Startup profile · funding coverage
Sakana AI funding, valuation and investors
Tokyo AI research company developing nature-inspired model optimization, collective intelligence and resource-efficient AI.
Verified company research · funding, investors and latest developments
Keep track of Sakana AI
Save this profile to your VCT watchlist for a quick return.
Funding, valuation & investors
Answer-first snapshotLatest funding
Strategic investment
Undisclosed · January 23, 2026
Latest known valuation
¥432B post-money (~$2.7B at announcement; company Japanese release)
Series B · November 17, 2025
Total disclosed equity funding
Not publicly disclosed
Excludes debt, grants, acquisitions, secondaries, and IPO proceeds.
Current status
private
series b · Tokyo, Japan
Investors in latest funding
Other: Google
Overview
Sakana AI, founded in Tokyo in July 2023, researches nature-inspired and compute-efficient approaches to AI, including evolutionary model merging, automated scientific discovery and model optimization for Japanese enterprises. Its September 2024 Series A was approximately $200M, led by NEA, Khosla Ventures and Lux Capital. A November 2025 Series B raised ¥32B (approximately $200M at announcement).
Google announced a separate strategic equity investment in January 2026 without disclosing the amount.
Why Sakana AI is interesting
Sakana demonstrates a funded alternative to brute-force pretraining: model evolution, model merging and agent-driven AI research. Its $200M-equivalent 2024 Series A and ¥32B 2025 Series B are separate company financings, not NCM investments.
Product & use cases
Research platform combining evolutionary model optimization, agent-based AI science, Japanese sovereign AI customization and industrial deployments.
- Improving existing models without full frontier-model pretraining
- Industrial, finance and sovereign AI applications in Japan
- Automated research and multi-model reasoning
Key facts
- Company-announced September 2024 Series A: approximately US$200M
- Company-announced November 2025 Series B: ¥32B (~US$200M, not a USD-denominated round)
- January 2026 strategic Google investment amount not disclosed
- Not a direct NCM developer
Funding history (newest first)
Series B
2025-11-17 ¥32B (~$200M at USDJPY 160) Valuation: ¥432B post-money (~$2.7B at announcement; company Japanese release)- Mitsubishi UFJ Financial Group (participant)
- Khosla Ventures (participant)
- New Enterprise Associates (participant)
- Lux Capital (participant)
- Factorial Funds (participant)
- Macquarie Capital (participant)
- Mouro Capital (participant)
- Mitsubishi Electric (participant)
- Salesforce Ventures (participant)
- Datadog Ventures (participant)
- Citi (participant)
- In-Q-Tel (participant)
Source: Read source · sakana.ai
Series A
2024-09-04 Approximately $200M- New Enterprise Associates (lead)
- Khosla Ventures (lead)
- Lux Capital (lead)
- Translink Capital (participant)
- 500 Global (participant)
- NVIDIA (participant)
- Mitsubishi UFJ Financial Group (participant)
- Sumitomo Mitsui Banking Corporation (participant)
- Mizuho Financial Group (participant)
- NEC (participant)
- SBI Group (participant)
- JAFCO (participant)
- Global Brain (participant)
- Miyako Capital (participant)
Source: Read source · sakana.ai
Competitive landscape
Edge: Nature-inspired model optimization research plus Japanese strategic investors and enterprise partnerships; unlike NCM, not a text-generation hierarchy model.
Sakana's ecosystem and experimental techniques compete with brute-force model development. Its business depends on converting R&D into repeatable enterprise contracts. Funding is company-wide and should not be attributed to a single technique.
Market / IPO context
Editorial / static context — not a live quote.
FAQs about Sakana AI
Practical answers founders, operators, and investors typically search for.
Last updated:
Editorial note: AI tools assisted with research, structure, or drafting. Venture Capital Tracker retains human editorial responsibility for factual accuracy, relevance, and source quality before publication.