The news
General Intuition, a startup building AI world models, has raised $220 million at a $6.2 billion valuation, SiliconANGLE reported on September 29, 2026. Investors include Valor Equity Partners, Atreides, Seven Seven Six, Point72, Khosla Ventures and General Catalyst.
A world model is an AI system that learns how an environment looks and changes over time, so it can predict what happens next or generate realistic simulated footage. Such models are seen as a foundation for robots and other systems that act in the physical world.
General Intuition spun out of Medal B.V., which makes a free tool for capturing and sharing video game footage, about a year before this round, according to SiliconANGLE. Pulse 2.0 reported that the company describes its work as foundation models focused on actions, physical environments, robotics and world modeling, trained on action-labeled video. It said Medal is on track to reach about 3 billion uploaded videos a year, and claimed that leading robotics models train on less than 1% of the action data it can access.
In June 2026 the company released MIRA, a world-model algorithm that generates synthetic video for training AI systems, SiliconANGLE reported. The outlet said MIRA has 5.6 billion parameters, uses a latent diffusion approach and produces 20 frames per second at 720x576 resolution on a single Nvidia (NVDA) B200 graphics processor. So far it has been demonstrated on a single video game. The company called MIRA a stepping stone to physical AI.
General Intuition has run limited tests with customers in robotics, simulation and entertainment and has opened a commercial waitlist, SiliconANGLE reported. The company plans to spend the money on training, computing capacity and hiring researchers, and to grow teams in New York and Europe, according to Pulse 2.0.
The numbers
- New funding
- $220 million
- Valuation
- $6.2 billion
- MIRA model size
- 5.6 billion parameters
- MIRA output
- 20 fps at 720x576 on one B200 GPU
- Medal uploads (company estimate)
- About 3 billion videos a year
Why CEOs should care
For leaders in manufacturing, logistics and robotics, world models matter because training robots in the real world is slow and expensive. Simulated environments that look and behave realistically could cut that cost. Ask robotics vendors what simulation data their systems train on and how well results carry over from simulation to real sites.
For investors and boards, the round shows capital following data advantages. General Intuition's case rests on access to Medal's gaming video, a supply of human action data that is hard to copy. Its own demos so far cover a single game, so the gap between the valuation and proven commercial use is wide. Diligence should focus on how the data transfers to physical tasks and on licensing rights to that footage.
For CTOs, the practical step is to watch the commercial waitlist and early pilots. Companies with simulation needs, such as training, testing autonomous systems or game development, may be able to test these models before broad availability.
The bigger picture
General Intuition is part of a wave of well-funded startups and big labs pursuing physical AI, the effort to move AI beyond text and images into systems that act in the world. Large rounds at multibillion-dollar valuations for companies at an early commercial stage show how much investors are willing to pay for a possible foundation layer.
The use of gaming data is notable. Games offer large volumes of human decisions in fast-changing settings, which Pulse 2.0 reported the company sees as useful training material.
What’s next
Watch for General Intuition to move customers off the waitlist, for MIRA results beyond a single game, and for partnerships with robotics or simulation companies that would test the claims behind the valuation.
What “Fact-checked” means
Fact-checking means testing a story’s facts against the evidence before it is published. This story went through at least two separate checks before this version was published.
- What we checked
- Its names, figures, dates, job titles, quotes and who said what were checked against the story’s sources, including its main source where it could be opened. The headline was checked for accuracy and overstatement.
- How
- A first check reviewed the whole story. If it passed, a second, skeptical check went back to the sources to look for mistakes in the most important facts. If a check flagged the story, it was edited to fix the problems found, and a separate re-check then reviewed the whole story again.
- Who
- The checks are made by our newsroom, as steps kept separate from the writing, under rules set by our editor, Hussein Mukhtar. A story the checks still flag is held for the editor, who decides whether it is fixed, published or dropped.
- If something is wrong
- “Fact-checked” does not mean error-free. If a material error is found after publication, we correct the story and add a note saying what changed. Report an error







