Sakana AI Raises $135M at $2.65B Valuation in Series B

Sakana AI secures ¥20B Series B at a $2.65B valuation, advancing Japan-focused model development and expanding projects with MUFG, Daiwa and other domestic groups.

Nov 17, 2025 - 08:01
Nov 17, 2025 - 08:02
 20
Sakana AI Raises $135M at $2.65B Valuation in Series B
Sakana AI Raises $135M at $2.65B Valuation in Series B

Key Points

 Sakana AI Raised ¥20 billion (≈ $135M) in Series B funding.
Valuation set at approximately $2.65 billion.
MUFG, Khosla Ventures and In-Q-Tel participated in the round.
Funds directed toward model development and hiring in Japan.
Plans for 2026 deployments in finance, manufacturing and government.

Sakana AI has completed a ¥20 billion (around $135 million) Series B funding round, valuing the company at approximately $2.65 billion. CEO David Ha confirmed the figures, noting that the round reflects continued interest in Japan-based AI research at a time when the global industry remains dominated by U.S. companies such as OpenAI, Google and Anthropic. The round comes one year after the company’s previous fundraise and further establishes Sakana as one of Japan’s most heavily funded AI startups.

Company Background

The firm was set up in 2023 by former Google researchers Llion Jones, Ren Ito and David Ha. The three founders built the team with a clear goal: to create language models that handle Japanese sentence structure, tone and formal writing conventions with greater accuracy than foreign systems. Their approach uses smaller, well-curated datasets and model architectures built to run efficiently on limited compute, making the technology more accessible to Japanese enterprises that cannot deploy extremely large models.

Research Approach

Sakana focuses heavily on refining its models after the initial training stage. Much of the work involves adjusting how the system interprets vocabulary, honorifics, writing style and contextual cues unique to Japanese. This process allows the company to release versions adapted for business documentation, customer communication and internal workflows without needing to train massive new models each time.
This methodology also separates Sakana from U.S. model developers that prioritize scale above specialization.

Investor Participation

The funding round includes participation from Mitsubishi UFJ Financial Group (MUFG), one of Japan’s largest financial institutions. MUFG has been exploring domestic AI infrastructure to support internal automation projects.
International investors joined as well, including Khosla Ventures, Macquarie Capital, NEA, Lux Capital and In-Q-Tel (IQT). Their involvement indicates continued global interest in region-specific model development, particularly in countries where data regulations and language characteristics require localized solutions.

Use of Funds

The company will increase hiring across model development, evaluation, engineering and customer deployment teams.
Sakana is also expanding its commercial operations in Japan, where it has already worked with groups such as MUFG and Daiwa to design AI systems for document processing, internal communication review and structured data analysis.
Additional funds will be used to build internal datasets, strengthen testing pipelines and release updated versions of its Japan-tuned models.

Sector Expansion

The firm plans to widen its enterprise offerings in 2026 to include industrial automation, factory operations and public-sector environments. These sectors require precise language handling and strict data-handling controls—areas where Sakana believes its locally developed models can be used without major adaptation.
The company is also examining opportunities in national security-related work, including information analysis and specialized model deployment, though no formal programs have been announced.

Strategic Direction

Sakana is open to collaborations with Japanese corporations and may consider acquisitions if they support its model research or enterprise deployments. The company is also studying potential overseas markets where demand for national-language AI systems is increasing, particularly in regions with strict data laws or complex local writing systems.

Sovereign AI Positioning

Japanese companies, especially in finance, manufacturing and regulated industries, have shown rising interest in models built within Japan’s regulatory environment. Data residency requirements, linguistic accuracy and internal risk policies have made domestic model providers more appealing for organizations that previously depended on U.S. systems.
Sakana aims to supply models that meet these requirements without the extensive modifications typically needed when working with foreign LLMs.

Previous Funding

The company’s earlier Series A, completed about a year ago, raised close to ¥30 billion (roughly $214 million) and valued Sakana at about $1.5 billion. PitchBook data shows that Sakana had raised around $244 million by March 2025. With the new Series B, its total funding now stands near $379 million, placing it among the most well-capitalized AI developers in the Asia-Pacific region.

Also Read: Why Top YouTubers No Longer Depend on Ad Revenue — and How They are Building Real Businesses

iShook Opinion Curated by iShook Opinion and guided by Founder and CEO Beni E Rachmanov. Dive into valuable financial insights at ishookfinance.com for expert articles and latest news on finance.