Revenue Accelerates to +48% and Shopee's FY26 Profit Goal Rises to $1B, But Tax and Monee's Spread Eat the Net Income Line

Results vs. Consensus

Revenue

Revenue of $7,787.8M grew 48.1%, a step up from Q1's 46.6%. Service revenue carried it, rising 48.6% to $7,130.1M, with sales of goods up 42.8% to $657.7M. The composition matters more than the headline. Shopee contributed $5,587.6M (up 48.2%), Monee $1,402.8M (up 58.9%) and Garena $746.6M (up 33.5%). Within Shopee, core marketplace revenue, which is transaction fees plus advertising, rose 65.6% to $4,255.5M, while value-added services revenue fell 9.0% to $676.4M because more logistics revenue is now netted against shipping subsidies.

Netting the two, total Shopee marketplace revenue as a share of GMV moved from roughly 11.1% to roughly 12.9%, an expansion of about 176 basis points. Management attributes just over 90 basis points of that to advertising. The remainder is commission increases and the value-added services netting change. That distinction is worth holding onto: roughly half of Shopee's take-rate expansion this quarter came from raising the price of the platform, not from selling sellers a better product.

Margins

Gross margin was 45.6%, down 24 basis points from 45.8%. That is a good outcome given the mix, because Shopee cost of revenue rose 51.9% on logistics investment while Monee, the highest-gross-margin business, grew fastest. Operating margin fell 92 basis points to 8.4%, and the driver is visible in one line: sales and marketing expense rose 64.5% to $1,660.1M, well ahead of revenue. Monee's own sales and marketing spend rose 139.8% to $293.9M as the company pushed harder on new borrower acquisition.

Discipline held everywhere else. General and administrative expense grew 22.0% and research and development grew just 5.6%, both far below revenue. Unallocated corporate expense actually fell 16.7% to $141.7M, helped by share-based compensation declining to $131.9M from $161.9M. This is a company spending into customer acquisition, not into overhead.

EPS and below-the-line

GAAP diluted EPS was $0.70 against $0.65, up 7.7%. Basic EPS was $0.72 against $0.68. The gap between 33.3% operating-income growth and 7.7% diluted EPS growth is almost entirely tax, with a small assist from lower non-operating income ($65.5M against $83.3M, as interest income fell). Share count worked in shareholders' favour: diluted weighted average shares for the half were 634.5 million against 636.2 million a year ago, helped by 4.7 million shares repurchased in the quarter for $416.8M at an average near $88.68.

The tax move is the open question of this print. A ten-point effective-rate step in a single year is either a change in the geographic mix of taxable profit, the exhaustion of a concession or loss carryforward, or a discrete item. The release discloses the number and nothing else. Until Sea explains it, the prudent assumption is that a meaningful part of it is structural, which is how we have treated it in the model section below.

Year-over-Year Comparison

The first half now reads: revenue $14,885.3M against $10,100.6M (up 47.4%), gross profit $6,695.5M against $4,646.0M (up 44.1%), operating income $1,243.3M against $944.1M (up 31.7%), and net income $896.3M against $825.0M (up 8.6%). The pattern is identical to the quarter. Everything above the tax line compounds in the thirties or forties. Everything below it grows single digits.

Sequential Comparison

Consolidated adjusted EBITDA fell 11.3% sequentially and dropped back below the $1B threshold that Q1 crossed for the first time. The cause is almost entirely Garena, where seasonality reverses hard: Q1 carries the Lunar New Year and anniversary content cycle, and Q1 2026 additionally carried a $234.9M swing in deferred revenue against just $16.9M this quarter. Shopee moved the other way, adding 14.4% sequentially and giving the FY26 goal its credibility. Read the two together and the sequential decline is a mix artefact, not a deterioration.

Segment Performance

Shopee: monetisation is doing the work, and the margin is inflecting

GMV reached $38.3B, up 28.4%, on 4.2 billion gross orders, up 27.5%. Management framed this as the eighth consecutive quarter of sequential growth. Revenue of $5,587.6M rose 48.2%, comfortably ahead of GMV, which is the take-rate expansion described above. Adjusted EBITDA of $255.4M grew 12.2% year on year and 14.4% sequentially off the Q1 low.

"GMV grew 28% year-on-year, marking eight consecutive quarters of sequential growth, and we again achieved new highs in gross order volume and revenue."
— Forrest Li, Chairman and CEO

The composition of Shopee's growth changed this quarter in a way worth flagging. In Q1, growth leaned on existing-buyer frequency, which rose 12% year on year, with monthly active buyers up 16%. This quarter frequency slowed to 8% while monthly active buyers rose 18% and, more strikingly, average monthly new active buyers rose more than 35%, which management called a significant acceleration from previous quarters. Shopee has switched from deepening its existing cohort to widening the funnel, and it is paying for that switch through the 64.5% increase in consolidated sales and marketing expense.

