Sakana AI Secures $100M Funding to Compete with OpenAI and Anthropic as a Leading AI Lab

The AI frenzy is sweeping the globe. Following China’s Alibaba announcement of Qwen2-VL, Sakana AI, a Japanese startup founded by former Google researchers David Ha and Llion Jones, along with ex-diplomat Ren Ito, has successfully raised $100 million in a Series A funding round.

This investment, led by industry giants like New Enterprise Associates, Khosla Ventures, and Lux Capital, with participation from Nvidia, highlights the impressive traction of the year-old company. Sakana AI plans to utilize this capital and Nvidia’s infrastructure support to advance its technologies and evolve into a world-class AI research lab, competing with frontrunners like OpenAI and Anthropic.

In an announcement on X, Ha emphasized that the company is operating at an accelerated pace compared to most global AI labs and has ambitions to “push the frontiers of what’s possible with AI” in the near future.

What is Sakana AI Doing?

Sakana AI made waves last year with its innovative nature-inspired collective intelligence approach to developing high-performing foundation models. This strategy involves integrating various smaller AI models, akin to a swarm, to achieve complex outcomes.

Leveraging this unique method and native Japanese datasets, Sakana AI created models capable of generating traditional ukiyo-e artwork. Recently, the company introduced the AI Scientist, an LLM-based system that automates the entire research cycle from ideation and coding to running experiments and conducting peer reviews.

Sakana AI aims to scale its nature-inspired AI development. With capital from leading venture firms and Nvidia, the company plans to accelerate hiring to build a talent-rich AI research organization while expanding its infrastructure through collaboration with Nvidia.

In a blog post, Sakana confirmed that Nvidia will provide infrastructure support in two ways: delivering cutting-edge GPU systems for advanced model development and granting access to Nvidia-powered data centers in Japan for experimental purposes. Additionally, Nvidia will assist in organizing community initiatives like AI hackathons and university outreach programs.

The ultimate goal is to establish a premier AI lab in Japan that produces advanced, energy-efficient AI technologies to tackle 21st-century challenges, such as declining population, reduced competitiveness, and geopolitical tensions.

“This daunting task will require years of R&D and building relationships with key stakeholders. We believe our technology can help Japan regain its technological edge in AI and boost its global competitiveness. Our aim is to assist Japan, its institutions, and its citizens in overcoming their AI challenges,” the company stated.

Competitors Eye Japan’s Growth

As Sakana AI gains momentum with its distinct approach, competitors are also making strides in Japan. OpenAI, recognized as a leader in the field, recently opened a Tokyo hub and launched a custom GPT-4 model optimized for the Japanese language. OpenAI Japan’s president, Tadao Nagasaki, has hinted at a new AI model, GPT-Next, claiming it will be 100 times better than its predecessor.

Additionally, Canada-based enterprise AI startup Cohere is developing a custom Japanese model in collaboration with Fujitsu.

Despite the competitive landscape, investor confidence in Sakana remains strong. Vinod Khosla, founder of Khosla Ventures, highlighted that while many labs globally are trying to catch up using the same foundation model training techniques, Sakana AI is paving the way for true innovation.