ByteDance, Alibaba, and DeepSeek Are All Launching AI Models in February

Three of China’s biggest AI players are preparing to launch new models in February, timed to the Lunar New Year holiday. ByteDance, Alibaba, and DeepSeek are all reportedly ready to ship, turning what was once a quiet festival period into the opening round of 2026’s AI arms race.

The lineup is aggressive. ByteDance is releasing three products at once. Alibaba is going after the consumer market. DeepSeek is teasing a next-generation architecture that could reset benchmarks.

What Each Company Is Shipping

ByteDance plans to launch Doubao 2.0, the next version of its flagship large language model. Doubao already has 163 million monthly active users, making it China’s largest AI application by user count. Alongside it, ByteDance is releasing Seeddream 5.0 for image generation and Seeddance 2.0 for video generation. This is a full multimodal push.

Alibaba is launching Qwen 3.5, its next-generation model optimized for mathematical reasoning and code generation. Alibaba is also rolling out large-scale marketing campaigns for Qwen’s consumer-facing chatbot, directly targeting ByteDance’s Doubao in the consumer AI market.

DeepSeek has been quieter, but GitHub repository updates revealed a new architecture identifier called MODEL1, widely seen as the foundation for DeepSeek V4. Sources say the model could drop as early as mid-February. DeepSeek’s last major release sent shockwaves through global markets – the kind of disruption that rattled crypto alongside it.

The Lunar New Year Strategy

The timing is not accidental. Tencent announced it will distribute 1 billion yuan ($140 million) in cash through its Yuanbao AI chatbot during the holiday, copying the “red envelope” campaigns that made WeChat Pay dominant. ByteDance and Baidu are running similar AI promotions.

This is a user acquisition land grab. Chinese tech companies are using the holiday the way American companies use the Super Bowl – a cultural moment to capture mass attention and convert it into daily AI app usage.

The “Months Behind” Question

At Davos in January, Google DeepMind CEO Demis Hassabis said Chinese AI firms are now just “months” behind Western frontier models – down from the two-to-three year gap estimated as recently as 2024. Bloomberg reported the number at approximately six months.

But Hassabis drew a distinction between copying and inventing. “They’ve shown they can catch up and be very close to the frontier,” he said. “But can they actually innovate something new, like a new transformer, that gets beyond the frontier? I don’t think that’s been shown yet.”

Nvidia CEO Jensen Huang offered a blunter assessment at CES: “China is well ahead of us on energy. We are way ahead on chips. They’re right there on infrastructure. They’re right there on AI models.”

Laterstack exists to sharpen critical thinking by connecting tech, policy, and power to everyday life – across class, industry, and influence. The “months behind” framing from Hassabis is already outdated by the time it reaches Western audiences. Three major Chinese firms are releasing next-generation models simultaneously, backed by consumer platforms with hundreds of millions of users. American policymakers still debating export controls need to understand that the gap is not closing – it is being closed deliberately, at scale, with state coordination Western labs cannot match. The question is no longer whether China can compete in AI. It is whether the West’s lead was ever as large as its leaders claimed.

What This Means for Everyday People

The AI app war is now global. If you use ChatGPT, Gemini, or Claude, you are using products that are being benchmarked against Chinese competitors most Americans have never heard of. That competition drives faster releases, lower prices, and more aggressive data collection on all sides.

For investors, the February launches could move markets. DeepSeek’s last major release triggered a $1 trillion sell-off in U.S. tech stocks. If DeepSeek V4 matches or beats frontier Western models again, expect similar volatility.