The Challenge: Teaching Machines to Understand Physical Space
While most AI applications focus on text processing or static image analysis, there's a less-explored but increasingly well-funded frontier: training machines to understand and move within physical space in real time. General Intuition, the New York-based startup founded by Pim de Witte (previously leading Medal, a gaming video platform), is in talks to raise approximately $300 million at a valuation slightly over $2 billion.
This new funding round comes just eight months after the startup's independent launch from Medal, when it closed a $134 million seed round. Backing comes from high-profile investors including Jeff Bezos and Eric Schmidt, plus firms like Khosla Ventures and General Catalyst.
World Models: The Key Differentiator
General Intuition's proposition isn't to build chatbots or image generators, but rather "world models"—AI systems capable of perceiving, anticipating, and interacting in real environments. The engine behind this capability is a unique dataset: 2 billion videos annually generated by 10 million monthly active Medal users.
"First-person spatial-temporal reasoning—the perspective you gain playing video games—is the perfect training material for teaching machines how to navigate physical worlds." Unlike competitors in the world models space (Runway, Decart, World Labs), General Intuition doesn't sell the models themselves; trained agents are the actual product. This strategic pivot is crucial: they don't compete in a saturated tools market but rather in autonomous automation—a far larger opportunity.
Medal's dataset appeal has already captured OpenAI's attention—the company previously attempted to acquire the platform. Sources suggest other major AI labs have knocked on their door too.
From Gaming to Manufacturing and Logistics
While General Intuition's initial focus is clearly training agents for gaming and robotics (technically adjacent sectors), the company is building foundations for something more ambitious: embodied AI that could transform physical-presence industries.
Manufacturing, logistics, and retail are sectors where a system understanding spatial dynamics in real time could unlock significant productivity gains. Imagine agents that don't just process orders but understand how to move a pallet through a congested warehouse or navigate a production line with dynamic obstacles.
The competitive landscape is heating up fast. Google recently integrated Google Maps data into its Genie 3 model, enabling more realistic simulations. But General Intuition has an advantage: a dataset generated by users interacting from first-person perspective, not static maps.
What This Means for Technology Leaders
For decision-makers evaluating automation and AI tools, this massive investment in world models and spatial agents signals a market shift. The next generation of production AI won't just read emails or analyze databases—it will navigate and act within physical spaces.
The company plans to invest new funds in scaling compute capacity to release a new product by late summer or early fall. Monitoring these launches could reveal early opportunities to integrate spatially-aware AI into manufacturing operations, logistics, or tasks currently requiring human presence.
This is the kind of technology capital is voting for today. Those who understand this early can position themselves for a competitive edge in the next automation cycle.