FULL STORY

Sakana AI Launches RSI Lab for Embodied AI

Sakana AI has launched its Recursive Self-Improvement (RSI) Lab in Tokyo to automate AI development. Led by CEO David Ha, the lab will also focus on advancing physical AI and embodied intelligence.

2026-08-04 ~ 2026-08-12 · 2 episodes · 8 posts

Episode 1 · Sakana AI Launches RSI Lab and Joins Japan's AIRoA (2026-08-04, 2 posts)

Sakana AI launched RSI Lab to focus on recursive self-improvement technology and joined the AI Robot Association (AIRoA) as a regular member to advance world models and physical AI.

Episode 2 · Sakana AI Opens RSI Lab in Tokyo, Ventures into Physical AI (2026-08-10, 6 posts)

Sakana AI announced the establishment and expansion of its Recursive Self-Improvement (RSI) lab in Tokyo, led by CEO David Ha. The lab's core goal is to transform AI development from static, human-led processes into autonomous self-improving intelligent engines, with plans to bring RSI principles into physical AI and embodied intelligence. The lab is currently actively recruiting full-time staff and interns.

Confirmed

  • The lab is located in Tokyo, led by Sakana AI CEO David Ha, and is currently hiring full-time employees and interns.
  • Research scope includes frontier language models, world models that learn the dynamics of the real physical world, and recursive self-improvement loops that combine the two.
  • The core technical path is building world models that enable agents to recursively improve their abilities by simulating, planning, and executing tasks in the real world.

Why it matters

  • Combining recursive self-improvement systems with physical world models means AI is not just iterating at the software level, but attempting autonomous evolution in real physical environments. According to official statements, this could be the next major revolution in physical AI and embodied intelligence.