A decade after WeChat’s 2014 “red‑envelope” gambit upended China’s payments market, the Lunar New Year has once again become a theatrical battleground — this time for consumer AI. What began as cash giveaways embedded in social chat has evolved into a complex, multi‑platform experiment in product education and user acquisition, with Tencent, Alibaba, ByteDance, Ant and Baidu together spending more than 45 billion yuan during the 2026 Spring Festival.
The mechanics of the 2026 campaign were sophisticated. Tencent’s Yuanbao kicked off with 10 billion yuan in cash prizes that required users to log in daily, hold AI conversations, generate images and write couplets before accessing draws; Alibaba’s Tongyi Qianwen offered some 30 billion yuan worth of 25‑yuan milk‑tea vouchers redeemable at hundreds of thousands of physical stores; ByteDance’s Doubao dominated the CCTV New Year Gala slot with AR scanning and voice‑activated interactions; Ant and Baidu deployed smaller but sustained offers, extending some activity windows into late February and even March.
These activations were designed less as simple subsidies than as purpose‑built funnels to teach users how to interact with chatbots and generative tools. Platforms traded one‑off payouts for task thresholds and timed withdrawals, turning a short burst of holiday attention into a sequence of first‑use experiences. The Spring Festival and the gala remain a rare national-scale distribution moment, and firms used it to force trial of AI features across age groups and regions.
Yet the arithmetic that underwrites such campaigns is unforgiving. Industry benchmarks treat roughly 10 yuan as a baseline customer‑acquisition cost and aim for a seven‑day retention of around 20 percent for a viable mainstream product. Historical precedent is bleak: after Baidu’s 2019 Spring Festival spend, third‑party data showed seven‑day retention near 2 percent. If Alibaba’s 30 billion yuan produced 100 million new downloads but only 2 percent survived a week, the implied cost per retained user could exceed 1,500 yuan — a figure that would be unsustainable absent rapid monetisation or durable engagement.
Which companies are best placed to turn ephemeral spikes into lasting habits depends on distribution and ecosystem leverage. ByteDance benefits from Douyin’s native traffic and content recommendation engine, Alibaba’s milk‑tea vouchers attempt to bind online attention to offline spending and Ant taps payments stickiness, while Tencent leans on the intimacy of social graphs in WeChat. But high initial installs and peak daily active user figures tell only part of the story; the real prize is habitual use and functional dependence on a given assistant.
The broader significance is clear: this was not merely a marketing splurge but a large‑scale product‑market test for consumer AI. The 45 billion yuan experiment will produce fast answers about which stimulus designs convert trial into habit, which ecosystems can lower acquisition costs, and how much value users place on AI features beyond novelty. When the Spring Festival lights dim and group chats return to normal, the companies that engineered genuine changes in day‑to‑day behaviour — not merely momentary headline metrics — will have won the strategic round.