AUR — Q2 2026 Financial Report Analysis

A bigger fleet on the balance sheet, still almost no revenue

Aurora Innovation, which runs driverless Class 8 trucks in Texas, booked $2 million of revenue in the second quarter of 2026 against a net loss of $270 million, up from a $201 million loss a year earlier. Two things drove the wider loss. Spending rose: research and development grew $21 million to $211 million and selling, general and administrative costs grew $14 million to $50 million. And a $32 million swing in a non-cash accounting line: the warrants and earnout shares left over from the company's 2021 SPAC listing are revalued every quarter as the share price moves, and they produced a $16 million charge this year against a $16 million gain a year ago. The quarter matters for what it set up more than for its results. Vehicles on the balance sheet went from $42 million at year-end to $106 million. The second-generation truck fleet (International LT trucks with Aurora's new hardware kit) launched in July with no one behind the wheel. And management kept its 2026 targets: $14–16 million of revenue and more than 200 driverless trucks running by year-end.

At a glance

• $1.22 billion in cash and short-term investments ($136 million cash plus $1,081 million of investments). At the company's own guided spending rate of $190–220 million a quarter, that covers roughly six quarters before any new share sales.
• $7 million cost of revenue against $2 million of revenue. Every dollar Aurora earns hauling freight still costs about $3.50 to deliver (terminals, staff, fuel), before any R&D. The new hardware kit is meant to fix this: management says it cuts Aurora Driver hardware costs by more than 50%.
• Shares outstanding up 10.7% in a year (1,976 million weighted average vs 1,785 million). Aurora is funding itself by selling stock. It sold 30 million shares this quarter at an average $7.50 for $215 million net, so each shareholder owns a steadily smaller slice.

The numbers

A note on the percentages. The filing itself calls revenue growth "100%", but $2 million vs $1 million are figures rounded to the nearest million, so the true growth rate could be very different. We don't treat it as a growth figure. Operating margin (the share of revenue left after running the business, before interest and tax) works out to around -13,000%, which says nothing useful at this stage, so we leave it out. Growth rates for the loss and loss per share are also omitted from the headline figures, because a percentage change between two losses misleads more than it informs.

Where the money went

Research and development ($211 million, +11%) is still the bulk of spending. The 10-Q attributes the increase to "hardware development costs, cloud computing costs, non-cash stock-based compensation costs, and personnel costs." Some costs moved out of R&D, which partly offsets this: trucking-operations costs now sit in cost of revenue (since the April 2025 commercial launch), and some staff moved into SG&A under "a realignment of resources." Excluding stock-based compensation (pay in shares, a real cost to shareholders but not a cash outlay), R&D was $164 million.

SG&A ($50 million, +39%) rose faster than R&D, but part of that is the same reclassification: costs that used to be booked as R&D are now booked as SG&A. The 39% jump overstates how much overhead actually grew. Excluding stock pay, SG&A was $37 million.

Cost of revenue ($7 million, +40%) rose on "terminal, personnel, and fuel expenses due to increased utilization and geographical expansion." Revenue rose for the matching reasons: "increased utilization, geographical expansion, and higher fuel surcharges." Fuel surcharges are pass-through pricing, so part of the revenue gain reflects diesel prices, not more freight hauled.

Operationally, Aurora's shareholder letter says the Aurora Driver had completed nearly 440,000 driverless miles since launch through June. It reports a 100% on-time record and zero collisions attributed to the Aurora Driver, both company-reported figures. Cumulative commercial miles, including loads with a safety operator on board, passed 6 million. New customer agreements in the quarter included Charger Logistics and Value Truck (Dallas–Laredo, Fort Worth–Phoenix). Aurora also began driverless runs for Detmar Logistics hauling frac sand in West Texas.