"Average monthly new active buyers grew more than 35% year-on-year, a significant acceleration from previous quarters."
— Forrest Li, Chairman and CEO

The advertising engine remains the cleanest part of the story. Ad revenue rose more than 70% and the ad take rate improved by more than 90 basis points, matching Q1's expansion. Underneath it, the number of ad-paying sellers rose around 45% while average ad spend per seller rose more than 15%. Compare that with a year ago, when seller count rose about 20% and spend per seller rose more than 40%. The engine has shifted from deepening a small advertiser base to broadening a large one, which is the more durable of the two patterns but also the lower-yield one per seller added.

Logistics and fulfilment. Instant and same-day delivery gained traction, with order volumes using instant delivery up around 80% year on year while cost per order fell around 20%. Fulfilment order volumes rose more than 20% sequentially, and in some markets more than 60% of fulfilled parcels arrive next day. Management repeatedly stressed the capital-light structure: Sea rents rather than owns fulfilment space and does not own trucks or land.

ShopeeVIP. Membership passed 15 million at end-June, up 25% sequentially, and is now live across Asia and Brazil. VIP members contributed 24% of GMV across Asia, up from roughly 20% in Q1. Monthly retention held around 80%. Brazil, launched in April, already passed 1 million members. The newer detail is that sellers and external partners have started co-funding VIP benefits, which management says improved the programme's unit economics in Asia. No numbers were attached to that claim.

Brazil. Still the fastest-scaling market, for a fifth consecutive quarter, growing GMV above the market on gains in active buyers, purchase frequency and basket size. Average buyer waiting time fell 15% year on year, fulfilment penetration doubled, nearly 500 new official brands were onboarded, and Shopee Mall GMV more than doubled. On the competitive question, management said pricing leadership held even after a major competitor cut both take rates and its free-shipping threshold.

Assessment: Shopee's quarter validates the operational thesis and marks the turn in Shopee's year-over-year earnings, which is what the guidance raise formalises. The reservation is the denominator. Adjusted EBITDA equals 0.67% of GMV this quarter against 0.76% a year ago, so on the metric management itself uses, Shopee is more profitable in dollars and less profitable per unit of volume than it was twelve months ago. The sequential improvement from Q1's 0.60% is real and is the right trend. But the recovery is measured from a depressed base, not from the prior peak, and that is a materially less impressive claim than the $1B headline implies on its own.

Monee: the balance sheet is compounding faster than the earnings

Revenue rose 58.9% to $1,402.8M and the loan book reached $11.1B, up 62.5%, split $10.0B on-book and $1.1B off-book through channelling arrangements. The 90-day non-performing ratio was 1.0%, which the company described as stable sequentially. Active credit users passed 40 million, up around 34%, with roughly 5.3 million unique first-time borrowers added in the quarter and average loans outstanding per user up around 20%. The Philippines became the newest market with a loan book above $1B.

Against all of that, adjusted EBITDA grew 12.8% to $288.0M. The segment EBITDA margin fell from 28.9% to 20.5%, a compression of roughly 840 basis points. Provision for credit losses at the segment level rose to $553.7M from $315.6M, moving from 35.8% of segment revenue to 39.5%. Sales and marketing rose 139.8%. On the balance sheet, the allowance for credit losses across current and non-current loans receivable reached $1,114.8M against $842.0M at year-end, taking coverage of gross loans to roughly 10.4% from 9.6%.

Q: "We noticed that the provisions for credit losses have increased quite a bit this quarter. What are the trends that you are seeing in delinquencies? And how does that affect your loan growth going forward?"
— Ranjan Sharma, JPMorgan

A: "One is the off-Shopee SPL and second one is the Brazil loan outstanding. Although Brazil, we have a very good ROA there, but it's a high interest, high risk market, so the higher mix of this two components contribute to the higher provision that you see."
— Tony Hou, Chief Financial Officer

The mix explanation is coherent and internally consistent. Off-Shopee SPayLater now exceeds 20% of the total SPayLater portfolio, reaching as high as 35% in some markets, and carries structurally lower return on assets than the on-Shopee product. Brazil is high-rate and high-loss. Management also confirmed it is deliberately reaching into prime segments that accept lower interest rates. Each of those is a rational trade of yield for volume and durability.

Assessment: The credit franchise is not deteriorating. NPLs are stable, coverage is rising rather than falling, and management's stated guardrail is stable non-performing rates within each product, country and segment. What is happening is a permanent repricing of the business toward a lower-spread, larger-balance model. That is probably the right long-term choice, but it breaks the mental model that carried Monee through 2025, when revenue and EBITDA compounded together above 50%. Investors underwriting Monee on a percentage of the loan book need to mark that percentage down. On disclosed figures, quarterly adjusted EBITDA against the period-end loan book fell from roughly 3.7% to roughly 2.6% year on year.

Garena: monetisation depth against a flat user base

Bookings were $763.5M, up 15.5%, decelerating from 20.1% in Q1 and well below the more than 30% pace guided for full-year 2025. GAAP revenue rose 33.5% to $746.6M, and adjusted EBITDA rose 16.7% to $429.8M, equal to 56.3% of bookings against 55.7% a year ago. Free Fire continues to draw more than 100 million average daily active users.

