The Thesis: Driverless Freight Is Now a Manufacturing Problem
Most autonomy capital chased the robotaxi. Aurora bet the other way: long-haul trucking, where highways are structured, freight repeats the same lanes, and Hours-of-Service rules cap how far a human can legally drive in a day. A machine that can run overnight does not just replace a driver. It raises the duty cycle of an 80,000-pound asset.
That bet is no longer a demo. Wikipedia records that Aurora began commercial driverless Class 8 hauls on the Dallas-Houston corridor in late April 2025. At Analyst and Investor Day on 23 September 2026 the company said the Aurora Driver had completed more than 500,000 driverless miles since commercial launch, that 2026 Transport-as-a-Service capacity is fully allocated to 200 driverless trucks, and that McLane, Werner and other customer trucks were already averaging more than 225,000 miles a year.
From a General-Purpose Driver to Freight First
Aurora was founded in 2017 in Pittsburgh by Chris Urmson, formerly of Google's self-driving program, Drew Bagnell, formerly of Uber ATG, and Sterling Anderson, formerly of Tesla Autopilot. Urmson remains chief executive. Anderson left in May 2025 and later joined General Motors.
Two transactions still shape the company. In December 2020 Aurora acquired Uber's Advanced Technologies Group in a stock deal that left Uber with a large equity stake. Aurora went public in November 2021 through a SPAC merger with Reinvent Technology Partners Y, then sequenced trucking first and pushed ride-hailing later. Uber has spent 2026 selling that ATG paper. After a 15 September block sale of 29,369,611 Class A shares at $6.205, Uber reported beneficial ownership of 157,103,800 shares, about 9.2% of Class A stock.
Hardware Built for a Million Miles
The stopping distance of a loaded Class 8 truck is why Aurora built its own lidar. FirstLight is an FMCW sensor that measures range and velocity directly. Wikipedia describes the second-generation kit as intended to double FirstLight range to 1,000 meters and roughly halve hardware cost, on platforms including the International LT Series. Analyst Day said that kit is engineered for a one-million-mile operating life and more than a 50% cost cut, with upfitter Roush targeting 20 trucks a week in October.
The system question on the original Texas lanes is largely answered. Through the end of June the Driver had completed nearly 440,000 driverless miles with what Aurora reported as 100% on-time performance and zero Aurora Driver-attributed collisions; the September figure is more than 500,000 miles. The remaining hardware question is industrial: whether second-generation kits can be built fast enough to fill a 200-truck year-end allocation.
TaaS Is the Proof. DaaS Is the Business.
Aurora does not want to own trucks for long. The 23 September 8-K deck treats Transportation-as-a-Service as a proof of concept priced at about $2 per mile including fuel surcharge, with Aurora owning a fleet it expects to cap around 500 trucks. Driver-as-a-Service, at $0.85 or more per mile, is the growth engine: customers own the vehicles and Aurora sells the driver toward tens of thousands of trucks.
2026 revenue is still TaaS. Analyst Day said that capacity is fully allocated to 200 driverless trucks and that Hirschbach intends to own and operate 500 DaaS trucks with deliveries from 2027. That Hirschbach plan is an intention, not a disclosed definitive order. Aurora said it is in talks with additional DaaS customers. Vehicle-side, the same deck shows International LT trucks upfitted by Roush, lineside integration with Volvo's VNL Autonomous, and a future PACCAR path.
The Financial Reality: Dilution Is Still the Customer
Reported scale is still tiny against the loss. Second-quarter 2026 revenue was $2 million. Operating loss was $266 million. Net loss was $270 million. Full-year guidance remains $14 to $16 million, with a targeted TaaS run-rate of about $80 million if Aurora exits 2026 with more than 200 driverless trucks. The company ended June with about $1.2 billion in cash and short-term investments after selling 30 million ATM shares for $215 million net. Average quarterly cash use is still guided at $190 to $220 million.
Those figures are funded by equity, not freight bills. The Analyst Day stack is a target, not a result: run-rate gross-margin breakeven in the first half of 2027 at about 500 trucks, more than $200 million of revenue and more than 1,000 trucks exiting 2027, positive free cash flow on a run-rate basis exiting 2028, and a 2030 vision of more than 30,000 trucks, $5 billion-plus annual revenue and about 60% gross margin. Uber's cut to about 9.2% is a reminder that a founding partner is still recycling the 2020 ATG position.
Constraints: The Line, the MOU, and the ATM
The original Dallas-Houston safety argument is no longer the bottleneck. The constraints are now industrial and financial. Roush's 20-trucks-a-week October rate is a target. The 200-truck 2026 book is sold, not yet built. Hirschbach's 500-truck DaaS plan can still fail to become a binding order. ATM equity remains part of the funding mix while quarterly cash use stays near $200 million.
Ride-hailing stays in the future. The investment case is whether Aurora can print and deploy second-generation trucks fast enough to turn a fully allocated 2026 TaaS book into 2027 DaaS mix before shareholders tire of funding the burn. Uber's third large 2026 block sale is one data point on that clock, not a verdict on the freight product.
What to Watch
The next operating test is the truck count. Management has said 2026 TaaS is fully allocated to 200 driverless trucks and that Roush is aiming for 20 trucks a week in October. The last disclosed numerical interim target, from the 29 July shareholder letter, was 20 to 25 second-generation trucks by the end of the third quarter. Analyst Day did not publish an updated fleet count, so that Q3 figure is unverified as of 27 September 2026.
Further out, watch three conversions: Hirschbach's 500-truck intention into a definitive DaaS contract, the 2027 shift from Aurora-owned TaaS toward customer-owned DaaS at the $0.85-plus per-mile price, and whether run-rate gross-margin breakeven in the first half of 2027 at about 500 trucks remains the company's target after the next earnings print.