The a16z Show

Snapshot · The a16z Show

Building the Physical AI Stack | Travis Kalanick on TBPN

Explore episode Jul 23, 2026

Where this was said

Federal Preemption, Regulatory Capture, and the Trial Lawyer Problem

At 34:50 · chapter starts 32:10

The conversation moves to regulation, prompted by a host asking whether Kalanick ever developed a philosophy on federal versus state-by-state regulation from his Uber years. His answer is blunt: federal preemption is good when you want regulatory capture — you're trying to lock out competitors. Uber never played that game, he says; they always tried to open markets and let the best company win. He warns viewers to watch which 'closed-weight' AI companies are suddenly eager to be regulated. He then connects this to autonomous vehicles, noting that trial lawyers fighting self-driving cars because they might be 'too safe' is entirely predictable. He argues that every systemically bad transport rule traces back to two lobbying forces: trial lawyers, who make money from accidents, and insurance companies, who profit from predictable actuarial outcomes. He reveals that when Uber launched in D.C., it was hit with a $1.5 million per-ride liability requirement — versus $25,000 for taxis — a number that delighted both industries.

Government
Federal Preemption Is Regulatory Capture

Building the Physical AI Stack | Travis Kalanick on TBPN · Jul 23, 2026 Government

Federal preemption sounds principled, but Kalanick calls it out plainly: it's regulatory capture. Companies pushing for federal AI rules want to squeeze out competitors. Uber never proposed rules that benefited them over others — they just opened markets and competed. He warns listeners to watch which closed-weight AI companies are suddenly eager to be regulated.

Government
Trial Lawyers and Insurers vs. Autonomous Vehicles

Building the Physical AI Stack | Travis Kalanick on TBPN · Jul 23, 2026 Government

Every systemically bad transport rule traces back to trial lawyers and insurance companies. Insurers love predictable accidents — they show up in actuarial tables and generate premiums. Autonomous vehicles threaten that model. Kalanick revealed Uber was hit with a $1.5M-per-ride insurance requirement in D.C., vs. $25K for taxis — a gift to both industries.

Technology
Autonomous Burritos and the Wheelbase for Robots

Building the Physical AI Stack | Travis Kalanick on TBPN · Jul 23, 2026 Technology

Atoms isn't building humanoids — it's building wheeled industrial robots at scale. Kalanick calls this 'wheelbase for robots.' In food, that means an autonomous delivery vehicle he calls an 'autonomous burrito' that drops food for $0.75 instead of the $12 Uber Eats charges. In mining, it's haul trucks. Every industry that moves things needs this layer.

Similar snapshots