China’s mid-year shopping marathon turned 138 billion dollars this year, and almost nobody cheered. According to retail data firm Syntun, this year’s 618 edition generated roughly 934 billion yuan (about US$138 billion) in gross merchandise value between May 13 and June 18. Its growth, however, slowed to 4 per cent from 15.2 per cent last year. A festival without the frenzy This year’s event on platforms including JD.com and Alibaba’s Tmall ran roughly 40 days from mid-May, a
id-May, and three to four shopping days longer than last year, depending on the platform. That extended window was widely expected to lift headline revenue by a low single digit, even as underlying demand remained soft.
The campaign has simply gone on too long compared to the 24 hours when it initially started. The perpetual-discount model has dulled the sense of urgency that once made it irresistible. When everything is on sale all the time, nothing feels like a deal.
In the same month the festival was ramping up, the country recorded its first decline in monthly retail sales since it emerged from Covid lockdowns, a 0.6 per cent year-on-year contraction in May, according to the National Bureau of Statistics, the first such drop since December 2022.
Over the first five months of the year, total retail sales of consumer goods rose just 1.4 per cent, and a broader gauge of goods and services combined rose 2.8 per cent, with services up 5.4 per cent but goods up only 1.2 per cent.
Into that subdued mood, the e-commerce platforms poured artificial intelligence. Industry insiders dubbed 2026 “the first AI-native shopping festival” and for once, the marketing was not entirely hyperbole. Alibaba rolled out its Wanxiang engine and the full integration of its Qwen model into Taobao, letting shoppers chat their way to product selection, virtual try-ons and price calculation. JD.com deployed proprietary large language models across logistics and operations, and ByteDance rolled out AI tooling and agentic checkout features on Douyin.
The dissenting read: GMV grew, net revenue didn’t
“This time around, we feel that it is quite quiet. I believe this is a good thing for the market. This shows that people’s consumption patterns are normalised, people don’t just stock up during shopping carnivals,” Derek Deng, Bain and Co’s head of consumer products in Greater China, told Reuters.
However, not everyone buys the maturity story, and the sharpest pushback comes from inside the industry. Jacques Roizen, managing director of consulting at Digital Luxury Group (DLG), argues that the celebratory headlines mistake an accounting artefact for a behavioural shift. “You will read that China’s second largest online shopping festival was a great success, that it has moved beyond discounting, and proven that Chinese consumers now reward brand and innovation over price,” he wrote in a post-festival assessment. “It’s a great story, but it’s wrong.”
GMV, Roizen notes, is “the only KPI that’s publicly accessible”, and this year’s, in his view, “is the most distorted it has ever been.” Strip out returns and cancellations, he contends, and most brands did not grow net revenue at all. He points to two mechanics that, he says, pumped up the top line. The promotional calendar was quietly stretched: although the platforms restored the optics of a cleaner schedule – returning the 5/20 shopping moment to its own slot and “officially” opening 618 on May 21 – brands were asked to begin discounting as early as May 6, making for “a cleaner calendar in theory, but longer discounting in reality.”
The second distortion, in his telling, is a coupon mechanism the platforms leaned on this year but not last June: discounts that unlocked only once an order crossed a spending threshold. To qualify, Roizen says, many consumers “padded their baskets beyond what they meant to buy, captured the discount, then cancelled the excess”, leaving GMV “structurally overstated vs net sales, this year even more than last year.”
The beauty category lost a crucial source of demand, he adds. With most prestige brands still dependent on top livestreamers for a large share of online revenue, the decision by China’s biggest beauty streamer, Austin Li (Li Jiaqi), to cut his number of sessions across May and June “pulled real demand out at the worst possible moment” for those most reliant on him.
And the brands that did hold flat or grow, Roizen argues, did so by offering more discounts, more promotional days, or a higher share of their assortment on markdown.
Measured in net revenue rather than GMV, his read is blunt: most brands are down high-single to double digits versus last year, and the category leaders in sportswear and prestige beauty are down more than 20 per cent. A handful, he allows, turned the revenue softness into genuine profit improvement “which should always be the real objective.”
His verdict on the festival as a whole is the line the official data cannot supply: “618 didn’t move beyond the price war. Platforms just got better at describing it.”
Roizen’s critique does not contradict the topline figures so much as reframe what they measure, and it sharpens the central tension of this year’s 618. The platforms simplified the discounts and dressed up the calendar; the consumer stayed cautious and value-driven. And the gap between a festival’s reported scale and its real profitability has rarely been wider.
Further reading: Subsidies spark a retail surge, but can China’s momentum last?