Tesla disclosed during its second-quarter earnings call that its robotaxi fleet has logged more than 380,000 unsupervised miles across six cities in two states without what the company described as a "notable" safety incident2.
Ashok Elluswamy, Tesla's vice president of AI software, highlighted the fleet's safety record during the call. "Zero notable incidents over 380,000 miles," Elluswamy told investors. Any reported incidents involved "other actors impacting us when we were stationary," he added.
"I'd like to emphasize how safe the operation has been so far," Elluswamy said.
Elluswamy said the results support Tesla's camera-based approach to autonomous driving, pushing back against the view that lidar, radar, and HD maps are necessary for safe autonomy. "Historically, the so-called experts have always claimed that you need lidars, radars, HD maps," Elluswamy said, according to the excerpt of his remarks.
Separately, Tesla announced it will no longer reach volume production of its Cybercab, Tesla Semi, and Megapack 3 in 2026, shifting focus instead to increasing battery production — particularly the 4680 cell1. The company said it is still building manufacturing lines for these products.
ANALYSIS The 380,000-mile figure and the "zero notable incidents" framing give Tesla a concrete operational data point to present to investors and regulators as it scales its robotaxi program. However, the company's own characterization of what constitutes a "notable" incident is doing significant work in that claim — the qualifier leaves open the possibility of minor incidents that fell below Tesla's internal threshold.
The delay of Cybercab volume production beyond 2026 creates a tension with the robotaxi safety narrative: Tesla is demonstrating autonomous driving capability in limited deployment while simultaneously pushing back the timeline for the dedicated hardware platform intended to carry that capability at scale.
Tesla's emphasis on a camera-only sensor suite — without lidar or radar — remains a distinctive technical bet among autonomous vehicle developers.