The user metrics deserve attention. Quarterly active users were 666.3 million against 664.8 million, essentially unchanged. Quarterly paying users rose 10.2% to 68.1 million, lifting the paying ratio from 9.3% to 10.2%, and average bookings per user rose from $0.99 to $1.15. Effectively all of Garena's growth this quarter came from converting and monetising the existing audience rather than from expanding it.

"Free Fire's longevity comes from a single discipline: we keep the experience fresh with the new game play and the content and we make it feel both local to the communities who play it and enjoyable for a global audience."
— Forrest Li, Chairman and CEO

The content cadence this quarter was in-house rather than licensed: an ocean-themed Songkran campaign called Undersea Mystery, and a World Cup tie-in whose campaign song generated over 350 million social media views. That is a change from the last four quarters, each of which was anchored by a major external intellectual property collaboration. Looking forward, Sea announced two licensed titles: Palworld Online, self-developed and globally published under licence from Pocketpair, and Monster Hunter Outlanders, developed by Tencent on Capcom's franchise and targeted at Southeast Asia, Latin America, Taiwan and potentially the Middle East, with a launch targeted this year.

Assessment: Garena is executing well on a harder comparison base and its profitability per booking is still improving. But the growth algorithm has narrowed to one variable. A flat active user base means the paying ratio and spend per user have to keep rising indefinitely, and both have finite headroom. The two announced titles are the right structural answer, and both rest on genuinely global intellectual property, which is a step up from the collaboration model. Neither has launched. Until one does, Garena is a single-title business monetising harder, and that is a materially different risk profile from the diversified franchise the segment has been marketed as.

Key Operating Metrics

Entries marked "n/d" are not disclosed at that grain. Take rates are computed on the disclosed GMV figures, which the company rounds to $0.1B, so they are approximate to within a few basis points.

Key Topics & Management Commentary

Overall management tone: Confident and unusually specific on operating detail, noticeably less specific on anything below operating income. Prepared remarks led with growth investment framed as already paying back, and for the first time in four quarters management volunteered a hard full-year profit number rather than a floor, which is the clearest signal of internal confidence in the print. The posture on questions about profit conversion was different: answers on Monee's spread and on the path to a 2% to 3% Shopee margin were directional and mechanism-based rather than quantified or time-bound, and two direct questions went unanswered entirely. Tone was more assured than Q1 on operations, and less forthcoming than Q1 on earnings quality.

1. The Shopee full-year profit goal moves from a floor to a target

The most consequential disclosure of the quarter is a single sentence in the prepared remarks. Last quarter Sea committed only that FY26 Shopee adjusted EBITDA would be no lower than FY25's $880.6M. This quarter it named $1.0B.

"I'm pleased that Shopee has delivered a strong first half of 2026. With this solid momentum we are optimistic that Shopee will achieve the milestone of $1 billion in adjusted EBITDA for the full year."
— Forrest Li, Chairman and CEO

First-half Shopee adjusted EBITDA was $478.6M ($223.2M in Q1 plus $255.4M in Q2). Reaching $1.0B therefore requires roughly $521M in the second half, about 9% above the first half. Management confirmed the shape explicitly when pressed, and noted that last year the second half was lower than the first.

Assessment: The raise is modest against the Street's $980.7M but large against Sea's own prior floor, and the required second-half step-up is undemanding given seasonally higher fourth-quarter GMV. This is a guide management should clear. The more interesting question is why they chose to name a number now rather than at Q3, and the most likely answer is that they wanted the market to stop treating 2026 as an open-ended investment year.

2. The switch from frequency to new-buyer acquisition

Shopee's growth mix rotated materially. New active buyers accelerated past 35% growth while purchase frequency slowed to 8% from 12% in Q1. Management explained the rotation as a consequence of improved unit economics: better economics let Shopee profitably serve users it previously could not, which in turn justifies leaning harder into acquisition.

Assessment: This is the correct sequencing and it is what the 64.5% jump in sales and marketing buys. But it front-loads cost against back-loaded revenue, and it means reported margin will understate underlying margin for as long as the cohort is being built. It also raises the stakes on retention data that Sea does not disclose. The claim that acquisition is profitable is currently unfalsifiable from the outside.

3. Take-rate expansion is now partly a price increase

Roughly 176 basis points of marketplace take-rate expansion, of which just over 90 came from advertising. The balance is commission increases, which an analyst raised directly, noting that both Shopee and a major competitor had raised commissions across several Southeast Asian markets this year.

"Going forward, we still see opportunities to increase our take rate, not only from commission, but also from the paid ads we have been able to penetrate more and more over time. Although you can argue that the fixed commissions probably has -- the pace of the fixed commission increase probably will be less than we observed before."
— Tony Hou, Chief Financial Officer

Assessment: This is the most valuable admission on the call. Management is telling you the commission lever is largely pulled and that future take-rate expansion has to come from advertising and seller-efficiency products. That is a higher-quality source of growth but a slower one. Anyone extrapolating this quarter's 176 basis points forward is extrapolating a component management has just guided down.

