NextFin News - Ant Group's profit appeared to flatten in the latest reported quarter, but the headline stability is an accounting illusion. The fintech giant's earnings held steady only because one-off investment revaluations swelled to cover a deepening drag from artificial-intelligence and overseas expansion spending — a composition shift that points to a structural reset of Ant's profitability, not a temporary dip.
Alibaba Group, which owns a 33% stake in Ant and reports the fintech's results one quarter in arrears, said its share of Ant's profit for the three months ended June 30 was RMB1.563 billion, essentially unchanged from RMB1.547 billion a year earlier. On the surface, that is a story of stabilization after a brutal stretch in which Ant's contribution to Alibaba collapsed. Look beneath it, and the picture is different: Alibaba explicitly attributed the flat result to "the net increase in fair value gains" being "offset by the increased investments in new growth initiatives and technologies." In other words, the operating engine weakened while paper gains on investments did the heavy lifting.
The setup raises the question this piece pursues: is Ant's profit stagnation a cyclical growth-phase cost bulge that will snap back, or the first visible symptom of a structural margin reset as the company trades its old high-margin lending economics for an AI-and-infrastructure future? The evidence points to structural — and that has consequences for how investors should read not just Ant, but Alibaba's own AI pivot.
The Number That Wasn't Flat
The reported figure — RMB1.563 billion of profit attributable to Alibaba's stake — is the cleanest entry point, because it comes straight from Alibaba's June-quarter earnings release dated August 20, 2026. But three layers of context turn a "flat" headline into something sharper.
First, the timing. Alibaba records Ant with a one-quarter lag, so the June 2026 share of profit reflects Ant's quarter ended March 31, 2026. That March quarter is where the damage shows: Ant's underlying net profit fell roughly 60% year over year to about RMB4.74 billion, according to calculations drawn from Alibaba's filing. The prior December quarter was worse — an estimated 79% decline to RMB1.13 billion. Contrast that with the June 2025 quarter, when Ant's net profit still grew 10.3% to RMB8.37 billion. The trajectory is not stabilization; it is a step-down in the earnings base that fair-value gains are papering over.
Second, the annual view removes any doubt. For the fiscal year ended March 31, 2026, Alibaba's share of Ant's profit fell to RMB5.048 billion from RMB12.648 billion in fiscal 2025 — a decline of about 60%. A single quarter of flat contribution does not reverse a year in which the contribution more than halved.
Third, the mechanism is disclosed in the company's own words. The flat share of profit was not the product of resilient operations; it was the arithmetic of two offsetting forces. Fair-value gains — the revaluation of investments Ant holds on its balance sheet — rose. Investments in "new growth initiatives and technologies" rose faster. This is not a company coasting; it is a company spending its way into a new identity.
There is also a cash counterpoint worth holding alongside the profit line. During fiscal 2026, Ant paid Alibaba a dividend of RMB3.293 billion. A business that can still distribute billions in cash is not distressed. But dividends are drawn from accumulated earnings and balance-sheet strength; they do not speak to the forward earnings power that the AI spending is meant to create.
A Structural Margin Reset, Not a Cyclical Dip
The central judgment here is cyclical versus structural, because it decides everything that follows. A cyclical cost bulge — a marketing push, a temporary hiring wave, a one-off regulatory fine — fades on its own. A structural reset does not revert; the margin profile changes permanently because the business model changes.
Three pieces of evidence argue this is structural.
The first is the scale and composition of the spending. Ant's research and development outlay reached a record RMB23.45 billion in 2024, up 10.7% year over year, with roughly 40% directed at AI and machine learning, the company said in its 2024 sustainability report. That is not a pilot budget; it is a core-capability build-out. AI infrastructure — model training, inference compute, data centers — carries a cost structure that does not switch off. Unlike a sales campaign that ends when the quarter closes, compute spend compounds with usage. Once Ant commits to running large models across payments, credit scoring, insurance claims, and wealth advice, the expense line is structurally higher.
The second is the business-model shift underneath the spending. Before China's 2020-2021 regulatory crackdown, Ant's profitability rested on a lightly capitalized lending model that routed consumer credit through its own balance sheet. Regulators forced a reversal: partner banks now fund the overwhelming majority of Ant's consumer loans, while Ant retains only a small residual share. That shift reduced regulatory risk — a genuine achievement — but it also compressed the take-rate Ant captures on each loan. The company is now rebuilding growth on the other side of that compression, through technology services and international expansion rather than balance-sheet leverage. A lower-margin revenue mix is a structural fact, not a cycle.
The third is the international build-out. Ant has been expanding abroad through payments and cash-management services, its cross-border payments platform Alipay+, and a move into securities brokerage: in April 2025 it agreed to buy a controlling stake in Hong Kong-listed Bright Smart Securities & Commodities Group for about HK$2.8 billion, clearing Chinese regulatory review and moving toward completion. The scale of what Ant is building overseas is substantial. Alipay+ connects more than 1.8 billion user accounts across 40 international payment partners to merchants in more than 100 markets, and Ant International — the overseas unit spun out in 2024 — reported cross-border transaction volumes surging threefold in 2024. WorldFirst, its global business-to-business payment service, reached US$100 billion in annual total payment volume. Geographic expansion at this scale carries upfront costs that precede profits by years, and it is a deliberate multi-year bet, not a quarterly fluctuation.
Put together, these are not mean-reverting items. They are a new cost base attached to a new revenue model. The fair-value gains that kept the June-quarter profit line flat are the only genuinely cyclical (and unreliable) component in the mix — revaluations can reverse as quickly as they appear, and they tell you nothing about operating leverage.
