General Intuition, a New York-based AI startup, has drawn fresh investor interest after reports said the company is in talks for a new funding round at a $6 billion pre-money valuation. The talks show how fast investor interest has grown around physical AI, a field that seeks to give machines the ability to understand and act in the real world.
The possible deal has attracted Valor Equity Partners, Point72 Ventures, and Seven Seven Six as new investors, according to sources cited by TechCrunch. Existing backers Khosla Ventures and General Catalyst may also take part. The round has not yet reached a final close, so the $6 billion figure remains a reported target rather than a confirmed final valuation.
The scale of the possible deal is what makes the story notable. General Intuition had only a few months earlier raised $320 million at a $2.3 billion valuation. That sharp rise has placed the company among the more closely watched young firms in the physical AI space.
A Huge Jump in Company Value
In June 2026, General Intuition raised $320 million in Series A funding at a $2.3 billion valuation. Khosla Ventures led that round, with support from General Catalyst and other major investors. The deal gave the young company a large financial base at an early stage of its development.
Now, only about two months later, the company is in talks at a pre-money value of $6 billion. That would put the new figure at more than twice the value from its June round.
Such a rise reflects a wider shift in the AI market. For several years, much of the biggest AI investment went toward chatbots, language models, software agents, and data tools. Investors now see another major opportunity in machines that can act outside a computer screen.
Robots need a different type of intelligence from a chatbot. A robot must know where objects are, how they move, what may happen after an action, and how to respond when something does not go as planned. General Intuition believes its approach can help solve part of that problem.
The Unusual Role of Video Games
The company has built its AI work around a simple but unusual idea: video games can provide useful data for physical AI.
At first, that may sound strange. A game world is not the real world. A digital character does not face the same limits as a robot. Yet games contain a huge amount of information about movement, space, timing, cause, and effect.
General Intuition uses gameplay footage as part of the data for its models. It also has access to information about the actions that players take, such as which buttons they press and when they press them. That extra information matters because the system can see not only what happens on screen but also the action that caused it.
The company came out of Medal, a platform for game clips. That background gives General Intuition access to a large supply of gameplay material. Earlier reports said Medal had a dataset of about 2 billion videos per year from 10 million monthly active users.
The central idea is simple. If AI can learn how people move through digital spaces, it may gain some useful knowledge about space and time before it ever controls a physical machine.
Why This Could Matter for Robots
One of the hardest problems in robotics is data.
A language model can learn from a vast amount of text that already exists online. A robot does not have the same luxury. A robot needs physical examples. It must see how a hand reaches for an object, how a machine walks across a floor, how a body reacts to an obstacle, and how an action changes the world.
That type of data can cost a lot of money and time to collect.
A company may need robots, special facilities, cameras, human operators, and safety systems just to create useful training data. General Intuition believes video games can reduce part of that burden.
The company does not claim that game data can replace all real-world data. Instead, its approach aims to give an AI model a strong base before it receives a smaller amount of physical data.
That could become a major advantage if the method works at scale.
An Eight-Minute Robot Test
One of the most striking examples from General Intuition came from a test with a quadrupedal robot.
According to TechCrunch, the company used about eight minutes of real-world robotics data to fine-tune its model for the robot. The machine used a front-facing camera as it moved through an office. It faced obstacles such as chairs and a trash bin, made mistakes, and changed its path.
The result does not prove that General Intuition has solved robotics. The test was small, and a short demo cannot show how well a system will work across many locations, machines, and tasks.
Still, the result supports the company’s larger argument. A model that already has some understanding of space, motion, and time from digital environments may need less physical data to adapt to a real machine.
That is the part investors appear to find attractive.
A Bigger Bet on Physical AI
General Intuition is part of a much larger race.
Many AI companies now want to build systems that do more than produce text or images. They want models that can understand an environment and take action inside it. This idea sits at the heart of physical AI and robotics.
The possible $6 billion valuation shows that investors are willing to place very large bets on this next stage of AI.
The interest also reflects a basic economic opportunity. Software can reach millions of people with little extra cost. Robots are harder to scale because they need hardware, factories, supply chains, batteries, sensors, and safe physical spaces.
Better AI could make those machines more useful and easier to control. If that happens, the market could extend far beyond humanoid robots. It could include factories, warehouses, homes, transport, healthcare, and many other areas.
The Risks Behind the $6 Billion Bet
The valuation also comes with major risks.
The biggest question is whether skills learned from games can transfer well to the real world. A game can teach useful ideas about space and movement, but physical reality has many details that a digital world may not capture.
Real objects have weight, friction, damage, and unpredictable behavior. A robot can also face lighting changes, poor camera views, uneven floors, and other problems that are hard to model perfectly.
There is another issue: competition.
General Intuition is not alone in the race for better robot intelligence. Several well-funded companies are also working on foundation models for machines. These firms have access to large amounts of real-world robot data, advanced hardware, and major investors.
General Intuition therefore has to prove that its unusual data strategy can create a real advantage.
What the New Funding Could Mean
If the reported $6 billion deal closes, General Intuition would gain a much larger financial base for research, hiring, computing, data, and robot tests.
The company could use that capital to improve its models and move from small demonstrations toward broader real-world use.
For investors, the bet is not simply on a robot that can walk through an office. It is on the idea that one AI model could learn general skills across many environments and later transfer those skills to physical machines.
That is a very ambitious goal.
A Key Moment for General Intuition
The reported $6 billion valuation is still subject to change because the funding talks are not final. Even so, the jump from $2.3 billion in June to a possible $6 billion now is a clear sign of the excitement around physical AI.
General Intuition has chosen a different path from many robotics companies. Rather than rely only on costly physical data, it wants to use the vast world of video games as a source of knowledge about movement, space, and action.
The idea is bold, but it has a clear reason behind it. If digital experience can give AI a useful sense of how the world works, then robots may need less time and money to learn physical tasks.
The next stage will decide whether that promise is real. A $6 billion valuation would show that investors believe the answer could be yes. But the company still has to turn its unusual data advantage into reliable machines that work outside the game world.
ALSO READ: WATER Raises $2.5M to Build AI for Adaptive Surfaces