4. The 2% to 3% Shopee margin ambition, and the distance to it

Asked to lay out the path and the time frame to the medium-term aspiration of 2% to 3% adjusted EBITDA as a share of GMV, management put the current level at roughly 0.67% and described the gap as bridgeable through three levers: investment programmes maturing and requiring less spend, structural cost improvement in logistics and fulfilment, and better take rates.

"Part of that comes from our cost improvement, fundamental cost structure improvements, for example, our logistics, our fulfillment, cost structure improvements over time. Part of that comes from better take rates from either ads or other forms. I think if you put that number together, we are really not too far."
— Forrest Li, Chairman and CEO

Assessment: The levers are credible and the arithmetic is enormous. Tripling to quadrupling the GMV margin on a base growing 25% a year would take Shopee's annual EBITDA from roughly $1B toward $4B or more. What management did not supply, despite being asked directly for it, is a time frame. Treat 2% to 3% as a destination without a schedule, and note that the metric moved backwards year on year this quarter.

5. Logistics and fulfilment as the capital-light moat

Instant delivery order volumes rose around 80% year on year while cost per order fell around 20%, with delivery in as little as one hour in urban areas. Fulfilment volumes grew more than 20% sequentially, and in some markets over 60% of fulfilled parcels arrive next day. Management stressed repeatedly that Sea rents fulfilment space rather than owning land or warehouses, and is experimenting with automation at an early stage.

Assessment: The capital-light framing is supported by the cash flow statement: first-half purchases of property and equipment were $495M against $2,563.9M of operating cash flow. This is the most defensible part of the Shopee story, because falling cost per order at rising volume is a genuine scale effect rather than a pricing decision. It is also the lever least visible in any single quarter's margin.

6. ShopeeVIP crosses 15 million and starts being co-funded

Membership passed 15 million, up 25% sequentially, with VIP members contributing 24% of Asia GMV and retention around 80%. Brazil passed 1 million members within a quarter of its April launch. The new element is that sellers and external partners have begun co-funding benefits across travel, dining and entertainment.

Assessment: Co-funding is the mechanism that turns VIP from a margin cost into a margin asset, because it shifts benefit expense onto counterparties who value access to the cohort. Management says this improved the programme's unit economics in Asia. No figure was given, and VIP's contribution to segment EBITDA has never been disclosed. Directionally right, currently unverifiable.

7. Brazil holds its lead against a repricing competitor

Brazil was the fastest-scaling market for a fifth consecutive quarter, with Shopee Mall GMV more than doubling, nearly 500 new official brands onboarded, buyer waiting time down 15% and fulfilment penetration doubled. An analyst noted that a key competitor had lowered take rates and its free-shipping threshold.

"We still see that our growth is well above the market in the country. And if you look at the price competitiveness, we are still very price competitive, a lot stronger than the competitors in the region, even after their change on the take rate and free shipping threshold."
— Forrest Li, Chairman and CEO

Assessment: Holding share and price leadership through a competitor's deliberate margin sacrifice is the strongest competitive datapoint in the quarter. The move upmarket, through Mall and brand onboarding, is also the right response, because it competes on assortment rather than on price. Brazil remains the single most under-modelled asset in the group.

8. Monee's risk models move to a transformer architecture

Management devoted unusual airtime to underwriting technology, describing models pretrained on behavioural and transactional data across the ecosystem, drawing on external sources including mobile-operator partnerships in Indonesia and open-finance data in Brazil, and using AI to verify income documents with review time cut by around 95%.

"Our latest risk models are pretrained on our broad set of behavioral and transactional data across our ecosystem using transformer architecture similar to those following today's large language models."
— Forrest Li, Chairman and CEO

Assessment: Better underwriting is the enabling condition for everything else Monee is doing, because reaching further from the Shopee transaction graph means underwriting users with thinner first-party histories. Management claims approval rates rose about 10% at similar risk. The stable 1.0% non-performing ratio at 62.5% book growth is consistent with that claim, though a fast-growing book flatters the ratio by construction, and that flattery unwinds when growth slows.

9. Monee's spread compression is a repricing, not a credit event

Revenue up 58.9%, adjusted EBITDA up 12.8%. Management attributed the divergence to product, country and segment mix: off-Shopee lending at structurally lower return on assets, later-stage countries at lower yields, and a deliberate push into prime borrowers who accept lower rates.

"Our guardrail is very simple. We want to maintain stable NPL for the segment, the product, the countries when we look at it. And when we grow new segments, new product, new countries, we wanted to make sure it brings a positive return of the assets to us."
— Tony Hou, Chief Financial Officer

Assessment: Two things are simultaneously true. The credit book is healthy, coverage is rising, and the guardrail is a sensible one. And Monee's earnings power per dollar of loans is structurally lower than it was a year ago, which is not a temporary investment phase but a change in what the business is. The market has been valuing Monee off loan-book growth. That is now the wrong denominator.

