Customers table gives the combined share of customers that each contributed 5% or more of revenue, at 99.7%, 98.0% and 84.9% for FY2023 to FY2025. The risk factors give top-five customer concentration, at 99.7%, 99.8% and 89.9%. The series diverge because the fourth and fifth largest customers each fell below the 5% threshold in the later years [1].

5. Peer multiples are indicative. They rest on traded prices as at 10 July 2026 and will have moved, while All-Link’s rest on an offer price and audited figures. The comparison is limited to freight forwarders worth between a third and four times All-Link, since multiples are size-sensitive, and excludes loss-making and distorted-earnings names. Widening the size band would raise the earnings median from 9.9 to 11.0 times; the narrower and less flattering figure is the one used [4].

6. The three Singapore names are context, not comparables. Vibrant Group, GKE Corporation and Chasen Holdings each run freight forwarding alongside property, materials or relocation businesses and own their asset base, so they sit outside the medians [4].

7. The industry figures are from the commissioned report. Air-cargo corridor shares and market commentary come from the Frost & Sullivan report reproduced in the prospectus, not from independent verification [3].

8. No third-party research was used. No sell-side or analyst research was an input. The article draws no forward earnings estimate, valuation target or recommendation.

References

[1] All-Link Air & Sea Limited, “Prospectus dated 28 July 2026,” lodged with and registered by the Monetary Authority of Singapore (audited FY2023-FY2025 financial statements, offer terms, capitalisation, use of proceeds, major customers and major suppliers, interested person transactions, moratorium undertakings, dividends, risk factors). Primary source for all company figures.

[2] All-Link Air & Sea Limited, “Appendix 4 Product Highlights Sheet dated 28 July 2026.” Both documents are available via the SGX-ST website and the MAS OPERA portal.

[3] Frost & Sullivan (Singapore) Pte Ltd, “Independent Industry Report,” reproduced as Appendix G in the All-Link Air & Sea Limited prospectus (ASEAN air-cargo growth rates, corridor tonnage and market share, cross-border e-commerce commentary).

[4] Peer valuation data for listed ASEAN freight forwarders, from exchange filings and market data as at 10 July 2026. Median set: MPJ Logistics (SET MPJ), TASCO (Bursa 5140), Triple i Logistics (SET III), Sonic Interfreight (SET SONIC), FM Global Logistics (Bursa 7210), South Logistics (HOSE STG), AGX Group (Bursa 0299), Sino Logistics (SET SINO), WICE Logistics (SET WICE). Core forwarders outside the size band: Tri-Mode System (Bursa 0199), Transimex (HOSE TMS). Singapore-listed logistics groups shown for local context: Vibrant Group (SGX BIP), GKE Corporation (SGX 595), Chasen Holdings (SGX 5NV).

[5] United States executive action removing the de minimis exemption for imports valued under US$800, effective 29 August 2025, as described in the All-Link prospectus risk factors and industry report.

Important Disclaimers

This article is published for informational and educational purposes only. It does not constitute financial advice, a recommendation, or a solicitation to buy, sell or hold any securities. The author is not a licensed financial adviser under the Financial Advisers Act 2001 of Singapore and this content is exempt under Regulation 34 of the Financial Advisers Regulations as a generally available publication. Consult a licensed adviser before investing. Past performance is not indicative of future results. The author holds no position in the securities discussed.

On the surface it looks bad.

• FY2025 revenue fell 9.2 percent to S$177.4 million,

• Net profit attributable to shareholders dropped 43 percent to S$5.8 million (S$6.1 million including non-controlling interests, the figure on the company’s highlights page [4]),

• Earnings per share went from 3.30 cents to 1.89, and

• Vietnam, until recently the second-largest market, collapsed 30.8 percent.

At S$0.35 [1] the market has filed Hyphens under “small-cap distributor in decline.”

Now the other line. The gross profit did not go down; rather, it increased to a new high of S$72.2 million from a new high margin of 40.7 percent, while the operating cash flow increased sharply to S$18.7 million, from practically nothing. Businesses in decline do not set new highs in gross profits and cash flow during the very year that revenues peak and decline. The problem is not whether FY2025 was bad. It was. The question is whether the bad year upgraded the earnings base.

Begin with understanding what Hyphens is – three businesses:

1. It in-licenses specialty drugs and aesthetics products and distributes them in more than five Southeast Asian countries;

2. It owns a growing set of brands like Ceradan and Ocean Health that carry a fatter margin; and

3. It operates docmedtech.com, a money-losing digital and wholesale business unit.

The point to remember here is that during the FY2025 period the company moved two products into the category of owned-brands and hence the reported figures showing 33% growth for that category are misleading along with an exaggerated 18% drop in the pharmaceuticals category.

Split the year in half, and the story becomes more compelling. Gross margin ended FY2025 at 42.0% in the second half of the year, versus around 39% in the first and 36% a year ago, the best in six years. The startling profit decline has been almost entirely contained within the first half of the year, when net profit plunged 66% compared to just 17% in the second half of the year. The first half decline is not attributable to a trading issue. While gross profit increased in the first half, this is because of problems further down the profit line due to Sterimar inventory write downs and foreign exchange losses, not loss of business.

The revenue bars are lumpy for reasons outside the thesis: the 2022 step up is the Novem acquisition consolidating for its first full year, and the 1H2024 peak was a post-disruption restocking bulge that then unwound. Gross profit, the steadier navy line, is the one that matters.