Ant as Alibaba's AI Bet in Miniature
Ant's story is not isolated. It is Alibaba's own strategy reflected in a subsidiary. In the same June-quarter release, Alibaba reported net income down 75% year over year and non-GAAP net income down 38%, while adjusted EBITA fell 30% — a decline the company attributed "primarily to the investment in technology." At the same time, Alibaba Cloud's external revenue growth accelerated to 45%, with AI-related product revenue posting triple-digit growth for the twelfth consecutive quarter.
The parallel is exact: both companies are accepting a near-term earnings hit to buy AI optionality. Alibaba's chief executive, Eddie Wu, framed it directly: "We delivered a strong quarter, driven by the improving commercialization of our full-stack AI capabilities... With our full-stack AI strategy, we have put Alibaba in a superior position to capture the substantial growth of demand for artificial intelligence and AI compute." Chief financial officer Toby Xu added that as "AI monetization ramps up," the group has "greater strategic and financial flexibility to make disciplined and sustained investments in full-stack AI capabilities."
That framing matters for Ant because it signals the parent's tolerance. Alibaba is not going to pressure Ant to squeeze margins back to earlier levels if doing so would starve the AI build-out. The profit flattening is, in effect, sanctioned by the owner. But tolerance is not the same as proof of return. Alibaba's own cloud AI revenue is growing at triple digits from a small base; the question of whether that growth converts into profit at the old margins remains unanswered.
This is where the second-order implication bites. The market has largely cheered Alibaba's AI pivot — the stock had climbed more than 40% from its 2026 low ahead of the August 20 earnings release. Investors are pricing the revenue story. What they are not pricing, or are discounting lightly, is the possibility that the AI economy in China becomes a high-revenue, low-margin utility: enormous top-line growth, relentless capex and opex, and a terminal margin well below the software-like levels that fintech once commanded. Ant is the early warning signal for that outcome. The same dynamic plays out across the group: revenue growth that looks like a recovery can coexist with profit that never comes back to where it was.
The Counter-Thesis, and What Would Prove It Wrong
The strongest case against the structural-reset reading is that Ant is doing exactly what a dominant franchise should do: reinvesting a mature cash flow into the next moat. From this angle, the profit decline is the price of admission to the AI era, and the payoff is a wider, more defensible business. The evidence is not thin. Ant holds 23,550 patents globally, more than 95% of them invention patents, according to the company. Its distributed database OceanBase more than doubled its client base in 2024 to over 2,000 enterprises, and its AI-driven services across health, finance, and daily life reached more than 130 million users. The regulatory overhang that destroyed the 2020 IPO has lifted; the company is no longer fighting for survival, but building for a second act. And it is not short of capital: Ant International raised US$1.2 billion in a 2026 funding round, giving outside investors a vote of confidence in the overseas strategy.
This counter-thesis deserves its weight. If Ant's AI and technology services scale with software-like gross margins — and if the international units reach profitability on a path measured in quarters, not years — then today's spending will look like a bargain in hindsight, and the "structural reset" call will have mistaken a growth investment for a margin impairment.
So the falsifying signal must be concrete. The structural-margin-reset thesis is wrong if, over the next two consecutive reported quarters, Ant's underlying operating profit (excluding fair-value gains on investments) grows year over year while AI-related revenue growth stays above 40%. That combination would prove the spending is buying incremental high-margin revenue fast enough to offset the lending-compression drag — the definition of a cyclical bulge resolving, not a regime shift. Conversely, if profit continues to fall or flatline while fair-value gains remain the balancing item, the structural read is confirmed. The metric to watch is not the headline share of profit Alibaba discloses, but the split between operating results and investment revaluations that the company itself highlights.
What to Watch: Scenarios Across Time Horizons
Short term (one to two quarters): The profit line stays volatile and partly opaque. Fair-value swings will keep obscuring the operating trend, so the number to watch is not the headline contribution but the company's own decomposition — how much of any change comes from investment revaluation versus operations. A quarter where profit rises without a fair-value tailwind would be the first real green shoot.
Medium term (six to eighteen months): The decisive variable is AI monetization. Alibaba's claim of triple-digit AI revenue growth must translate into Ant-specific metrics: AI-driven payment and credit volumes, insurance-claim automation adoption, and enterprise database contracts. If these scale while the cost base keeps rising, margin compression persists even as revenue grows — the classic "growing into the spend" problem.
Long term (two years and beyond): The structural question resolves one of two ways. In the upside case, Ant becomes a high-margin technology vendor to the financial system — payments, risk, database, and AI services sold to banks and enterprises globally, with the partner-bank lending model a stable, regulated annuity underneath. In the downside case, Ant becomes a lower-margin, capital-intensive AI utility: essential infrastructure, heavy spend, and a permanently thinner profit slice.
Base case: a multi-year margin reset with revenue growth that only gradually catches up to the cost base. Upside case: AI services achieve software-like margins faster than expected, validating the spend. Downside case: fair-value gains fade while operating profit keeps sliding, and the flat headline gives way to an outright decline.
The Takeaway
Ant Group's flattened profit is the kind of number that rewards skepticism. Stability built on investment revaluations is not stability at all — it is a bridge between a fading earnings model and one that has not yet proven itself. The company is spending heavily and deliberately, with the backing of an owner making the same bet at scale. That may yet be the right call. But the market should stop reading "flat profit" as "recovery." It is a transition, and transitions are where margins go before they come back — if they come back at all.
Explore more exclusive insights at nextfin.ai.
