The earnings season in autumn 2026 amounts to little more than a midterm exam.
One evening before the 2025 Lunar New Year, Liu Qiangdong hosted Wang Xing for dinner at a restaurant in Beijing. During the meal, Liu said, "Brother, I'm officially entering food delivery."
The remark landed like a stone in a pond. Four months later, Liu Qiangdong put on a red delivery uniform and made several deliveries on Beijing's streets. That night, he shared a hotpot with 40 or 50 riders. Raising his glass, he said, "I also made a point of delivering a few orders this afternoon."
At the time, few expected the battle to drag on so long, and even fewer anticipated that when the three platforms reported earnings together in autumn 2026, a year and a half later, the meaning of "victory" would have shifted fundamentally.
On Aug. 28, Meituan reported second-quarter 2026 results. Operating profit for core local commerce reached 5.67 billion yuan, a margin of 7.9%. A quarter earlier, the segment had lost 2.03 billion yuan.
For Wang Xing, the "stop-the-bleeding" report arrived not a moment too soon. In Q3 2025, Meituan's core local commerce recorded a loss of 14.07 billion yuan, the fiercest point of the war. Now profitability has returned, and although the 7.9% margin is less than four-tenths of its pre-war level, the defender's walls have not collapsed.
Eight days earlier, on Aug. 20, Alibaba turned in its own scorecard: China instant-retail revenue rose 45% to 53.295 billion yuan. In the results, CEO Wu Yongming played down peak order volumes, instead stressing that "unit economics for Taobao Flash Purchase improved sequentially," citing "higher ticket size and fulfillment efficiency."
Alibaba's win isn't about the top line; it's about the economics of each order. As JD.com pounded the market with billions in subsidies, Alibaba carved its own route: rather than who delivers more, it's who delivers better value.
JD.com was the first to report. On August 13, its second-quarter 2026 results showed the new business segment, including food delivery, posted an operating loss of 9.85 billion yuan, narrowing by 4.92 billion yuan from the same period in 2025 and nearly 5.9 billion yuan from the 15.74 billion yuan peak in the third quarter of 2025. CEO Sandy Xu wrote in the earnings release: "The profit trajectory has reached a clear inflection point."
From 15.74 billion to 9.85 billion, JD cut losses by nearly 40% in three quarters. But Xu's wording was cautious: she said "inflection point," not "profitability." JD's delivery business is still bleeding, though the bleeding has slowed.
What Was Left Behind?
In the early stages of the war, scale was the only ticket in.
In April 2025, Richard Liu laid out the math plainly at JD's headquarters: "I can lose money on meals forever, as long as we make money on supply chain." He added: "With 40% cross-selling to our e-commerce, the money we lose is still cheaper than buying traffic from Douyin or Tencent."
That was the first narrative: JD's delivery business is not really about delivery; it's a gateway to supply chain and a customer-acquisition channel for e-commerce.
To sell that story, JD cited three figures in its second-quarter 2025 report: daily order peak above 25 million during the 618 shopping festival, more than 1.5 million "quality merchants" onboard, and over 150,000 full-time riders. Xu declared then that delivery had "successfully achieved our initial strategic goals" — meaning JD had earned a seat at the table.
Liu Qiangdong's supply-chain payment story has yet to be validated.
Alibaba's pivot is more direct.
In August 2025, Eddie Wu said on an earnings call: "We invested heavily in instant retail, quickly achieved initial results, and won consumer mindshare." At that time, Taobao Flash Purchase had been live for four months, and its primary task was to prove it could bring high-frequency consumption back to Taobao.
Three months later, the narrative was slightly adjusted: "Instant retail scale expanded and unit economics improved significantly." Another three months later: "Unit economics continued to improve sequentially while maintaining market share."
Alibaba's three statements were like a microscope steadily adjusting its focus—from consumer mindshare to scale, then to average order value and fulfillment efficiency. What Wu has been trying to prove is never Ele.me's scale expansion, but whether each Taobao Flash Purchase order can cover its own costs while generating additional customer-management revenue for Tmall.
Meituan, the defender, is also using the opportunity to take stock.
In May 2025, Wang Xing said on an earnings call: "Low-quality, low-price involution-style competition is unsustainable in the long run." This was not a sign of weakness but a declaration of war—Meituan aims to defend not just orders, but also the quality of its merchants, couriers, and delivery network.
Six months later, Meituan disclosed that its daily instant-retail order peak exceeded 150 million, with an average delivery time of 34 minutes. Wang Xing's language shifted to "consolidate core competitiveness and maintain industry leadership"—after paying out subsidies, he needed to prove that order and network density had not been broken up by the two attackers.
By the second quarter of 2026, Wang Xing had changed his tune again: "Meituan firmly opposes involution. We will cooperate fully with regulatory investigations. At the same time, while working to consolidate our market-leading position, we will also reduce resource investment in low-quality orders."
The War Ledger:
How Much Real Money Was Burned?
Narratives can shift, but the ledger doesn't lie.
Extend the observation window, and the price the three companies paid for this war becomes clear.
From Q1 2025 to Q2 2026, JD.com's marketing and fulfillment expenses were RMB 57.83 billion higher than the corresponding year-ago baseline; Meituan's logistics and promotional incentive expenses were RMB 82.74 billion higher. From the launch of Taobao Flash Purchase in Q2 2025 through Q2 2026, Alibaba's sales and marketing expenses were RMB 95.45 billion higher.
All three figures are on a broad group basis, including spending on retail, warehousing and distribution, other instant retail, overseas operations, and user growth. Even so, nearly RMB 240 billion in incremental costs would keep any CFO awake at night.