Let’s also be fair to the other side. The one thing that has mostly caused the decline in the revenue was Vietnam. This seems more like a regulatory problem caused by the obligatory reduction and not an optional one. According to management, the cause of the loss of the business in that country is due to them imposing stringent regulations on the pharmacies and the supply chain, which is evident through the new pharmacy laws in Vietnam and an inspection that will be carried out in 2025 [6].

The forward case rests on the owned brands. Hyphens in January 2026 did something that a pure distributor never does: it licensed its proprietary formula of Cerapro to a dermatology company from Switzerland for an up-front payment and royalties [2], demonstrating that owned technology has value. Medical aesthetics, via the Ardence acquisition, is the second pillar; though it remains small, it relies heavily on its founder’s presence. This is the best part of the portfolio, which explains why the gross margin can continue rising.

Now the number. Add back the genuine one-offs, chiefly a S$2.0 million Sterimar write-down plus smaller Vietnam and Ardence charges, about S$2.9 million before tax in total or roughly S$2.3 million after, and normalised earnings are about S$8.1 million against the S$5.8 million reported

On this basis, our FY2025 underlying earnings is down 15-20%, not 43%.

Given the currency headwind of S$2.8m needs to stay in the base as much as anything else (we’re assuming a currency effect annually whenever you do your buying in dollars and your selling in all sorts of other currencies around the region), a figure a bit above S$10.3m arises, which we will take to not be our base given that this is rather steep of a base to run on.

Yet, not all good business is a good stock. Hyphens, trading at 18.5x forward on S$0.35, sports P/E around the low teens to mid teens on our adjusted- earnings (and cash), while the market’s forward 8x P/E has it factoring a recovery of profits to S$13m-a level we think is too aggressive even on future predictions [2]. One free option sits on top: in 2022 a listed strategic, Metro Holdings, paid real cash for a slice of the digital-and-wholesale arm at a level the market now ignores [3], though the mark is stale and we credit it with nothing.

Three catches keep this honest.

To start, liquidity. Only 23.7 per cent of the stock, around S$25 million worth, is publicly held as two owners control the rest – more than two-thirds. This means no institution can add to a stake, or get rid of one, without affecting the market, and for long, so why shouldn’t the discount stay put and its causes perpetuate its presence.

Then, history. With S$10 million in net profit notched only twice in eight years as a listed company, the adjusted normalized base looks more like the midpoint of what it tends to deliver rather than a support floor.

Last, signals. Buying from CEO [5] deserves to be flagged, though against his existing control stake, it is trivial, and in a cash loaded, poorly performing public, it could mean either an inexpensive take private bid or re-rating. And don’t give the executive management’s promises much heed, as last year, for instance, it assured stakeholders the Vietnamese regulatory system would see no real shifts ahead - this was weeks before its collapse caused, it later insisted, by policy changes.

And what do you do with it? Not much, for a while. This is a watch-one-number scenario, and that one number is gross profit dollars. It is constructed to be a small position and watchlist candidate, not an investment; the float is just too thin to make a difference, the 4.3% dividend pays you to wait, and waiting may take a while.

The test is clean and close.

• The next results come out in mid-August, covering the first half of 2026. One number tells you most of what you need: is gross profit still growing, above the S$35.3 million the group made in the first half of 2025? But gross profit alone is not the answer. It only helps you if it reaches the bottom line, instead of being spent on the marketing needed to sell those brands. And that risk is real, not theoretical. In the second half of 2025 the owned-brands arm grew its sales but earned less, because growing it cost more than it brought in.

• So here is the good outcome: if gross profit keeps rising and this time reaches profit, then the market was ignoring a business that was quietly getting better, and that is exactly why you could buy it cheap.

• And the bad outcome: if the marketing eats the gain, then FY2025 just made the company smaller at a nicer margin, no richer, and the low share price was right all along.

Watch that line, all the way to the bottom one.

Data integrity notes

All figures are from Hyphens Pharma’s SGX filings, except share price and market data, which are from Yahoo Finance at the 23 July 2026 close [1]. Net profit is the attributable figure, S$5.8 million; the S$6.1 million on the company’s highlights page is the total before the 3.5 percent attributable to non-controlling interests [4]. Two caveats matter: the FY2025 segment reclassification means the 33 percent Proprietary Brands growth is not clean like-for-like, and the first-half figures we use are derived from full-year figures less the company’s reported second half. A basic governance and accounting red-flag screen did not show obvious issues, with goodwill tested without impairment, an unqualified audit, and operating cash flow above reported net profit.

References

[1] Hyphens Pharma International (1J5.SI), share price and market data via Yahoo Finance, at the 23 July 2026 close (S$0.35). https://finance.yahoo.com/quote/1J5.SI

[2] “Hyphens Pharma out-licences Cerapro MED Skin Barrier Cream for six European countries,” Hyphens Pharma media release, 5 January 2026. https://www.hyphensgroup.com/media-release-hyphens-pharma-out-licences-cerapro-med-skin-barrier-cream-for-six-european-countries/

[3] “Metro Holdings takes 10% stake in Hyphens Pharma subsidiary DocMed at $60 mil valuation,” The Edge Singapore, 27 May 2022. https://www.theedgesingapore.com/news/ma/metro-holdings-takes-10-stake-hyphens-pharma-subsidiary-docmed-60-mil-valuation

[4] Hyphens Pharma International, Financial Highlights. https://www.hyphensgroup.com/investor-relations/financial-highlights/

[5] “Directors increase stakes in Aspial Lifestyle, Raffles Medical Group, Centurion, among others,” The Business Times, 24 May 2026. https://www.businesstimes.com.sg/companies-markets/directors-increase-stakes-aspial-lifestyle-raffles-medical-group-centurion-among-others

Primary filings (SGX disclosures, no inline citation): Hyphens Pharma International Limited, FY2025 full-year results (24 February 2026); FY2020 to FY2024 results and half-year results; Annual Report 2025; and Responses to Substantial and Relevant Questions from Shareholders for the 2026 AGM.