What the headline numbers hide

• The cash loss is smaller than the accounting loss, but cash burn rose faster. Over six months, operating activities used $384 million against a $493 million net loss. The gap is mostly $106 million of stock-based compensation and the $17 million derivative charge, neither of which costs cash. But cash used rose $98 million year on year, faster than the loss. The 10-Q gives two reasons: "the settlement of the annual bonus in cash rather than Class A common stock, and hardware development programs." The CFO puts the bonus at $63 million, paid this quarter. Last year that bonus was paid in shares, so it never showed up as cash spent.
• The bonus was cash in form, but shareholders still paid for it. The CFO's letter says the cash bonus was "funded through our at-the-market (ATM) program", meaning new shares sold on the open market. Excluding the bonus, Aurora says Q2 cash spend was within its $190–220 million quarterly target. Including it, Q2 operating cash use was about $225 million plus $31 million of capital spending: roughly $256 million out the door in the quarter.
• Adjusted EBITDA leaves out a lot. Aurora's non-GAAP adjusted EBITDA (profit before interest, tax, depreciation and certain other items) was -$200 million. It strips out $60 million of stock pay, $6 million of depreciation, the $16 million derivative charge, and $12 million of interest income. Stock pay is the line to watch: $552 million of already-granted restricted stock is still to be expensed over about three years, so this cost will not shrink soon.
• The derivative line will mostly go away. The 21 million public and private SPAC warrants (exercise price $11.50) expire on November 3, 2026. The Q2 ATM sales averaged $7.50 a share, well below that strike. The earnout-share liability ($30 million, up from $15 million) remains and will keep moving with the stock price.
• Six-month other income was flattered by a one-off. H1 other income rose to $34 million from $25 million. The 10-Q says this came from "remeasurement of non-marketable equity securities resulting in unrealized gains," which offset lower interest income as rates fell. In Q2 alone, other income slipped to $12 million from $13 million.
• Loss per share understates the loss growth. Net loss grew 34%, but loss per share grew less (about -$0.137 vs -$0.113 unrounded) because there were 10.7% more shares to spread it across. Dilution is softening the per-share figure. Another 225 million potential shares (restricted stock units, options, warrants, earnout shares) are excluded from the share count because they would reduce the loss per share.
• Guidance held. The 2026 revenue range ($14–16 million) and the cash-use range ($190–220 million a quarter on average, including about $150 million of full-year capital spending) were both reaffirmed, not cut.

Takeaway: Aurora's 2026 depends almost entirely on the second half. Six-month revenue was $3 million, so hitting the $14–16 million target needs $11–13 million in July–December, with management saying Q4 alone will be over half the year. That means going from 20–25 second-generation trucks at the end of September to 200+ by December. The money for that ramp is coming from roughly $200 million a quarter of spending, paid for by a shrinking $1.22 billion cash pile and steady share sales. Execution on the truck ramp, not this quarter's revenue, is what decides whether the dilution pays off.

Outlook

What management says. The CFO's letter expects 2026 revenue of $14–16 million, "back-end loaded — with the fourth quarter projected to contribute over half of full year revenue." It expects to exit the year with more than 200 driverless trucks, which it says equals "an approximately $80 million revenue run-rate" for its current model, where Aurora owns the trucks and sells hauling (Transportation as a Service). The bigger shift is meant to come in 2027: Driver as a Service, where customers own the trucks and pay Aurora a per-mile fee for the software and hardware. Aurora calls that model "asset-light and high margin." Supply milestones: 20–25 second-generation trucks by the end of Q3, upfitter Roush ramping to a 1,000-truck annual rate in October, Volvo Autonomous Solutions starting driverless operations on Aurora-powered Volvo trucks in Q1 2027 with more than 300 trucks targeted by end of 2027, and AUMOVIO (formerly Continental) starting production of the third-generation hardware in the second half of 2027.

Since the quarter closed. At its September 23 Analyst and Investor Day (8-K, Exhibit 99.1), Aurora said it had passed 500,000 driverless miles. Hirschbach plans to own and run 500 trucks under a Driver-as-a-Service agreement, with deliveries starting in 2027, and Aurora set a 2030 target of more than 30,000 driverless trucks. Those are targets, not results, and the Hirschbach figure is stated as "intent."

Our read. The run-rate math is worth checking against the numbers the company has given. $80 million spread over 200 trucks is about $400,000 of revenue per truck per year. Aurora says its customer trucks already average over 225,000 miles a year, which implies roughly $1.75–1.80 of revenue per mile. That is plausible for a carrier-style service, but under this model Aurora also pays for the trucks, terminals and fuel. Cost of revenue was 3.5 times revenue this quarter, so the more than 50% hardware-cost cut from the second-generation kit is what has to close the gap.

The funding picture is adequate but not comfortable. $1.22 billion covers about six quarters at the guided burn, which runs to roughly the end of 2027: close to when Driver-as-a-Service revenue is supposed to start. The company says it expects to "opportunistically raise additional capital," so investors should assume share sales continue. Three things to check in the Q3 report: whether the second-generation truck count reached 20–25 by September 30, whether quarterly revenue moves meaningfully above $2 million, and whether cash use excluding one-off items stays inside $190–220 million while capital spending ramps toward the $150 million full-year plan.