China's Chatbot Wars: 144 Million Users but Retention Is the Real Battle

China's AI chatbot market has exploded in scale, with ByteDance's Doubao reaching 144 million daily active users — a number that dwarfs ChatGPT's global user base. But beneath the staggering headline figures lies a more complicated reality: much of that growth was driven by holiday promotions and platform bundling, and the industry is now confronting the far harder question of whether users will stick around.

The Doubao Surge

ByteDance's strategy was straightforward and effective. By integrating Doubao directly into Douyin — China's version of TikTok, with over 700 million daily users — the company gave its chatbot instant distribution that no standalone app could match. The Lunar New Year TV gala partnerships amplified the effect, introducing Doubao to hundreds of millions of viewers in a single evening. The result was a user acquisition curve that broke every existing record in the AI chatbot category.

But integration into Douyin is a double-edged sword. Many of those 144 million daily users are interacting with Doubao passively — encountering it within the Douyin experience rather than deliberately seeking out an AI assistant. The distinction between an active user and an incidental one matters enormously for long-term viability.

Alibaba's Promotion-Fueled Sprint

Alibaba took a different approach. On February 7, Qwen hit 73.5 million users through a milk tea delivery promotion run through Alipay. Order a drink, chat with Qwen. The tactic was clever — it embedded AI interaction into an existing consumer behavior — but it also raised the question of whether users were coming for the chatbot or the free milk tea.

The data suggests the answer was largely the latter. Daily usage metrics dropped sharply after the promotional period ended, a pattern that repeated across multiple chatbot platforms running similar campaigns during the holiday season.

The Retention Problem

China now has more AI chatbot companies than any other country, and nearly all of them are burning money to acquire users. The apps themselves are remarkably versatile by Western standards — handling shopping, payments, deliveries, and entertainment in addition to the conversational AI functions familiar to ChatGPT users. But versatility has not translated into habitual usage.

The core issue is that most Chinese consumers do not yet have a daily workflow that requires an AI chatbot. Super-apps like WeChat and Alipay already handle payments, messaging, and services. Adding an AI layer on top of these existing behaviors requires changing user habits, not just offering a new feature.

Enterprise and Vertical Applications

Some companies are finding more sustainable traction by focusing on industry-specific applications rather than the consumer chatbot wars. Ping An, China's insurance giant, has deployed AI assistants that handle claims processing and customer service at scale. Manufacturing companies are using chatbot interfaces for equipment diagnostics and supply chain management. These use cases generate recurring engagement because they solve specific professional problems rather than competing for casual consumer attention.

The vertical approach may prove more defensible than the consumer land grab. Enterprise customers who integrate an AI chatbot into their operations are unlikely to switch providers on a whim, creating the kind of retention that consumer apps struggle to achieve.

The US Comparison

In the United States, ChatGPT has maintained steadier usage patterns, in part because it established itself as a productivity tool before the competitive field became crowded. American users tend to interact with chatbots for work-related tasks — writing, coding, research — that generate habitual daily use. China's chatbot market, by contrast, has been shaped by consumer entertainment and e-commerce dynamics that produce dramatic spikes followed by equally dramatic declines.

The Monetization Question

The retention challenge feeds directly into the broader AI business model question. Companies that cannot retain users cannot monetize them, and the cost of continuous user acquisition through promotions is unsustainable. China's chatbot companies are spending billions on subsidies, free API access, and marketing partnerships, but the path from massive user numbers to profitable recurring revenue remains unclear.

The market will eventually consolidate. The companies that survive will be those that solve the retention problem — either by becoming genuinely indispensable to daily life or by finding enterprise niches where switching costs are high. The 144 million daily user figure is impressive, but it is the 30-day retention rate that will determine who wins China's chatbot wars.