Analysis of the First AI-Driven Chinese New Year User Battle

Deep News09:31

The inaugural large-scale, subsidy-driven competition for users in China's consumer AI sector occurred during the 2026 Lunar New Year holiday. According to a detailed review by Morgan Stanley, major players including Alibaba (Tongyi Qianwen/Qwen), ByteDance (Doubao), Tencent (Yuanbao), and Baidu (Ernie) engaged aggressively by offering cash rewards, consumption vouchers, and sponsoring televised New Year galas to attract users. Morgan Stanley estimates the total promotional expenditure across these platforms likely exceeded 8 billion RMB, covering cash rewards, per-order subsidies, ecosystem traffic redirection, and media exposure.

However, the outcomes varied significantly after the promotional frenzy. Utilizing Questmobile data, Morgan Stanley systematically evaluated the platforms based on three metrics: traffic scale, user engagement, and post-campaign user retention. The findings were stark: while all participants achieved short-term daily active user (DAU) surges, only Doubao and Tongyi Qianwen demonstrated a real ability to retain users. The report concludes that long-term success hinges on product utility, ecosystem integration, and creating genuine value beyond monetary incentives.

Morgan Stanley characterized this event as a strategic battle for the next "super gateway" in consumer AI. The platform that can maintain its user base after subsidies end is positioned to lead the subsequent phase of AI commercialization. Current data indicates ByteDance's Doubao and Alibaba's Tongyi Qianwen performed relatively well, while Tencent's Yuanbao showed concerning retention metrics, and Baidu was largely absent from the competition.

**Spending Scale: Alibaba Led with Over 80 Billion RMB Total** Spending intensity differed markedly among the companies:

Alibaba (Tongyi Qianwen) officially announced a "3 Billion RMB Spring Festival Treat Plan," centered on no-threshold 25 RMB vouchers applicable to food delivery, instant retail, movie tickets, flights, and hotels. This campaign was deeply integrated with platforms like Taobao, Hema, Tmall Supermarket, Fliggy, Damai, and Alipay. The voucher validity period was extended multiple times. Morgan Stanley estimates actual spending likely surpassed 5 billion RMB, based on approximately 200 million orders redeemed with the vouchers plus additional brand marketing costs.

ByteDance (Doubao) secured extensive media exposure through a deep partnership with China Central Television's Spring Festival Gala. The company concurrently ran a cash reward lottery and offered 100,000 high-value tech prizes. Morgan Stanley estimates total expenditure significantly exceeded 1.5 billion RMB, based on the prize value alone. The campaign period also saw the launch of Doubao 2.0 and Seedance 2.0.

Tencent (Yuanbao) deployed a 1 billion RMB cash reward program distributed via its app and chat groups, incorporating tasks like AI image generation. Notably, direct links to the reward activity were temporarily blocked on WeChat before being reinstated in an alternate format. The marketing emphasized the capabilities of the DeepSeek model.

Baidu (Ernie) allocated 500 million RMB for cash rewards and served as the premier AI partner for Beijing TV's Spring Festival Gala.

**Traffic Surge: Doubao > Tongyi Qianwen > Yuanbao** All three primary contenders achieved significant DAU growth, albeit at different scales:

Doubao saw its DAU surge from a pre-campaign 84 million to a peak of 144.5 million on New Year's Eve, a net increase of approximately 60.5 million. Its weekly active user (WAU) count also peaked at 253 million during the holiday week.

Tongyi Qianwen's DAU jumped from a baseline of around 17 million to a peak of 73.5 million, a net increase of about 56.5 million. Its WAU also showed substantial weekly increases during the campaign period.

Yuanbao experienced a rise from a January 2026 baseline of 7 million DAU to a peak of 40.5 million, a net increase of 33.5 million. Its official figures reported a peak DAU exceeding 50 million and a monthly active user (MAU) count of 114 million.

