Why Physical AI Is the Next Frontier | Applied Intuition with a16z

Key points

• Applied Intuition describes itself as a physical AI company with a mission to put intelligence on a billion machines; automotive is only about 30% of its business today.
• The company has over a thousand engineers, 18 offices globally, and has raised about a billion dollars, all still in the bank.
• Physical AI differs from digital AI because it requires proprietary data collection, safety-critical validation, and deployment onto roughly 50 hardware platforms.
• The founders argue the Cruise shutdown by GM was a multi-varied problem involving safety, business model conflicts, union negotiations, and mishandling of the government response.
• Self-driving in personally owned vehicles will follow a 'wait, wait, wait, then a lot' pattern, becoming default in new cars by the early 2030s.
• Robotaxis will be available in many US cities by 2030 and routine by 2032–2033; driver-out long-haul trucking is a few years away.
• Dana, the company's biggest launch ever, is an agentic platform that makes autonomy development accessible—aiming for a high school kid to build autonomous systems.
• World models span a spectrum from physics-based simulation to fully neural, reactive video generation, with real-time performance constraints being the key moat.
• Trucking and mining face severe labor shortages and dangerous conditions—mining is 1% of the global labor pool but 8% of work-related fatalities—so operators actively want autonomy.
• The founders argue society must embrace technological progress, citing the Confucian saying 'no hand can block the sun.'