Disclaimer and holdings

This article is published for informational and educational purposes only. It does not constitute financial advice, a recommendation, or a solicitation to buy, sell or hold any securities. The author is not a licensed financial adviser under the Financial Advisers Act 2001 of Singapore and this content is exempt under Regulation 34 of the Financial Advisers Regulations as a generally available publication. Consult a licensed adviser before investing. Past performance is not indicative of future results. The author holds no position in the securities discussed.

The mirage

In January 2026, Jollibee Foods Corporation said it would separate its international business from its Philippine operations and list the overseas arm on a United States exchange, handing existing shareholders stock in the new company. It was something no Philippine consumer company had attempted at this scale, and local brokers greeted it as a way to unlock the value of a global restaurant group the Manila market had never fully credited. Investors liked it at once, and the shares rose around 16 percent [2]. Two-thirds of Jollibee's more than 10,000 stores sit outside the Philippines, and the pitch was simple: put that fast-growing global business in front of American investors who pay up for restaurant growth, keep the steady Philippine cash machine at home, and the sum of the parts is worth more than the blended company trading in Manila. The logic was easy to follow, and for a few days the market bought all of it.

Six months later the enthusiasm had cooled. At its June annual meeting the company no longer described a spin-off and a 2027 US listing but an “ongoing strategic review” with “multiple pathways to unlock value over time,” one that would proceed “regardless of the outcome,” and the venue had drifted from New York toward Hong Kong [3]. The company’s annual-meeting messaging placed the separation inside a broader strategic review rather than repeating January’s firm transaction language. The shares gave back the excitement and now trade in the 140s, about a third below their level a year ago.

Most of the debate since has been about whether the spin happens. That is the wrong question. The more important one, which almost nobody asks, is whether it would deliver the thing it is premised on: a Wall Street re-rating of the international arm. And there is evidence to test it, because the market has already run several versions of the experiment.

The spin-off can crystallise value, but the listing alone cannot durably multiply it. The reason is simple: the closest available US-listed precedents, each Chinese or Chinese-origin, have not sustained US growth multiples. Chagee IPO’d on Nasdaq in 2025 at roughly 2.5 to 3 times sales on an enterprise basis and, as its growth cooled, fell to about 0.7; Yum China, spun off and NYSE-listed, sits near 1.3.

That matters because the bull case depends on a venue re-rating. Jollibee’s international arm may be valuable, but its value is already explainable on Asian marks. The mirage is not the business; it is the idea that New York changes the multiple. What follows builds that case from the filings up: what Jollibee is, where its profit sits, what it has bought, and then the test itself.

What Jollibee really is

Built from a single ice cream parlour the Tan family opened in 1975, Jollibee is now one of Asia’s largest restaurant operators, with more than 10,400 stores across 33 countries and around 20 brands. When customers at that ice cream parlour kept asking for hot meals, the founders switched the menu to burgers and fried chicken and incorporated the business as Jollibee in 1978. The defining moment came in 1981, when McDonald’s entered the Philippines and, against every expectation, failed to dislodge the local upstart: Jollibee had tuned its food to a Filipino palate, sweeter burgers and saltier fried chicken, that global chains never quite matched. To this day the Philippines is one of the very few markets on earth where the home-grown chain outsells McDonald’s on its own ground, and Brand Finance ranks Jollibee the most valuable restaurant brand in Southeast Asia [4]. That victory taught the company two lessons that still shape it: that local taste and habit beat global scale, which is why the domestic business is so hard to dislodge; and that the formula does not automatically travel, because the home-turf advantage is exactly what it lacks everywhere else.

Underneath the brands sit three businesses. The first is operating restaurants, some company-owned, most franchised, and the distinction drives the economics. A franchised store sends Jollibee a royalty with almost no incremental cost or capital; a company-owned store carries the full weight of rent, staff and fit-out. Jollibee’s mature domestic brands lean heavily franchised, which is why the home business throws off cash, while several overseas businesses are more company-operated, which is one reason they consume it. The flagship makes the split visible: of the 1,279 Jollibee-branded stores in the Philippines at the end of 2024, about two-thirds were franchised; of the 480 abroad, three-quarters were company-owned. Same brand, inverted model.

The second business is the commissary and supply chain, a network of central kitchens and distribution centres that manufactures the standardised sauces, marinades and frozen components which make a Chickenjoy in Davao taste like one in Manila. It is unglamorous, and it is a real moat: it enforces consistency, captures manufacturing margin a pure franchisor would hand to suppliers, and cannot be cheaply copied. A competitor can clone a recipe; it cannot clone a nationwide commissary built over four decades.

The third business, increasingly central, is capital allocation, which is to say buying restaurant companies. The domestic portfolio was assembled partly by acquisition; the international arm almost wholly so. This makes Jollibee as much an acquirer and operator of brands as a single chain, and it is the fact that most shapes how the overseas business should be valued. Management has spoken for years of becoming one of the largest restaurant companies in the world, so the spin-off is the logical endpoint of a decades-long ambition rather than a sudden idea, which is part of why the company is reluctant to abandon it even as the mechanics turn out to be hard. Two familiar Singapore touchpoints, for orientation: Jollibee owns The Coffee Bean & Tea Leaf, controls Tim Ho Wan through an investment fund, and through a joint venture took Tiong Bahru Bakery and Common Man Coffee Roasters to the Philippines [5].

