Sources: Sakana AI, a Tokyo-based LLM developer founded last year by Google alumni, is raising ~$100M at a $1B valuation in a round led by NEA, Khosla, and Lux
The Information :
Context & Ripple Effects
Sakana AI was launched in Tokyo by Google research alumni, including a co-author of the transformer paper, and had already raised a $30M seed round focused on smaller AI models. This reported round would sharply raise the capital available to a young lab pursuing that approach.
The same investors later backed Sakana's $100M Series A alongside Nvidia, tying this report to a broader investor commitment rather than a one-off financing discussion.
First-order effects
- If completed, the roughly $100M round would give Sakana AI substantially more runway for model development and commercialization while valuing the company at about $1B.
- NEA, Khosla and Lux would become the lead financial sponsors of a Tokyo-based LLM developer founded by former Google researchers.
Second-order effects
- A $1B valuation this early would raise the financing benchmark for other Japan-focused LLM developers, increasing pressure to demonstrate differentiated models, language coverage or capital efficiency.
- The round would also widen Sakana's ability to compete for AI researchers and compute access against better-funded global labs, even as its smaller-model focus may offer a different operating profile.
Third-order effects
- If follow-on financings sustain these valuations, Japan's AI ecosystem could develop a more distinct set of well-capitalized domestic labs rather than relying solely on foreign model providers.
- The pattern points to venture capital concentrating behind a limited number of technically credible frontier-model teams; whether that produces durable local AI champions will depend on commercialization, not fundraising alone.
The trend: This is one data point in the globalization of frontier-AI financing, as major US venture firms fund locally positioned model labs outside the traditional US hub.