Valor, Point72 back General Intuition at $6B valuation as AI startup pushes into robotics

2026-08-25·8 min read

On August 24, 2026, TechCrunch reported that New York-based startup General Intuition is in talks to raise funding at a $6 billion pre-money valuation from new investors including Valor Equity Partners, known for backing SpaceX, and hedge fund giant Point72, with existing shareholders Khosla Ventures and General Catalyst participating. The round comes just weeks after General Intuition raised $320 million at a $2.3 billion valuation. General Intuition is building a foundation model that trains generalized AI agents to act through space and time — one of the hottest projects in the 'Physical AI' race.

General Intuition's uniqueness lies in its data approach. CEO Pim de Witte spun the company out of Medal, a video game clip-sharing platform, last October, bringing with him a 'data goldmine' almost no one can replicate: hundreds of millions of hours of gameplay recordings, plus the accompanying 'action labels' — precise records of which button a player pressed and when. This data forms the initial dataset for General Intuition's 'large action models'. Unlike traditional large language models that learn 'sequences of words', action models learn 'sequences of actions': given a situation, what should be done next. Investor Vinod Khosla recently told TechCrunch that he believes these action labels will be key to 'the emergence of intuition' — the ability of a model to truly generalize across tasks it was not explicitly trained on.

Why are investors willing to push the valuation from $2.3 billion to $6 billion within weeks? The answer lies in the grand narrative of 'Physical AI'. In recent years, the main battlefield of the AI industry has been the digital world — text, code, images, video. But applications like robotics, autonomous driving, and industrial automation — the 'physical world' — need models that can perceive, decide, and act in three-dimensional space, not models that just chat. That is exactly the direction General Intuition is targeting: it plans to use the funds to improve its general model and focus on 'robotic embodiments' — letting models truly 'live inside' robot bodies. The company also has a partnership with compute giant CoreWeave, with funds earmarked for expanding compute infrastructure and large-scale hiring. TechCrunch notes that the deal confirms earlier reporting on the project, and General Intuition is becoming one of the most iconic startups in the physical AI race.

The investor lineup is also highly symbolic. Valor Equity Partners is famous for backing SpaceX — if confirmed, this would be the fund's first investment in an AI lab since SpaceX. Point72 is Steve Cohen's giant hedge fund, which has been actively deploying in AI in recent years. Their entry, alongside continued participation from Khosla Ventures and General Catalyst, shows mainstream Wall Street capital is shifting from 'investing in AI applications' to 'investing in AI bodies'. A source close to the deal said the round is oversubscribed, and the company continues to field interest from more investors. Notably, the speed of General Intuition's valuation surge is itself a signal: when capital markets are willing to hand a startup less than a year old a $6 billion pre-money valuation, opportunity and froth coexist in physical AI — but it also shows that investor confidence in 'embodied intelligence' is gathering rapidly.

From a broader industry perspective, General Intuition's story is a microcosm of AI competition entering the 'second half'. The first half starred language: OpenAI, Anthropic, and Google used large language models to redefine how humans interact with information. The second half is about action: getting AI out of the screen and into the world. Nvidia CEO Jensen Huang has repeatedly said 'physical AI' is the next wave of artificial intelligence; Tesla's Optimus robot, Figure AI's humanoid robots, and every autonomous-driving company are competing for the same endgame — agents with general action capabilities. General Intuition's chosen path is 'learn to play games first, then learn to do things': gameplay data naturally contains vast 'perceive-decide-act' loops, making it the best corpus for training action models. Of course, transferring from the game world to the physical world is not easy — the physics, safety constraints, and long-tail scenarios of the real world are far more complex than games. But capital has already voted with valuations: if this path works, it will redefine the intelligent transformation of robotics, autonomous driving, and the entire physical economy.

📌 Source: TechCrunch (August 24, 2026) — 'Valor, Point72 back General Intuition at $6B valuation as AI startup pushes into robotics' by Rebecca Bellan. Link: techcrunch.com/2026/08/24/valor-point72-back-general-intuition-at-6b-valuation-as-ai-startup-pushes-into-robotics/

🤔 Frequently Asked Questions

Q1: What does General Intuition do?

General Intuition is a New York-based startup building a foundation model that trains generalized AI agents to act through space and time. It trains 'large action models' using hundreds of millions of hours of gameplay recordings and 'action labels' (records of which button a player pressed and when), aiming to give AI general action capabilities for robotics and the physical world.

Q2: What is the valuation for this round?

General Intuition is in talks to raise at a $6 billion pre-money valuation, with new investors including Valor Equity Partners and Point72, and existing shareholders Khosla Ventures and General Catalyst participating. Just weeks ago, the company raised $320 million at a $2.3 billion valuation — a roughly 2.6x increase in a very short time.

Q3: What are 'large action models'?

Large action models are General Intuition's core technical direction. Unlike traditional large language models that learn 'sequences of words', action models learn 'sequences of actions': what to do next in a given situation. Training data comes from the 'perceive-decide-act' loops of hundreds of millions of hours of gameplay, which investors believe is key to 'the emergence of intuition' — generalization to tasks the model was not trained on.

Q4: Why are investors so bullish on physical AI?

Because AI competition has entered the 'second half': the first half used language models to redefine human-machine interaction; the second half is getting AI out of screens and into the real world. Robotics, autonomous driving, and industrial automation all need models that perceive, decide, and act in three-dimensional space. Nvidia CEO Jensen Huang has repeatedly called physical AI the next wave, and Tesla, Figure AI, and others are competing for the same endgame.

🛠️ Recommended Tools

  • Text Summarizer - Quickly distill key points from funding news, industry reports and technical papers to track the physical AI race
  • Image Converter - Handle image format conversion for robotics vision and demo materials, making them easy to present in documents and slides
  • Word Counter - Count words in technical documents and investment memos to ensure reports meet length requirements

Summary

General Intuition's funding talks — leaping from a $2.3 billion valuation to a $6 billion pre-money valuation within weeks — are the most direct barometer of the physical AI race's heat. This startup spun out of gaming clip platform Medal holds a unique data asset: hundreds of millions of hours of gameplay and 'action labels', betting on 'large action models' — teaching AI to learn 'sequences of actions' rather than 'sequences of words'. The lineup of Valor (SpaceX backer), Point72, Khosla Ventures, and General Catalyst marks mainstream Wall Street capital formally entering 'embodied intelligence'. Of course, transferring from the game world to the physical world is full of challenges: real-world physics, safety constraints, and long-tail scenarios are far more complex than games. But capital has already spoken with real money — when AI competition enters the 'second half', whoever can teach AI to 'act' holds the ticket to the next era.