Around the flagship sits a portfolio built to cover the whole Filipino table: Chowking for Chinese-style fast food, Mang Inasal for grilled-chicken value, Greenwich for pizza, Red Ribbon for cakes and bread, together roughly 3,500 outlets. That breadth captures the eating-out peso across occasions and price points and gives the commissary the volume that makes it efficient. Stack the layers and the domestic moat comes into focus: a brand that is a cultural default, a portfolio that spans the table, a franchised structure that turns dominance into high-return royalties, and a supply chain that enforces quality while capturing manufacturing margin. It is the combination, not any single layer, that a competitor finds hard to match, and it still has room to run as the group pushes into higher-growth provincial markets behind new commissary capacity in the Visayas.

The pattern to carry forward is that all of this is local. Brand affection, portfolio breadth, franchised density and commissary scale are built on Filipino habit and Philippine geography, and those things do not travel. Everywhere except Vietnam, Jollibee competes abroad without the home-turf advantage, as one more challenger buying its way into crowded markets. That is the deeper reason the profit sits where it does.

The profit is still Philippine

Almost all of Jollibee’s profit is Philippine, a fact the “clean unlock” framing skates over.

The six-year record contains a violent shock and an incomplete recovery. Revenue fell to about ₱129bn in the 2020 pandemic year, when dining rooms shut and the group posted its first-ever annual loss, roughly ₱11.5bn. It then recovered hard, to ₱154bn, ₱212bn, ₱244bn, ₱270bn and ₱305bn in the five years to 2025, more than doubling the top line. The bottom line came back more slowly and has since flattened: net income to shareholders climbed from that loss to about ₱10.9bn, but the step from ₱10.3bn in 2024 to ₱10.9bn in 2025 was small, and momentum turned negative in early 2026. Reported profit is also flattered slightly by deferred tax assets recognised on the overseas units’ accumulated losses, an entry that lifts net income without adding a peso of cash, and one more reason to read the segment detail rather than the headline.

On the company’s own FY2025 geographic figures, the Philippine business earned about 112 percent of group net income to shareholders; the international arm, taken together, was a small net loss, about minus 12 percent. The overseas arm made an operating profit, roughly 19 percent of group operating income, but a net loss once financing and acquisition costs were counted, so the Philippine business had to out-earn the whole group to offset it. The pattern held the year before, at plus 111 and minus 11 percent, and it worsened at the net line in early 2026: in the first quarter the international operations lost roughly ₱0.7bn, about minus 52 percent of a depressed group net income, and attributable net income fell about 39 percent, to roughly ₱1.5bn [8].

The home engine is strong but maturing. Philippine same-store sales growth has stepped down from about 7.5 percent in 2024 to 5.2 percent in 2025 to 3.2 percent in the first quarter of 2026, the last against a base lifted by election-related spending, and domestic margins felt cost pressure early in 2026. This is a dominant, franchised, cash-generative business that is also slowing, not an annuity that compounds untouched, which is part of why management keeps reaching abroad.

The point this sharpens is what the separation actually is. It is not two profitable halves going their separate ways. It is a profitable Philippine business that has quietly subsidised a global build-out, and a proposed separation that would hand at least part of that build-out to public investors abroad. The question the rest of this piece works through is what the shareholder left holding the Philippine business receives in exchange, and what it costs to deliver.

The international arm is real, but mixed

This does not make the international arm bad. It makes it early, and real. International EBITDA grew about 20 percent in FY2025 to ₱14.1bn, its operating margin rose from 2.0 to 3.0 percent, and its share of group operating income climbed from 13 to 19 percent. The engine is coffee: of that ₱14.1bn, ₱8.6bn, more than 60 percent, is Coffee and Tea, led by Compose in Korea (EBITDA up 156 percent in its first full year inside the group), The Coffee Bean & Tea Leaf, and Highlands in Vietnam. Around it sit the newer Asian brands and a scatter of turnarounds still dragging, China (international EBITDA down 69 percent) and Smashburger (EBITDA of about minus ₱1.0bn) among them. Jollibee’s international business is, first and foremost, a coffee platform with a Vietnamese quick-service winner attached, not a scaled-down replica of the chicken chain at home.

One caveat on the brand itself, precisely because it is easy to overstate. Where the Jollibee brand has travelled abroad, it has historically done so on the back of Filipinos rather than by winning locals: its overseas stores cluster where the diaspora is thick, and the company’s own push to reach mainstream American customers concedes as much [9]. Vietnam is the telling exception, with only a small Filipino population but more than two hundred stores selling to local diners. Kept in proportion, this applies to the Jollibee brand abroad, a few hundred of the arm’s thousands of stores; the arm’s growth case rests on the acquired local brands, Compose, Highlands and The Coffee Bean & Tea Leaf, where Filipino demand is beside the point. The value is not the brand’s diaspora pull but the businesses it has bought.

