As U.S. giants like Google, OpenAI and Anthropic push ahead in the race to build massive language models, emerging startups such as Sakana AI, Mistral AI, DeepSeek and AI21 Labs are carving out distinct niches with specialized, region-specific AI systems.
Tokyo-based Sakana AI has now raised a ¥20 billion (about $135 million) Series B round, valuing the company at $2.65 billion post-money, up from a $2.5 billion pre-money valuation, CEO David Ha told TechCrunch.
Founded in 2023 by former Google researchers Llion Jones, Ren Ito, and David Ha, Sakana AI builds affordable generative AI models optimized for Japanese language, culture, and small-dataset performance.
The new funding came from a mix of global and Japanese investors, including MUFG, Khosla Ventures, Macquarie Capital, NEA, Lux Capital, and In-Q-Tel.
Ha says the funding will support R&D, model development, and team expansion across engineering, sales, and distribution. The company has been deepening its Japanese presence this year, partnering with major institutions such as Daiwa and MUFG while building AI solutions for local enterprises.
Sakana also plans to expand into industrial, manufacturing, government, defense, and intelligence sectors by 2026, broadening its enterprise footprint beyond finance.
The startup says global demand for sovereign AI is rising, systems that reflect national culture, values, and language. Sakana believes its strategy of focusing on post-training product development allows it to efficiently tailor advanced AI models to Japan, distinguishing it from companies building massive general-purpose LLMs.
Related: OpenAI Launches Group Chats For ChatGPT
This Series B comes roughly a year after Sakana’s ¥30 billion ($214M) Series A, which valued the company at $1.5 billion. With this new raise, Sakana has now secured about $379 million in total funding.
