General Intuition’s valuation jumps to $6 billion as top investors bet on physical AI

General Intuition is in talks to raise new money at a $6 billion pre-money valuation, just weeks after the New York startup closed a $320 million round at a $2.3 billion valuation. The rapid markup underscores how aggressively investors are now pricing companies that aim to build AI systems for the physical world, especially robotics.

According to people familiar with the deal, the new round would include fresh backing from Valor Equity Partners, Point72 Ventures and Seven Seven Six, alongside continuing support from existing investors Khosla Ventures and General Catalyst. The startup, which was spun out of video game clip-sharing company Medal last year, is positioning itself as a foundation-model company focused on generalized agents that can learn how to move through space and time.

The financing, which is still being finalized, arrives amid a broader rush into so-called physical AI, where companies are trying to move large-model capabilities from language and image generation into robots, autonomous systems and other embodied applications. General Intuition’s pitch combines large-scale gameplay data, compute-heavy training and robotics ambitions in a way that has clearly caught investor attention.

Why General Intuition is drawing so much investor attention

General Intuition is attracting capital because it is selling a vision that goes beyond text or video generation: it wants AI agents that can understand motion, sequence and decision-making in physical environments. That promise has made it one of the most closely watched startups in the emerging robotics AI market.

The company’s appeal also comes from its unusually rich starting dataset. Rather than building from scratch with only synthetic or robotic sensor data, General Intuition began with years of gameplay footage and interaction logs collected through Medal, the gaming and clip-sharing platform from which it was spun out in October 2025.

Those logs include what the company describes as “action labels” — records showing which inputs players made and when. In practical terms, that gives the startup a way to study how humans make choices in fast-changing environments, which is relevant to training systems that may one day control robots or other machines in the real world.

Vinod Khosla has argued that action labels could play an important role in the rise of what he calls “intuition” in AI, meaning the ability for models to generalize across tasks they were never explicitly trained to perform.

How did General Intuition get here so quickly?

General Intuition reached this point by moving from a consumer gaming product into a frontier AI bet in less than a year. The company was carved out of Medal by CEO Pim de Witte, who repurposed the platform’s data advantage into a machine-learning company aimed at training general agents for physical tasks.

That transition mirrors a broader pattern in AI, where startups are increasingly being built around data moats rather than only model architecture. In General Intuition’s case, the moat is not just the scale of the data but the type of behavior it captures: timing, intent and action in dynamic environments.

Investors have also been encouraged by the speed with which the company has already raised capital. The current financing discussion comes only weeks after the startup secured $320 million at a $2.3 billion valuation, indicating that demand for access to the company has continued to build.

Key fundraising milestones

General Intuition’s recent funding activity suggests a startup moving through valuation tiers at unusual speed.

What is General Intuition building?

General Intuition is building a foundation model designed to train generalized AI agents that can navigate space and time. In other words, the company wants models that can understand movement, environment, timing and action in a way that could transfer to robotics and other embodied systems.

That ambition places the startup at the intersection of several fast-growing AI categories: foundation models, robotics, compute infrastructure and advanced agentic systems. The company’s focus is especially relevant at a moment when many investors believe the next wave of AI value creation will come from linking intelligence to action.

General Intuition has said it intends to use the new funding to improve its core model with a specific emphasis on robotic embodiments. That likely means more research, more training runs and more hardware-heavy experimentation designed to make the system useful in machines that operate in the physical world.

Where the data advantage comes from

The startup’s dataset stems from Medal, which built a large archive of gaming clips and related interaction signals. Those records include not just what happened on screen, but also the sequence of human inputs that produced each action.

That distinction matters because action data can reveal patterns that are harder to extract from video alone. A model trained on such material may learn how humans decide, react and adapt under pressure, which is useful for tasks requiring control and planning rather than simple recognition.

In essence, General Intuition is betting that the path to better embodied AI may run through gaming behavior at scale.

Why robotics investors are paying attention now

Robotics has become one of the most competitive areas in AI because it offers a concrete business case for models that can perceive and act. Unlike chatbots, robots must deal with messy real-world conditions, from physical obstacles to changing environments and delayed feedback.