Because the arm is bought rather than built, judging it means judging Jollibee as an acquirer, and the record is genuinely mixed. The domestic template worked: Mang Inasal, bought in 2010, became one of the group’s best growth engines, a heavily franchised grilled-chicken chain that still posts some of the fastest same-store growth in the portfolio. That is the template the bulls hope repeats abroad, buy a strong local concept and scale it through franchising, though the domestic deals had the home-turf advantages working for them, which is exactly why they are weak evidence for what happens overseas. Abroad the results are harder to read. Smashburger, the American “better burger” chain taken to full ownership by 2018, has underperformed for years and was still being reformatted in 2026; The Coffee Bean & Tea Leaf, bought in 2019, has been a turnaround more than a growth story; China has been a repeated cycle of openings and closures. Against those sit the clear winners, and they cluster in Asia and in the newest deals. Vietnam is the standout, with systemwide sales up more than 40 percent last year; Tim Ho Wan more than doubled its Hong Kong store count within a year of full integration; and above all Compose, bought in 2024 for about US$340m [6] at roughly eight times its annual operating earnings, grew its underlying product sales more than 10 percent in 2025 [15], even as its reported contribution to the group jumped far more on its first full year of consolidation. Tellingly, Highlands is pursuing a parallel Vietnam listing at a reported US$400m [7], a separate monetisation path that could complicate, or potentially thin, the eventual international perimeter, depending on how the two transactions are structured. The honest reading is that the arm is some winners and some turnarounds wearing a single growth label, and the separation is partly an attempt to let the market price the winners without the drag.

The exits tell their own story. Over the past decade the group has quietly walked away from a US subsidiary, China’s Dunkin’ and a hotpot venture, and Vietnam’s Pho24, none large alone but together the mark of a company that buys often and is willing to cut its losers. That is healthier than clinging to them, but it is not the effortless compounding the “global growth company” label implies. Compose is the counter-example the bulls lean on hardest: almost entirely franchised, close to debt-free and high-margin, it is the asset-light model the group wants to define its international future. And the buying has not stopped, which is a tell in itself. In February 2026, well into the “strategic review,” the group agreed to acquire the South Korean hot pot chain Shabu All Day for about US$87m [6], a second Korean deal in under two years that sits awkwardly against the idea that the arm is being frozen and readied for a clean carve-out.

Management knows this ceiling, and its answer is instructive. It has been expanding its US franchising programme, aiming for a majority-franchised American footprint and courting mainstream customers rather than only Filipinos [9]. That is a genuine point in the separation’s favour. The open question is whether the strong US unit economics the company advertises hold beyond a handful of high-traffic destination outlets in areas of dense Filipino settlement; that is the diaspora question restated, and we flag it as our reading rather than a proven claim, since we lack the store-level customer data to settle it either way.

All of this sits on the balance sheet as about ₱78.8bn of goodwill and brand intangibles, more than a quarter of the group’s total assets, the accumulated price of a decade of buying. It is not a cost that recurs, but it is capital that has to earn a return, and at the group’s net line the international business has not yet earned one on it. Judging Jollibee therefore means judging its record as a buyer, which is why the mixed record above weighs as much as the growth rate.

What the closest US-listed Asian precedents show

Here is the mechanism the bull case needs, stated plainly. Today the international arm is buried inside a Manila-listed chicken chain and the market gives it little; the whole company trades at roughly what the Philippine business alone is worth. Separate it, list it in New York where investors pay rich multiples for restaurant growth, and the same business re-rates to a US growth multiple. The gap between “nothing” and “US growth multiple” is the unlock, and the venue-driven bull narrative maps that supposed re-rating onto the gap between the current price and bullish valuation scenarios, from the consensus mean near ₱211 [1] to the ₱287 top of the current range and the ₱330-plus targets seen before the recent de-rating.

Everything hinges on one link: a US listing produces a US multiple. It is the single testable claim on which the whole bull case rests, and it has been tested, because several Asian restaurant and coffee companies have already reached a US exchange by different routes, and we can see what the market paid. The closest to Jollibee’s proposed transaction is Yum China, which Yum Brands spun off and listed on the New York Stock Exchange in 2016: a carved-out, Asia-focused restaurant operator, made independent and primarily US-listed. It is large, profitable, dividend-paying and cleanly run, and it trades at about 1.3 times sales and nine times EBITDA. A fresher test arrived in April 2025, when Chagee, a Chinese premium-tea chain, listed directly on Nasdaq in a primary IPO; it trades near 0.7 times sales. And Haidilao carved out its own international arm as Super Hi and added a Nasdaq line in 2024; it sits near one times sales. Three routes onto a US exchange, a spin-off, a primary IPO and a carve-out, and the same answer each time:

Read down the Asian names and the pattern is hard to miss: across these three precedents, the current valuations sit at roughly 0.7 to 1.3 times sales, while the American growth names command three to eight. Chagee is the sharpest test, and it cuts both ways. When it IPO’d on Nasdaq in April 2025 the market did hand it a substantial US-style listing premium, valuing it at roughly 2.4 times sales on an enterprise basis at its IPO price and nearer 3 times at the opening pop, depending on the treatment of its IPO cash; as its growth cooled, that premium drained away and it now trades near 0.7 times EV/sales, like the rest. So the honest lesson is not that a US listing can never confer a substantial listing premium, it plainly can at the moment of listing, but that the venue cannot sustain one the underlying economics do not support.

Each Asian name has its own story, Chagee’s cooling growth, Yum China’s China-market risk, Super Hi’s dual listing, and that is exactly why the pattern persuades: the explanations differ, but the durable multiple is missing in every case. We are not claiming the venue is worth nothing; a genuine US listing can add liquidity, disclosure and investor access, and the research on cross-listing premiums says those are worth something. We are claiming something narrower and sturdier: no plausible venue benefit durably turns a one-times-sales Asian platform into a five-times-sales American one. Absent materially better growth, margins and capital efficiency than it shows today, Jollibee’s coffee-led Asian arm belongs nearer one to two times sales than to the multiples of US growth chains. The re-rating the bull case needs is the part that does not last.