10. The adjusted EBITDA definition changes in Q3

Disclosed in a footnote rather than in prepared remarks: from the third-quarter release, Sea will exclude the net effect of changes in Garena's deferred revenue from adjusted EBITDA, while continuing to disclose it separately. This quarter that effect added $20.3M to Garena's adjusted EBITDA. A year ago it added $88.3M.

Restating both periods on the new basis, Garena adjusted EBITDA would be $409.5M against $279.8M, growth of 46.3% rather than the reported 16.7%. Consolidated adjusted EBITDA would be $896.9M against $740.9M, growth of 21.1% rather than the reported 10.6%.

Assessment: The change is defensible on the merits, because a deferred-revenue swing is an accounting timing effect rather than cash generation, and Sea will keep disclosing it. But note the timing and the direction. The new definition lowers the reported level and roughly doubles the reported consolidated growth rate, and it takes effect in the quarter after the prior-year comparison stops flattering the old basis. Investors building a model should restate history now rather than absorbing a growth-rate step change in November as if it were performance.

11. The buyback bought the drawdown

Sea repurchased 4.7 million shares for $416.8M during the quarter under a $1.0B programme, at an average near $88.68. First-half cash used for repurchases was $578M, and treasury stock rose to $510.8M from $14.5M at year-end.

Assessment: Buying at roughly $89 against a stock that closed at $118.32 on 2026-08-27 is a good outcome and, more usefully, evidence that management treats the share price as a valuation input rather than a scoreboard. With diluted share count down slightly year on year despite $131.9M of quarterly share-based compensation, the programme is doing real work against dilution rather than merely offsetting it.

Guidance & Outlook

The GMV reaffirmation came with two caveats management volunteered rather than being pressed into. Currency is the first, with several operating markets weaker against the dollar, and the second-half base is the second.

"If we look forward for the full year, we remain well on track and confident of achieving our full year growth outlook of around 25%. And that said, we want to make sure that we also anticipate the potential ForEx headwind as well."
— Tony Hou, Chief Financial Officer

Implied second-half ramp. First-half GMV was $75.6B. Roughly 25% full-year growth against a first half that grew faster than that implies the second half decelerates modestly, consistent with the tougher base management flagged. On profit, first-half Shopee adjusted EBITDA of $478.6M against a $1.0B full-year goal requires about $521M in the second half, a step up of roughly 9%, and management confirmed the second half will exceed the first in absolute terms. That is the opposite shape to last year and is the single clearest statement of confidence in the guide.

Where the Street sits. Consensus carried roughly $980.7M for Shopee's full-year adjusted EBITDA before the print, so the new $1.0B goal sits about 2% above it. This is not a large raise in absolute terms. Its significance is qualitative: management replaced a floor with a target, which removes the open-ended investment-year framing that had weighed on the stock since March.

Guidance style. Sea guides sparingly and conservatively. It gives one annual GMV growth number for Shopee, now one annual Shopee profit number, and nothing at all for Monee or Garena. Historically the company has cleared what it guides and has raised mid-year when the trend supports it, as it did with Garena bookings in Q2 2025. The absence of any Garena figure this year, after last year's mid-year raise, is conspicuous.

Analyst Q&A Highlights

Whether peak investment is behind Shopee

The opening question of the call went straight to the crux: with margins improving sequentially, is the heaviest spending done, and what does that mean for GMV growth and second-half margins. The answer reaffirmed the annual GMV outlook, flagged currency, and pointed to sequential improvement in unit economics across the newer initiatives without quantifying any of them.

Q: "Firstly, on Shopee, your investments are delivering results on the growth, so can you talk a little bit about outlook for GMV growth? And are we behind peak investments as margins have improved sequentially?"
— Piyush Choudhary, HSBC

A: "In general, we see our unit economics have been improving quarter-on-quarter. I think for our content businesses, which we shared that we did an investment for a period of time. If you look at the unit economic has been as good as the platform already."
— Tony Hou, Chief Financial Officer

Assessment: The content-commerce claim is the concrete part, and it is a real milestone, because live and video commerce is now more than 25% of physical goods orders in Southeast Asia at platform-level economics. On the broader question of peak investment, management declined to say yes. The $1B full-year target answers it implicitly and more credibly than any verbal assurance would.

Whether upside to the GMV guide converts into second-half profit

A direct attempt to establish whether outperformance on volume flows through to earnings, or gets reinvested. Management declined to commit in either direction, framing the allocation as a live decision driven by market growth and competitive conditions.

Q: "In the case that if the GMV were to further exceed the guided growth rate, is it suggested that there is also further upside on the EBITDA for the second half?"
— Alicia Yap, Citigroup

A: "Regarding the balance between growth and EBITDA, it's always a question on what's the best balance between the growth and EBITDA, and I don't think there is simple answer there."
— Forrest Li, Chairman and CEO

Assessment: An honest non-answer, and the correct one for a business genuinely optimising the trade-off in real time. The practical implication for models is asymmetric: treat the $1B as a ceiling for the year rather than a base case to beat, because incremental GMV upside is more likely to be reinvested than banked.

