Sakana AI
A cutting-edge AI lab founded by former Google researchers in Tokyo, focusing on bionic, collective intelligence, and innovative research methods.
Visit Website ↗What is Sakana AI
Sakana AI is a research and development company based in Tokyo, founded in 2023 by David Ha (CEO), Llion Jones (CTO, co-author of the Transformer paper), and Ren Ito (COO). The company's name is derived from the Japanese word for "fish" (さかな), inspired by the collective behavior of fish schools, which represents the concept of "collective intelligence" that the company advocates for. In 2025, Sakana AI completed a Series B funding round of approximately 13.5 billion yen, with a post-investment valuation of $2.65 billion, making it a flagship of Japan's cutting-edge AI research.
Unlike many AI companies that focus on developing a single large model, Sakana AI takes a different approach, exploring bionic, evolutionary, and collective intelligence research directions, aiming to combine models in a more efficient and intelligent way. For researchers and technical decision-makers, Sakana AI represents a force in "Japan's homegrown cutting-edge basic research", rather than just another applied technology.
Features and Use Cases
Sakana AI's most notable product is AI Scientist, which claims to automate the entire scientific research lifecycle: from generating innovative research ideas, writing necessary code, conducting experiments, to summarizing results, visualizing data, and producing a complete draft of a scientific paper. This has sparked significant discussion in the academic community, demonstrating that AI can not only assist research but also potentially complete a research cycle on its own.
Who is it suitable for? Scholars and researchers interested in AI-automated scientific research, those who want to understand cutting-edge research methods, and engineers and decision-makers who care about the technical route of "progress without relying on massive computing power". It's essential to note that Sakana AI is a research laboratory rather than an out-of-the-box commercial SaaS, and its outcomes are often released in the form of papers, open-source projects, and models, making it more suitable for research, tracking, and experimentation rather than immediate productivity tools.
Key Features
- Founded by former Google researchers, including a co-author of the Transformer paper
- Focuses on bionic, evolutionary, and collective intelligence research directions
- AI Scientist can automate the entire scientific research lifecycle
- Outcomes are often released in the form of papers, open-source projects, and models
- A Japanese cutting-edge AI laboratory with a valuation of $2.65 billion
Pros
- Takes a smart and efficient approach, rather than simply scaling up
- The team has a strong academic background, and research outcomes have significant impact
- Often releases open-source projects, making it convenient for researchers to track and experiment
Cons
- Primarily a research laboratory, not an out-of-the-box commercial SaaS
- Outcomes are mainly in the form of papers and models, which may be difficult for general users to apply directly
- AI-automated scientific research is still in its early stages, and practical maturity is limited
Use Cases
- Tracking the latest developments in AI-automated scientific research
- Researching bionic and collective intelligence AI methods
- Using AI Scientist to explore automated research workflows
- Following Japan's cutting-edge AI technology trends
Editor's Note
A flagship of Japan's cutting-edge AI research, taking a distinctive approach with bionic and efficient methods in a field where many companies focus on scaling up. While it's not a ready-to-use tool, AI Scientist's vision for automated research is worth tracking. We give it a rating of 4.3.
FAQ
Can AI Scientist really conduct research on its own?
It can automate the entire research lifecycle, from generating ideas to producing a paper draft, demonstrating the possibility of AI-automated research, but it's still in its early stages and requires human quality control.
Is Sakana AI a tool for general users?
No, it's primarily a research laboratory, and its outcomes are often released in the form of papers, open-source projects, and models, making it more suitable for researchers and engineers to track and experiment with.