The Robotaxi Reality Check: How Close Are We Really to Driverless Cities?

01 The Waymo Lead

Waymo is the clear leader in robotaxi deployment. 200,000 paid rides per week across Phoenix, San Francisco, Los Angeles, and Austin. The company has driven over 50 million autonomous miles with no at-fault serious accidents. But the service operates only in geofenced areas — mapped, tested, and validated. Expanding to a new city takes 6-12 months of mapping and testing.

02 The Tesla Paradox

Tesla has the most vehicles on the road with 'Full Self-Driving' software — over 2 million. But Tesla's system is Level 2 (requires human supervision), not Level 4 (fully autonomous). Tesla's approach (end-to-end neural networks, no HD maps, camera-only) is fundamentally different from Waymo's (lidar, radar, HD maps, rule-based + ML). Tesla's bet is that scale of data will eventually solve autonomous driving. Waymo's bet is that you need the right sensors and careful engineering. Both could be right — or both could be wrong.

03 The Economics of Driverless

A robotaxi needs to cost less per mile than a human-driven Uber. Current cost: Waymo is approximately $2-3 per mile including vehicle, maintenance, remote support, and amortized R&D. Uber is approximately $2-3 per mile including driver. They're at parity. But robotaxi costs decrease with scale while Uber costs increase with labor inflation. The crossover is near — but only in geofenced areas. A driverless city requires solving the last 5% of driving scenarios, which may take another decade.