There is a real limitation to this comparison, and it is worth stating plainly. Each of the three closest US-listed precedents used here is Chinese or Chinese-origin: Yum China operates almost entirely in China, Chagee is a Shanghai-based chain, and Super Hi is the carved-out international arm of a Chinese hotpot operator. Part of their discount may therefore be China-specific, reflecting US-China delisting risk, governance perceptions around Chinese issuers, capital controls and single-market concentration, rather than a generic Asian penalty. Jollibee’s arm is Philippine-controlled and spread across Korea, Vietnam, China and North America, and could on that basis command a somewhat higher multiple than a pure-China name. So our claim is not that these precedents prove an immutable Asian ceiling. It is narrower and harder to dispute: they offer no support for assuming an automatic leap to US growth-chain valuations, and any durable premium Jollibee’s arm earns will have to come from its own growth, margins and capital efficiency, not from the ticker. On that score its own record, loss-making at the net line, minority-laden and built by acquisition, argues for caution rather than a premium.

So a US listing crystallises value without multiplying it. That leaves the two questions that decide whether the stock is actually cheap: what are Jollibee’s parts really worth, and how much of that ever reaches the person holding the shares in Manila?

Below, for paid subscribers: the sum-of-the-parts, and how far the per-share value drops once the arm carries its real share of JWPL's debt; the leaks between enterprise value and a Manila shareholder's pocket; the bull case at full strength; and the evidence that would prove us wrong.

Read more

The most expensive silicon in the world does not ship when it leaves the fab. It ships when someone proves it works.

For AI accelerators, that proof is getting harder, longer and more valuable [2]. A modern accelerator has to be probed, packaged, burned in, and increasingly run through a full system-level rehearsal before it goes anywhere near a data centre. The result is that test is moving from a back-end afterthought to a front-line bottleneck, and as customers diversify their back-end supply chains away from Greater China, more of that bottleneck is being relieved in Southeast Asia.

That sets up a simple investment question: who actually captures the AI-test dollar, the independent tester, the turnkey OSAT, the equipment supplier, or the customer that owns the tools? The answer is not “every OSAT.” Pure-play testers, integrated packaging houses and equipment suppliers sit in different layers of the stack, run on different qualification clocks, and trade at very different valuations. Some independent testers still trade like cyclical back-end shops; some AI-adjacent names already trade at equipment-like multiples.

This is not a screen of Southeast Asia-listed companies only. It is a map of who captures the test economics of Southeast Asian back-end migration: local listed names across independent test, OSAT and equipment (KESM, Inari, Unisem, MPI, AEM), foreign names building regional capacity (KYEC, ASE, Chipbond, Trio-Tech), and the equipment leaders and specialists whose tools set the margin pool (Advantest, Teradyne, Aehr). We map who does what, who catches the first wave, and, for paying subscribers, where the valuation gap looks most interesting.

What semiconductor test is, in plain English

Think of the back end as four gates. Probe the die while it is still on the wafer, package it, test the finished package, then stress it under real-world conditions. AI makes those gates more valuable because the parts passing through them are worth thousands of dollars each, and because a modern accelerator is not one chip but a stack of chiplets and high-bandwidth memory (HBM) bonded onto an interposer, the small slab that wires them together. Once those pieces are bonded, rework is essentially impossible, so every die has to be proven good first. Some industry estimates put test at roughly 5% to 10% of the cost of a leading-edge AI part, up from low-single digits for a conventional chip [6][7]. Test has moved from the end of the line toward the front of it.

Two distinctions run through the whole piece. First, whoever owns the tester matters: the core testers, the automated test equipment (ATE), are dominated by Japan's Advantest and America's Teradyne, on the order of 80% of the high-end tester market between them, though share estimates vary by how the market is defined [1][3], and a test house either owns its machines or runs them on consignment for a customer, which changes the margin. Second, an independent test house (also called a merchant tester) sells test as a service and is not owned by or affiliated with a chip customer, while an integrated OSAT folds it into packaging. Those two compete for the same AI work by different routes.

It helps to see the whole thing as a stack, because the Southeast Asian listed opportunity is not spread evenly across it. It is thin in the tester layer, which is a foreign duopoly, and thickest in the test-service layer and in the handler and system-level-test niche where Singapore's AEM sits.

One honest caveat frames everything. The most leading-edge packaging, the CoWoS-class interposers (TSMC’s advanced chip-on-wafer-on-substrate method) and the HBM stacking closest to the AI die, largely stays captive inside the foundries and IDMs (TSMC, Intel, Samsung, SK hynix), because it is too tightly coupled to the front end to hand off [11][35]. What actually outsources to independent houses in Southeast Asia is chiefly final test, burn-in and mid-tier packaging. Even at the leading edge, testing is the minority of the dollar: ASE puts its own advanced packaging-and-test revenue at roughly 75% packaging and 25% testing [24]. That is why test, not packaging, is the right lens here.

The money is following the difficulty. SEMI expects semiconductor test-equipment sales to rise about 48% to roughly US$11.2 billion in 2025 [8]; the broader outsourced assembly-and-test market is put near US$47 billion now, heading toward US$71 billion by 2030 [9]; and the slice that matters most, burn-in and system-level test for accelerators, is smaller but growing faster still [10]. That is the demand pull under everything that follows.