Commission headroom, and an unanswered question on regional profitability

A two-part question asked how much further commission increases can run and, separately, for confirmation that Southeast Asian e-commerce is now profitable. Only the first part was addressed. The second, which would have been the most valuable single disclosure available on the call, was not returned to.

Q: "We've noticed that both Shopee and TikTok shop have raised commissions in several ASEAN markets this year. Could you talk about how much more upside do you think there is for this? And can you confirm if ASEAN e-commerce is now profitable?"
— Divya Kothiyal, Morgan Stanley

A: "From what we observed so far, we have been saying very healthy ecosystem even with the increase of take rate. And the reason for that is that we reinvest a large part of the take rate to the ecosystem growth as well."
— Tony Hou, Chief Financial Officer

Assessment: The framework offered for setting take rates is thoughtful, weighing price leadership against offline alternatives and seller profitability. But the profitability question went unanswered, and regional profitability is exactly the disclosure that would let investors judge whether the $1B target rests on mature markets funding new ones or on across-the-board improvement. Its absence is not neutral.

Where Monee's margin settles

The most analytically important exchange of the call. The question asked directly where fintech margins stabilise, noted that sales and marketing expense keeps rising, and asked when EBITDA growth reaccelerates. The reply reframed the issue as return on assets rather than margin, and located the compression entirely in mix.

Q: "Where do you expect Monee's margins to really stabilize? We did see sales and marketing expenses continue to rise. When should we expect Monee's EBITDA growth to reaccelerate to more healthy levels?"
— Divya Kothiyal, Morgan Stanley

A: "For example, some of the off-Shopee SPL lending growth, which is quite meaningful, more than 20% of the total SPL already, as Forrest shared in the opening, has naturally lower ROA compared to the on-Shopee SPLs. For example, we have been trying to penetrate more to prime segment users, which naturally have slightly lower interest rate, et cetera, and all those growth are intentional."
— Tony Hou, Chief Financial Officer

Assessment: Management answered the "why" completely and the "when" not at all. No level was named and no timeline offered, and the logic given implies compression continues for as long as the newer, lower-yield categories keep outgrowing the core, which is the stated plan. The correct reading is that there is no reacceleration to wait for, only a lower and more durable steady-state spread across a much larger book.

The path to a 2% to 3% Shopee margin

A request for the mechanism and, explicitly, the time frame to close the gap between the current GMV margin and the long-stated medium-term ambition. Three levers were named. No schedule was given.

Q: "I just wanted to understand from here on and back to your medium-term kind of aspiration of 2% to 3%. If you could help us understand the path, the time frame and how you get there."
— Navin Killa, UBS

A: "Part of that will come from -- over time, we don't need to invest so much in many of things we invest in. Like many of the programs we are doing right now, it will get mature over time. So I think we just invest less into it."
— Forrest Li, Chairman and CEO

Assessment: The bridge rests substantially on spending less later, which is the least verifiable of the three levers and the one most exposed to competitive conditions outside management's control. The cost-structure and take-rate legs are stronger. The refusal to date the target, when asked for the time frame in the question itself, is the honest reading of a company that intends to keep the option to reinvest.

The shape of second-half profit

A sharp observation that the new full-year target implies a second half larger than the first, inverting last year's pattern, and a request for the drivers. This drew the most concrete forward commitment of the call.

Q: "You talked about a full year '26 EBITDA to be over $1 billion. That would suggest a higher EBITDA for the second half than the first half, which is different from last year."
— Jiong Shao, Barclays

A: "If you do the math, it does mean that in the absolute terms, our EBITDA for second half of the year will be higher than the first of the year. Of course, partially because of the growth of the market. It's actually the overall GMV, we believe that we still see quarter-over-quarter growth."
— Forrest Li, Chairman and CEO

Assessment: This is the commitment to grade next quarter. It converts a full-year aspiration into a testable half-year statement, and it means a third quarter below roughly $250M of Shopee adjusted EBITDA would put the target at risk. The inversion versus last year is meaningful, because last year's second-half decline was the datapoint that seeded the investment-phase anxiety in the first place.

The provision build and what it signals about credit

The final question of the call asked about the sharp increase in provisions, delinquency trends, and the read-through for loan growth. The answer located the increase entirely in portfolio mix rather than in performance, naming the two components carrying higher loss rates.

Q: "We noticed that the provisions for credit losses have increased quite a bit this quarter. What are the trends that you are seeing in delinquencies? And how does that affect your loan growth going forward?"
— Ranjan Sharma, JPMorgan

A: "I think for the provision, it's primarily driven by the loan mix I think there are two components of our loan mix, we naturally have high provisions."
— Tony Hou, Chief Financial Officer

Assessment: The mix explanation is corroborated by the disclosures. Off-Shopee lending exceeds 20% of the SPayLater portfolio and Brazil is explicitly described as a high-rate, high-risk market, and the 90-day non-performing ratio held at 1.0%. What went unaddressed is the delinquency trend the question actually asked about. Balance-sheet allowance coverage rising to roughly 10.4% of gross loans from 9.6% at year-end is consistent with prudent front-loading against a riskier mix, and it is the number to watch if book growth decelerates.

