General Intuition has secured a $320 million funding round at a $2.3 billion valuation, positioning the startup to capitalize on a predicted paradigm shift in embodied artificial intelligence. Under CEO Pim de Witte and with lead backing from Vinod Khosla, the company argues that robotics is entering its foundation model era, mirroring the transition that large language models triggered in generative AI. Rather than engineering isolated systems for specific robots or environments, General Intuition is developing a generalized physical AI base layer focused on spatial-temporal reasoning. The startup’s methodology departs from conventional robotics data strategies. Instead of accumulating vast volumes of task-specific real-world footage, General Intuition trained its core architecture on extensive video game datasets containing precise controller inputs and human behavioral patterns. De Witte maintains that high-quality interactive data, rather than sheer volume, cultivates the intuitive reasoning required for robots to navigate complex, shifting environments. This approach fundamentally challenges industry norms, with de Witte asserting that future deployments will require only minutes of targeted real-world fine-tuning, rendering current multi-million-hour data collection efforts largely obsolete. Technical validation of the model occurred during a recent trial with a quadrupedal robot. After fine-tuning on just eight minutes of real-world video, the system successfully navigated an unstructured office setting using only a single front-facing camera. It maintained stability and pathfinding while accounting for dynamic obstacles and moving personnel, achieving zero-shot performance without auxiliary sensor arrays. Both de Witte and Khosla cite this demonstration as evidence that the model has internalized universal movement and interaction principles rather than memorizing static environments. General Intuition’s long-term strategy explicitly avoids hardware manufacturing. The company aims to serve as the foundational software infrastructure for the broader robotics industry, offering a standardized base model that third-party developers can adapt to various mechanical forms and operational tasks. By centralizing physical AI reasoning, the startup intends to compress development timelines and lower computational barriers for manufacturers across industrial, commercial, and consumer sectors. The $320 million injection reflects growing investor confidence that physical AI will consolidate around general-purpose foundation models, transforming robotics from a fragmented engineering discipline into a scalable software-driven platform.