Subscribe now

Why Southeast Asia, why now

The pull is policy as much as economics. US export controls tightened through 2023 and 2024, and the volatility that followed through 2025 and 2026 kept pushing customers to qualify test sites outside China [14]. The leverage in that is arithmetic. Greater China ran about 75% of global OSAT in 2023, so the pool of back-end work that policy could dislodge is enormous, while Malaysia handles an estimated 13% of global chip assembly and test, the largest established cluster outside China and Taiwan [9]. When the big pool is what moves and the smaller base is where it lands, even a modest share shifting south is an outsized gain for Southeast Asia. That asymmetry, not the region's size today, is why the timing matters.

Why not just keep test next to the fab? Because it does not have to be there. A finished wafer is small and hugely valuable per gram, so shipping it to a cheaper site costs little, whereas test and burn-in are labour-, power- and capacity-hungry and lower-margin, work that gravitates away from high-cost, resource-constrained Taiwan. The packaging that must stay coupled to the fab does stay put, the CoWoS interposer and the HBM stacking remain captive at the foundries and memory makers. Korea’s memory giants make the point in concrete: rather than outsource, SK hynix and Samsung are pouring billions into captive HBM packaging-and-test at home, SK hynix’s roughly US$13 billion Cheongju plant and Samsung’s Cheonan and Onyang lines [35]. What travels is the separable test and mid-tier packaging, and it travels to wherever is cheapest and most neutral.

The migration is not abstract. KYEC, Taiwan’s largest independent tester, divested its entire China testing operation during 2025 even as it built a new plant in Singapore [12]. Test-equipment maker Teradyne pulled roughly US$1 billion of production out of Suzhou under the same pressure [13]. Trio-Tech shows the other side of the same move: its China revenue fell about 84% over the nine months to March 2026 while group revenue rose 85%, consistent with AI final-test work leaving China for its Singapore and Malaysia labs [39]. The rules themselves keep shifting, though. A January 2025 rule that would have swept Malaysia into a tiered AI-chip export regime was rescinded that May, and Washington has drafted and pulled back further AI-chip export measures since, most recently a broad rule withdrawn in March 2026 [14]. No single rule has stuck. The signal is the pattern, not the particulars: the regulatory line can be redrawn at short notice, and that uncertainty pushes customers to spread their back-end work across several countries rather than bet on one, which is exactly what favours the region.

Southeast Asia is where that capacity is landing. The table below is the map at a glance.

The countries are not competing on the same terms; each pulls a different policy lever at a different point on the chain. Malaysia leans on the deepest existing ecosystem and its National Semiconductor Strategy, which had drawn RM85 billion of approved investment by the end of 2025 against a 60,000-worker target, and is nudging local firms up-market through programmes like SemiconStart [36]. Singapore competes on value and neutrality rather than cost, winning the equipment and high-end test slice, AEM’s tooling and KYEC’s neutral Singapore plant [4][21]. Vietnam is the aggressive challenger, with a national target of ten advanced test-and-packaging plants by 2030 and incentives that have pulled in Samsung, Intel and Amkor [28][31]. Thailand’s Board of Investment has drafted a national roadmap targeting more than 2.5 trillion baht of investment and a “chip made in Thailand” by 2050, skewed toward power electronics and automotive [37]. The Philippines offers CREATE-era incentives over a large but mostly foreign-owned assembly-and-test base [31], and Indonesia is earliest-stage, a government test house and a joint-venture push from a low, import-dependent base [28][31]. For an equity investor the map compresses further: Malaysia is where the listed cluster actually sits (KESM, Inari, Unisem, MPI), Singapore adds AEM and KYEC’s site extension, and Vietnam, the Philippines and Indonesia matter strategically but offer little directly listed exposure because their new capacity is largely foreign or captive.

Penang is the anchor, and it straddles the island and the mainland. On the island, at Bayan Lepas, sit KESM’s independent burn-in lines, and Cohu, a US test-equipment and handler company adjacent to Singapore’s AEM, opened a test-design centre there in June 2024, deepening the local engineering base [16]. Across the water on the mainland, in Seberang Perai and specifically Batu Kawan, the heavy new capacity is going up: SPIL’s roughly RM6 billion (about US$1.27 billion) packaging-and-test plant at Bandar Cassia, the captive TF-AMD joint venture that runs AMD’s Penang assembly and test, Inari’s P34 line serving a new US memory customer, Micron’s assembly-and-test operations, and now Chipbond’s new plant, with Intel investing on the order of US$7 billion in captive advanced packaging in Penang alongside them [9][17][18][22]. Smaller but riding the same thesis, Trio-Tech runs a new leased final-test facility in Perai, where the single AI customer behind its recent revenue surge is served [39]. The split matters: the island holds the legacy base, while the mainland at Batu Kawan is where most of the new capacity is rising, a strait apart rather than one campus. Not all of it is AI, ASE's newer Bayan Lepas plant, P5, is wire-bond for automotive and edge, not accelerators [19], and the AI share is selective. But taken together, it is one of the densest back-end clusters outside Taiwan.

Two business models

The capacity race is really a contest between two models.

The turnkey OSAT channel bundles packaging and test on one floor and captures the test work once its own line is online. SPIL, an ASE subsidiary, is the local example: it broke ground at Batu Kawan in May 2024 on a plant meant to offer turnkey packaging-and-test under one roof [17].

The independent-test channel is the standalone tester extending qualified capacity into a new geography. KYEC is the textbook case: test is roughly 95% of its revenue, and it opened a S$100 million plant in Singapore in 2026 [12][21]. It is no longer alone; Chipbond, a Taiwanese driver-IC bumping-and-test specialist, opened a plant of close to US$200 million at Batu Kawan in early 2026 [22].