That is exactly why companies with strong data pipelines are getting premium valuations. Investors are increasingly willing to pay for a startup that can combine frontier-model talent, a proprietary dataset and the compute budget needed to iterate quickly.

General Intuition appears to fit that template. Its planned use of proceeds includes more compute spending and additional hiring, both of which are critical for a company that needs to train large-scale systems and recruit researchers, engineers and robotics specialists.

The startup already has a relationship with CoreWeave, the cloud infrastructure provider increasingly used by AI labs to access high-end compute. A larger round would likely allow General Intuition to expand those resources further as it pushes deeper into model training.

Who is backing the round?

The proposed round includes a mix of returning and new investors, a sign that the company’s existing backers are being joined by funds that want exposure to the physical-AI theme.

• Valor Equity Partners — the firm best known for backing SpaceX and, if confirmed, making its first AI lab investment since that deal.
• Point72 Ventures — the venture arm of the hedge fund associated with Steve Cohen.
• Seven Seven Six — the venture firm founded by Alexis Ohanian.
• Khosla Ventures — an existing investor that has been outspoken about the promise of action data.
• General Catalyst — another returning backer participating again.

The participation of both new and existing investors suggests strong competition for allocation. A source close to the transaction said demand for the round has exceeded the available supply, which is often a sign that a startup’s narrative is resonating with investors even before the product is fully commercialized.

People close to the deal say the funding is oversubscribed, meaning more capital is being offered than the company is prepared to accept.

What does a $6 billion valuation mean for the market?

A $6 billion pre-money valuation for a startup still in the model-building phase would place General Intuition among the most richly valued young companies in the AI sector. It reflects investor conviction not just in the team, but in the category itself and the belief that robotics-focused AI could become a major platform market.

At the same time, such a sharp jump in valuation can raise expectations. Once a company is priced at this level, it must show that its technical progress, data advantage and go-to-market plan can justify the premium. For AI startups, those expectations usually center on model capability, infrastructure efficiency and the ability to recruit top talent before competitors do.

The speed of the valuation increase also illustrates how volatile the AI funding market has become. A company can move from being a promising startup to a multibillion-dollar platform bet in a matter of weeks if investors believe it has access to scarce data or a unique technical wedge.

Timeline of the deal and company formation

The following timeline captures the major public and reported steps in General Intuition’s short but eventful rise.

How does Medal fit into the strategy?

Medal matters because it gave General Intuition a dataset that most robotics startups would struggle to assemble on their own. Rather than starting with robots in warehouses or labs, the company has begun with recorded human behavior across hundreds of millions of hours of gameplay.

That foundation may help the startup build generalizable systems faster than competitors that rely only on narrow task data. The logic is that games create controlled but complex environments where humans repeatedly solve problems, make decisions and adjust strategies in real time.

If those patterns can be translated into model training, General Intuition may be able to teach AI not just to recognize what is happening, but to infer what to do next. That is the underlying bet investors are making.

Why this deal matters beyond one startup

This financing is a signal about where the AI market may be heading next. For the past two years, much of the capital in the sector has flowed into large language models, copilots and consumer chat interfaces. General Intuition shows that investors are now looking harder at companies trying to connect foundation models to the physical world.

That shift has implications for robotics, autonomous systems, manufacturing automation and other industries that need AI to do more than generate words. If the company succeeds, it could validate a new class of data-centric AI startups built around embodied intelligence rather than pure language processing.

It also reinforces the importance of compute as a strategic asset. Training models for physical environments will require substantial infrastructure, and companies that can secure both capital and cloud capacity may gain a lasting advantage.

What happens next?

General Intuition is still finalizing the round, so the exact size, structure and closing timetable remain subject to change. But the reported oversubscription suggests the company can likely choose among investors eager to participate.

For now, the more important story is the direction of travel. General Intuition is trying to turn gameplay data into a robotics-ready AI foundation model, and the market appears willing to pay a steep premium for that possibility.

If the company can convert its data advantage into demonstrable model gains, the current valuation may look prescient. If not, it will become another reminder that AI enthusiasm can outrun product readiness. Either way, the deal is a telling snapshot of the current race to build intelligence that can act, not just respond.