The question is: what did all this money buy?
National Bureau of Statistics data offers a sober answer. After revisions from the fifth national economic census, China's catering revenue was RMB 5.618 trillion in 2024, up 5.3% year on year; in 2025 it rose to RMB 5.798 trillion, with growth slowing to 3.2%; in the first half of 2026, catering revenue was RMB 2.826 trillion, up 2.8% year on year.
In July 2025, the peak of the food-delivery subsidy war, China's catering revenue reached 450.4 billion yuan, up just 1.1% from a year earlier. Revenue at catering companies above a designated size fell 0.3%.
In other words, platform orders can shift quickly under heavy subsidies, but overall dining demand doesn't double in tandem. The delivery war is essentially a redistribution of market share: orders that JD.com and Alibaba took from Meituan didn't make Chinese people eat more meals; they just moved the same meal from platform A to platform B.
That means the three platforms' improvements can't all be built on growth in the overall dining market. If someone wins, someone must lose.
In all likelihood, consumers captured the subsidy gains.
AI Takes Center Stage
Just as the three platforms were changing the scoreboard for food delivery, another, costlier narrative moved to the center of earnings: AI.
Alibaba was the most aggressive. Capital expenditure reached 67.68 billion yuan in the second quarter of 2026, up 75% from a year earlier, mainly for AI infrastructure. Alibaba Cloud's external commercial revenue grew 45%, and AI-related product revenue posted triple-digit growth for the 12th consecutive quarter. For CEO Eddie Wu, AI is not a cost-cutting tool for delivery but a second growth pillar alongside consumption.
Meituan is embedding AI into the capillaries of local life. R&D spending in the second quarter was about 7.7 billion yuan, up 22.5% from a year earlier, or 7.3% of revenue. 'Xiaotuan' helps users make decisions, CatPaw serves merchants, and 'Tuanbao' protects rider safety. CEO Wang Xing said: 'We will firmly increase investment in ecosystem and technology, and push AI into real business scenarios.'
JD.com didn't disclose separate AI spending but emphasized supply-chain efficiency. In the first half of 2026, JD Industrial deployed more than 70 AI agents across the procurement-to-fulfillment chain. Unlike Alibaba's search for a second curve or Meituan's revamp of local-life entry points, JD is still trying to put AI back into the gears of its supply chain.
All three companies are pushing AI to the forefront, but the rules of the ledger have not changed. AI must prove revenue or cost reduction; food delivery must prove retention, cross-selling, and supply-chain profit. Both future narratives ultimately converge on the same income statement.
Forecast: Three Strategies for the Second Half
A year into the food-delivery war, the three platforms' next moves may take three distinct approaches.
JD.com is entering a verification period for supply-chain payback.
The RMB 9.85 billion quarterly loss looks large, but Xu Ran's talk of an inflection point is not without basis. JD.com's real leverage lies not in front-end delivery but in its back-end logistics and supply-chain network. If the cross-selling and supply-chain efficiency gains from delivery show up in retail margins over the next four quarters, JD.com's narrative closes the loop. Conversely, if retail profit cannot cover the delivery cash burn, JD.com's 'supply-chain payback' story will face more skepticism.
Forecast: JD.com's delivery business will continue to narrow losses in the second half of 2026, but full-year new-business losses may still exceed RMB 30 billion. The real test comes in 2027.
Alibaba's next issue is the standalone P&L of Taobao Flash Purchase.
Instant-retail revenue of RMB 53.295 billion grew 45% year on year, but Alibaba has never separately disclosed Taobao Flash Purchase's profit or loss. 'Sequential improvement' in unit economics is a vague phrase — improved by how much? How far from breakeven? Alibaba needs to provide a clearer per-order economic model in the next two quarters. That said, the incremental business delivery brings to traditional e-commerce is clearly visible.
Forecast: Alibaba will first disclose Taobao Flash Purchase's standalone operating data in Q3 or Q4 2026. If per-order losses are near breakeven, Alibaba's instant-retail narrative will be re-rated.
Meituan is being driven by two forces: profit recovery and AI efficiency.
The quarterly profit of 5.67 billion yuan is a positive signal, but the 7.9% margin remains far from the 21% seen before the war. Meituan's recovery path hinges on two variables: whether industry competition truly turns rational—with regulators continuing to curb subsidy wars—and whether AI can deliver meaningful efficiency gains in rider dispatch, merchant operations, and user recommendations.
Forecast: Meituan's core local commerce margin will recover to the 10%-12% range in the second half of 2026, but is unlikely to return to pre-war levels within the year. AI-driven improvements in fulfillment efficiency will begin to show up in financial statements in 2027.
Conclusion: A War With No Endgame
At a dinner before the 2025 Spring Festival, when Richard Liu told Wang Xing, 'I will formally enter food delivery,' neither probably imagined the fight would become a war of attrition lasting 18 months and costing hundreds of billions of yuan.
Even more unexpected is that, with the war still ongoing, 'winning' has come to mean many things. JD.com must prove its supply chain can generate cash, Alibaba must prove every order is economically sound, and Meituan must prove its network can still make money after defending its turf.
All three companies say they are getting better. But getting better is not winning, cutting losses is not profitability, and improved unit economics is not the same as being profitable. The food-delivery war has no endgame; it has only moved from the first phase of burning cash to grab territory to a second phase of proving the business model.
This earnings season in autumn 2026 is merely the first midterm exam of the second phase. The real report card may not be released until spring 2027.
By then, the AI story may have already rewritten the rules of the entire game.