Ex-Qwen lead's Pragmatik Labs raises $220M at a $2B valuation
Lin Junyang, the researcher who built Alibaba's Qwen into one of the world's most-used open model families, came out of stealth early Wednesday with a company that has no product, no demo, and a $2 billion price tag. That gap is the story.
Lin Junyang, the former technical lead of Alibaba's Qwen family, announced the founding of Pragmatik Labs (语用科技) in Shanghai — backed by a $220 million angel round at a $2 billion valuation, closed before the company shipped anything. Gaorong and Sequoia China (HSG) co-led the round with roughly $100 million each; Tencent put in about $20 million, and the Shanghai Future Industry Fund participated. Lin, who stepped down from Qwen in March, retains about 88% of the company through his holding entities. The startup's stated focus is "next-generation intelligence spanning the digital and physical worlds," split into two tracks: Digital Agents for knowledge work and Physical Agents for real-world environments. Its website currently lists research directions — reasoning, tool use, long-horizon task execution — rather than a product.
The valuation is a bet on the man and the thesis, not the output. Lin joined Alibaba DAMO Academy in 2019, took over the Qwen line in late 2022, and became one of Alibaba's youngest P10 technical leads; by January 2026, Qwen derivative models had passed 200,000 with over a billion downloads globally. His March essay "From 'Reasoning' Thinking to 'Agentic' Thinking" laid out the shift he's now betting his career on: models that merely answer questions are being overtaken by systems that act, call tools, and adjust from feedback. Notably, he is not building a bigger base model — the deliberate move away from the pretraining arms race is itself the signal.
China's capital markets are paying attention: the same week Manus announced it would resume independent operations after Meta's acquisition was blocked by regulators, the country's most prominent model lead chose agents over scale. Lin joins a pattern set by Mira Murati's Thinking Machines Lab, which raised $2 billion at a $12 billion valuation before product launch — for a handful of proven technical leaders, track record now prices ahead of proof. The risk is equally visible: Pragmatik will have to show an agent system that works in the physical world, a far harder bar than a benchmark leaderboard, and investors who paid angel-round prices for a $2 billion company will be watching the first demo closely.
What to watch: whether Pragmatik ships anything concrete before its next raise — Lin was reportedly already fielding interest in a follow-on when the angel round closed.
Is a $2 billion valuation with no product the market pricing talent correctly, or the first sign of a froth cycle? Tell us in the comments.
Sources: Tencent News 深网 · Phoenix Tech (ifeng) · Sina Finance · Lianhe Zaobao