The two do not move at the same speed, and the reason is qualification. Putting a new test supplier into a customer’s flow is a months-long proof of yield and reliability, framed by KESM’s own management as a six-to-twenty-four-month exercise [23]. An independent tester opening a site for customers it already qualifies elsewhere runs a site extension, not a cold qualification, so it moves first. A turnkey OSAT only wins the test work once its own floor is built and proven. That head start is, in practice, the whole game.

The players, by role

One filter runs through every name below: how close is the AI link, really? It helps to keep three tiers apart. Direct AI test is burn-in, final test and system-level test for accelerators and HPC chips. AI-infrastructure adjacent is the optical transceivers, power, analog, sockets and inspection that surround those chips without testing them. General back-end is the automotive, RF, industrial, consumer and EMS work that dominates most of these companies’ revenue today. Very few of these names are in the first tier; the market too often prices them as if they were.

Independent testers (the purest bet on the theme). KESM Industries (Bursa: 9334) is the regional anchor, a listed independent burn-in and reliability-test house that describes itself as the world’s largest independent provider in that niche, defining independent as not related to any of its customers, with an automotive base, lines at Bayan Lepas on Penang island and a Malaysian footprint extending to Petaling Jaya, and for the first time it named “artificial intelligence related chips” as a demand driver in its 2026 filings while tripling capex [15]. KYEC (TWSE: 2449) is the Taiwanese pure-play now Singapore-relevant, serving 48% of the world’s fifty largest chipmakers with NVIDIA among named customers and building its own high-power burn-in ovens [12]. Trio-Tech International (NYSE American: TRT) is a small US-listed final-test and burn-in house with regional operations, and the clearest China-plus-one case in the group: its recent revenue surge is one customer's AI final-test work relocated out of China [39]. Chipbond (TWSE: 6147) is the newest entrant, the second Taiwanese tester to plant capacity in Penang [22].

Turnkey OSATs (packaging plus test). ASE Technology Holding (NYSE: ASX; TWSE: 3711), parent of SPIL, is the mega-cap shown for scale; its 2025 revenue of about NT$645 billion splits roughly 59% semiconductor packaging-and-test and 40% electronic manufacturing services, a reminder that even the group flagbearer is only part test [24]. Among listed Malaysian OSATs, Inari Amertron (Bursa: 0166) is the largest but is really an RF and optoelectronics specialist: radio-frequency work was about 61% of revenue in its most recent quarter, optoelectronics about a third. Inari does not disclose its largest customer’s share, but roughly 87% of FY2025 revenue was booked in Singapore, a widely used proxy for its heavy reliance on one principal customer, generally identified by analysts as Broadcom. Its AI angle is real but specific, 800-gigabit optical transceivers for data-centre interconnect rather than accelerator test [20]. Unisem (Bursa: 5005) is a broad assembly-and-test house leaning on a single large US analog customer that brokers identify as Monolithic Power Systems (its 2024 annual report discloses one customer above 10% of revenue), and it is raising up to RM742 million through a placement to expand [25][26]. Malaysian Pacific Industries (Bursa: 3867), through Carsem, is automotive-heavy and bought Infineon’s Bangkok/Nonthaburi back-end plant in 2026 with a long-term supply agreement attached, the second Infineon back-end site to pass to a Southeast Asian OSAT after ASE’s 2024 Cavite deal, which gives its Thai expansion built-in offtake [9][27][28].

Equipment and adjacent names. AEM Holdings (SGX: AWX) is the most direct equipment play, but it sits a layer below the testers: it builds the handlers, system-level-test cells and burn-in equipment that work with the tester. It has been Intel’s principal test-handling supplier for over a decade, with more than 35,000 system-level-test sites deployed, and its active thermal control, which keeps high-power AI die cool during test, sits directly in the accelerator’s path; a 2025 Intel Foundry partnership and a high-volume ramp for an AI/HPC customer lifted FY2025 revenue 5% to S$399 million, driven by its Test Cell Solutions segment. The flip side is concentration: one customer was about 38% of group revenue in 2025 [4]. ViTrox (Bursa: 0097) and Pentamaster (Bursa: 7160) are inspection and automated-test equipment makers; JF Technology (Bursa: 0146) makes the sockets and contactors [5]; Globetronics (Bursa: 7022) is a sensor-led OSAT now shrinking [29]. In Thailand, Hana Microelectronics (SET: HANA) is a hybrid EMS-and-OSAT: about 63% of 2025 sales was board-level PCBA (printed-circuit-board assembly) and 31% IC assembly-and-test, at an 8% gross margin. Tellingly, it cut capital spending sharply in 2025 as IC demand softened, a useful reminder that not every regional back-end name is riding the AI wave [30]. None of these is an independent test house, and treating them as one is the easiest mistake in this sector.

Put the roster against the numbers and the theme separates cleanly: the AI-exposed testers are growing, and the consumer- and RF-cyclical names are not.

So where does that leave the roster? Before valuation enters the picture, the names sort into three buckets by exposure:

The theme, in other words, is not “buy Southeast Asian back-end.” It is to find where the pricing and the durability disagree. That is what the rest of this piece is for.

The free map tells you who is exposed. The paid question is what the market is already paying for, where that looks wrong, and which exposures are durable enough to deserve the multiple. Below, for paying subscribers, we compare the names on valuation and EBITDA margin, weigh customer concentration and the hidden margin question of who owns the tools, and set out what we are watching into the 2027 capacity wall.

The moat, the clock and the 2027 risk

Read more