What They're NOT Saying

1. The effective tax rate. The single largest negative swing in the quarter, from 25.2% to 35.0%, costing roughly $70M of net income and explaining most of the gap between 33.3% operating-income growth and 10.6% net-income growth. It appears in the release only as a number. It was not mentioned in prepared remarks and no analyst raised it.
2. Whether Southeast Asian e-commerce is profitable. Asked directly and not answered. This is the disclosure that would establish whether the $1B target rests on mature markets subsidising expansion or on broad improvement.
3. A time frame for the 2% to 3% Shopee margin. The question specifically asked for the time frame. Three mechanisms were given and no schedule. The ambition has now been carried for several quarters without a date attached.
4. Any Garena full-year bookings guidance. A year ago management raised full-year Garena bookings guidance to above 30% at this same quarter. This year there is no figure at all, against bookings decelerating to 15.5%.
5. The Free Fire tenth anniversary. Named last quarter as a 2027 catalyst and treated as a thesis pillar. Absent from this call entirely, including from a question that asked about upcoming content.
6. The Naruto collaboration timing. Asked directly as part of a two-part gaming question. Management answered the publishing-geography half and did not return to the collaboration timing.
7. ShopeeVIP economics in figures. Membership, retention and GMV share are disclosed. The programme's cost, its contribution to segment EBITDA, and the size of the new partner co-funding are not, which makes the claim that co-funding improved unit economics impossible to verify.
8. The economics of off-book lending. Off-book principal reached $1.1B through channelling arrangements. The revenue share, the risk retention and the return profile versus on-book lending are not broken out, even though the mix shift toward lower-yield lending is the central question on the segment.
9. The adjusted EBITDA basis change, in prepared remarks. A definitional change that would restate consolidated growth from 10.6% to 21.1% on a like-for-like basis was disclosed as a footnote. It was not discussed on the call and no analyst asked about it.

Market Reaction

• Pre-print setup: Closed 2026-08-10 at $114.80, down 10.0% year to date against an S&P 500 up 13.3%, down 21.5% over twelve months, and up 3.3% over the prior thirty days. The 52-week closing range entering the print was $78.16 to $196.50, so the stock sat 41% below its high. A de-rated setup, not a crowded long.
• Reaction session (2026-08-11, before-open print): Opened $127.87, a gap of 11.4%. Traded $124.52 to $131.94. Closed $131.51, up 14.6% or $16.71. Volume 15.2 million shares against a 4.6 million thirty-day average, or 3.3 times normal.
• Benchmark: The S&P 500 fell 0.3% on the session, so effectively the entire move was stock-specific.
• Since the print: The stock closed at $118.32 on 2026-08-27, down 10.0% from the reaction close and up just 3.1% versus the pre-print level, with the S&P 500 flat over the same window. The retracement came on steadily declining volume.

The reaction-day move had one dominant cause. Management replaced an open-ended "no lower than last year" commitment on Shopee profit with a hard $1.0B number, and did so above where the Street was carrying it. That removed the specific uncertainty that had de-rated the stock 41% from its high: whether 2026 was an investment year with no defined end. The 9.9% revenue beat and the acceleration to 48.1% growth supplied the supporting evidence, and the pre-print positioning supplied the fuel.

The retracement is the more instructive half. Nothing negative happened in the twelve sessions that followed, and volume declined throughout, which is the signature of a re-rating being unwound by ordinary supply rather than by a new catalyst. What most likely happened is that the market worked through the income statement after the headline. A print that grows revenue 48% and earnings 10.6% cannot be underwritten on either number alone, and the resolution requires a view on whether the tax rate is permanent and whether Monee's spread compression is a phase or a repricing. Neither question was answered on the call, and neither can be answered from outside. The stock now sits close to where it started, which is a reasonable place for it to sit until Q3 supplies evidence.

Street Perspective

Debate: is Shopee's margin recovery structural or a guided beat?

Bull view: The year-over-year decline is over. Adjusted EBITDA rose 12.2% year on year and 14.4% sequentially and the full-year commitment moved from a floor to a target above consensus, requiring only a 9% second-half step-up into a seasonally stronger period. Cost per order is falling in both instant delivery and fulfilment at rising volume, which is a genuine scale effect that compounds.

Bear view: At 0.67% of GMV, Shopee's margin is below the 0.76% it ran a year ago, so on management's own metric the business has not yet recovered to where it was. The $1B is only 2% above where the Street already sat. Roughly half of this quarter's take-rate expansion came from commission increases that management has just signalled will slow, and the 2% to 3% ambition still has no date.

Our take: The bulls have the better of the near term and the bears have the better of the framing. The $1B will very likely be met, because the required ramp is undemanding and management named the number voluntarily. But it should be read as the year-over-year decline being behind rather than as the margin story resuming, and models that straight-line the 176 basis points of take-rate expansion are extrapolating a component that has been guided down.