Baidu (Ernie) lagged significantly. Despite promotional efforts, WAU and DAU for the Baidu App continued a weekly decline. The Ernie App's WAU peaked at just 3.4 million, and its DAU at 826,000, figures far below its competitors.

**User Engagement: Doubao > Yuanbao > Tongyi Qianwen** Rapid traffic expansion generally compressed average user session time, but the degree of compression varied, indicating differences in user quality.

Tongyi Qianwen saw the most dramatic drop, with average daily usage time per user plummeting by 51% to between 3 and 5 minutes during the peak period. This suggests its traffic was heavily concentrated on transactional behaviors like redeeming vouchers, representing a "low-engagement, transactional" usage pattern.

Yuanbao's average usage time decreased by approximately 32% compared to January 2026 levels, averaging around 6.5 minutes per week, indicating relatively low engagement from new users.

Doubao demonstrated the strongest resilience, with average usage time declining only moderately by about 15% to 8.7 minutes per day during the peak week, maintaining robust user stickiness despite the massive DAU expansion.

**Post-Campaign Retention: Tongyi Qianwen > Doubao > Yuanbao** Following the campaign peak on February 16th, traffic receded across all platforms, but the extent of the decline revealed fundamental differences in retention capability.

Yuanbao exhibited the weakest retention, with its DAU quickly falling back to near pre-campaign levels.

Doubao's user base receded from its peak but settled at a level structurally higher than before the campaign, indicating the acquisition of a cohort of genuine users.

Tongyi Qianwen's post-campaign DAU remained significantly above its pre-campaign baseline, showing the most impressive retention. However, Morgan Stanley noted that this resilience was partly supported by the extended validity of its consumption vouchers until March 3rd, suggesting the data may be artificially inflated and requires further observation after the voucher period concludes.

**Order and Ecosystem Data: Alibaba's AI-Ecommerce Experiment** Tongyi Qianwen's order data during the period was particularly striking, highlighting the advantage of Alibaba's ecosystem integration. The platform facilitated approximately 200 million orders, including over 102 million beverage orders. Significant growth was also seen in flight, hotel, and attraction ticket bookings. Notably, over 400,000 users aged 60 and above placed orders through the app, nearly 50% of orders came from county and rural areas, and over 130 million consumers used the Tongyi Qianwen App for AI-assisted shopping for the first time.

**The Super Gateway Contest: Outcome Unclear but Patterns Emerge** Morgan Stanley's core assessment is that this battle marks the official start of the contest for the next consumer AI "super gateway," but short-term traffic spikes do not guarantee long-term victory. A composite evaluation across the three dimensions reveals: - Traffic Scale: Doubao > Tongyi Qianwen > Yuanbao - User Engagement: Doubao > Yuanbao > Tongyi Qianwen - Post-Campaign Retention: Tongyi Qianwen > Doubao > Yuanbao

Doubao delivered the most balanced and outstanding overall performance, leveraging its top-tier Gala exposure. Tongyi Qianwen achieved impressive order volume through deep ecosystem integration, though user engagement quality needs improvement. Yuanbao's retention is a concern, suggesting its 1 billion RMB reward campaign primarily yielded short-term exposure. Baidu fell significantly behind, with its Ernie App's metrics not on the same scale as the others.

Morgan Stanley emphasizes that the key to long-term success lies in retaining users through product utility post-campaign, achieving seamless ecosystem integration, and creating authentic value beyond financial incentives. For investors, this event serves as a critical window for observing the user acquisition efficiency and retention capabilities of major AI platforms, with subsequent data trends warranting close monitoring.

Disclaimer: Investing carries risk. This is not financial advice. The above content should not be regarded as an offer, recommendation, or solicitation on acquiring or disposing of any financial products, any associated discussions, comments, or posts by author or other users should not be considered as such either. It is solely for general information purpose only, which does not consider your own investment objectives, financial situations or needs. TTM assumes no responsibility or warranty for the accuracy and completeness of the information, investors should do their own research and may seek professional advice before investing.

Comments

We need your insight to fill this gap
Leave a comment