Pragmatik Labs, the Shanghai AI startup founded by the former head of Alibaba's Qwen language-model team, has raised fresh funding at a $5 billion valuation, according to people familiar with the matter cited by Bloomberg. The new round, which raised tens of millions of dollars, more than doubles the $2 billion valuation the company reached in its first fundraiser in August — just three months ago.
New investors in the round include HyT Capital, CDH Investments and automaker Geely Automobile Holdings, the report said. The company, which is still in stealth mode, has released no product, model or research paper since it was founded.
A bet on the founder, not a product
Pragmatik Labs was set up over the summer by Junyang Lin, 33, who left Alibaba in March after leading the technical development of its Qwen models. Lin announced the venture on social media in August, saying it would focus on next-generation AI agents that can operate across both the digital and physical worlds — a bet that AI is moving from training models to training agents that can act on their own.
Investors have so far been buying the founder rather than a product. The company's first round, announced in August, raised around $220 million at a $2 billion post-money valuation, co-led by Gaorong Ventures and HongShan, the firm formerly known as Sequoia China. Tencent invested about $20 million, and the Shanghai Future Industry Fund, a government-backed fund, also took part. Lin and related entities hold roughly 88% of the company, with outside investors holding about 12%.
Lin was one of Alibaba's youngest senior engineers, reaching its top P10 technical rank while leading Qwen from a little-known project to one of the world's most widely used open-source model families. That track record is what investors are paying for: a research lab led by someone who has already shipped frontier models at scale.
The AI talent war, priced in
The jump from $2 billion to $5 billion in a single quarter captures how aggressively capital is chasing proven AI talent in China. Frontier-model engineers who can build and run large systems are scarce, and venture firms have been willing to pay steep prices for teams led by them — even before a product exists.
The trend is not unique to Pragmatik. Across China's AI sector, startups led by well-known researchers have raised large rounds at valuations that would have seemed implausible a few years ago, as investors race to back the teams they believe will define the next generation of AI agents.
Still, the risks are plain. Pragmatik has no revenue, no customers and no public technical work to judge, and its valuation now rests entirely on Lin's reputation and the belief that agentic AI will be the next big platform shift. The company has not commented publicly on the new valuation, and the figures come from people familiar with the deal rather than an official announcement.
What comes next
For now, the money buys time and talent. Pragmatik plans to use the funding to build out its research lab and design specialised model architectures, as it prepares for the work of turning its agent vision into something demonstrable. Whether a $5 billion valuation can be justified will depend on what the lab produces in the coming year — and whether investors' faith in the founder-led model survives contact with reality.
Either way, the round is a marker of how the AI race is being fought in 2026: not just with compute and data, but with cheques written to the engineers the market believes can build what comes next.