Debate: is Monee a scaling engine or a thinning spread?

Bull view: A loan book up 62.5% to $11.1B at a stable 1.0% non-performing ratio, more than 40 million active credit users, five markets above $1B, and rising allowance coverage. Credit penetration across the ecosystem remains low, and the underwriting stack has just taken a genuine technical step forward. Growth of this quality at this scale is rare.

Bear view: Revenue grew 58.9% and adjusted EBITDA grew 12.8%. Segment EBITDA margin fell roughly 840 basis points and quarterly EBITDA against the loan book fell from roughly 3.7% to 2.6%. Provisions grew faster than the book, sales and marketing spend more than doubled, and management named no level and no date at which the spread stabilises. A book growing this fast also flatters the non-performing ratio by construction.

Our take: The bears are right about the mechanics and wrong about the conclusion. This is a deliberate repricing toward a lower-yield, larger, more diversified book, and that is the correct long-term trade for a lender trying to escape dependence on a single transaction graph. But it invalidates the way the segment has been valued. Monee should be underwritten on absolute earnings and on return on assets, not on loan-book growth, and on that basis the segment is worth less per dollar of loans than it was a year ago while being worth more in total.

Debate: is Garena an evergreen franchise or a maturing one?

Bull view: Bookings still grew 15.5% against a demanding comparison, profitability per booking improved, and Free Fire holds more than 100 million daily users in its ninth year. Two titles built on globally recognised intellectual property are due, one self-developed for global publishing. The content engine is proven and the monetisation runway in paying ratio is real.

Bear view: Quarterly active users were flat year on year. All growth came from converting and charging the existing base harder, which has a ceiling. Bookings decelerated for a second consecutive quarter, no full-year guidance was offered where a raise came a year ago, the tenth-anniversary catalyst went unmentioned, and a definitional change is about to make the reported growth rate look better without the business changing.

Our take: The bears are describing this correctly. A flat user base with rising monetisation is a maturing franchise being harvested well, not an expanding one, and the missing guidance is the tell. That is not a crisis: Garena still generated $429.8M of adjusted EBITDA on modest capital and funds the rest of the group. But it should be modelled as a durable cash annuity with option value on two unlaunched titles, not as a growth engine. The right response is to lower the growth assumption and leave the cash assumption alone.

Model Update Needed

Resulting estimates. We carry FY26 total adjusted EBITDA of roughly $3.8B, built from the guided $1.0B at Shopee, roughly $1.18B at Monee, roughly $1.85B at Garena on the current basis, and roughly $215M of Other Services losses and unallocated cost. First-half actuals of $1,951.5M make the implied second half about $1.86B. For FY27 we carry roughly $5.25B, with the largest single swing factor being whether Shopee's GMV margin reaches 0.95% or stalls near 0.75%.

Valuation. At the 2026-08-27 close of $118.32 on 634.5 million diluted shares, the equity is worth about $75.1B. Cash and equivalents of $3,529.3M, short-term investments of $5,569.6M and long-term investments of $2,756.1M, less borrowings of $1,224.3M and convertible notes of $996.3M, give roughly $9.6B of net cash and investments. We exclude restricted cash and deposits payable on both sides as customer float, and exclude net loans receivable as working capital of the credit business. That puts enterprise value near $65.4B, or about 17.2 times our FY26 adjusted EBITDA and about 12.5 times FY27. On reported earnings the stock trades near 40 times our FY26 GAAP diluted EPS estimate of roughly $2.95, which is why enterprise value to EBITDA is the more useful lens while depreciation, share-based compensation and the tax rate are all in motion.

Fair value range: $150 to $165, equal to roughly 16 to 18 times FY27 adjusted EBITDA on an enterprise-value basis, adding back net cash. Against the $118.32 close on 2026-08-27 that implies 26.8% to 39.5% upside. We are raising the floor of the prior $130 to $170 range on the Shopee guidance raise and trimming the ceiling on the tax rate and Monee's spread, which together lower our FY27 earnings conversion even as revenue runs ahead of plan.

Thesis Scorecard Post-Earnings

Overall: The thesis is intact above the operating line and weaker below it. Four bull pillars confirmed outright, one confirmed on volume while its economics deteriorate, two neutral or unaddressed, and one challenged. The bear case has shifted rather than resolved: the margin compression that defined Shopee through 2025 has been contained there and has reappeared in Monee, and a new drag has opened up in tax. Revenue growth is better than we modelled and earnings conversion is worse, and on balance those roughly offset at the enterprise-value level while lowering reported earnings quality.

Action: Maintaining Outperform and moderating conviction from High to Moderate. The post-print retracement to $118.32 restores most of the risk/reward that the reaction session removed, and the fair value range of $150 to $165 implies 26.8% to 39.5% upside. The two things to grade next quarter are whether Shopee's third-quarter adjusted EBITDA clears roughly $250M, which is what the second-half commitment requires, and whether the effective tax rate stays near 35%. A miss on the first would break a commitment management volunteered. A confirmation of the second would permanently lower our earnings estimates without changing the operating story.