Reading Aurora's Q3 2026 10-Q like an auditor

The filing: Aurora Innovation Form 10-Q for the quarter ended September 30, 2026, filed with the SEC on October 28, 2026.

Last reviewed Nov 8, 2026 by Blake, CPA.

A 10-Q is the quarterly report a public company files with the SEC, and the thing you should know right away is that its financial statements are labeled "unaudited." Auditors do a lighter review of a 10-Q than the full audit the annual 10-K gets, so I read one a little more skeptically.

When I was auditing at EY, I never started with the press release. I started with the balance sheet, then the cash flow statement, then the footnotes, and every number below comes straight from the filing, with the note or section it's in.

I also covered Aurora in a video on November 4, Aurora Stock (AUR): What Has to Go Right for $11.55?, so I'll point you to the spot in the video where it lines up.

1. Cash, investments and the burn rate

The first question I ask about any company that loses money is simple... how much money does it have, and how fast is it going through it?

Aurora's balance sheet shows $141 million of cash and cash equivalents and $893 million of short-term investments at September 30, 2026. The investments are mostly U.S. Treasury securities and corporate bonds, according to Note 4, which is what you want to see, because it means the money is parked somewhere boring and easy to sell.

Source: condensed consolidated balance sheets in this 10-Q and in the June 30, 2026 10-Q. The totals are my addition. Restricted cash of $17 million is left out, which matches how MD&A (management's discussion and analysis, the section where the company explains its own numbers) describes liquidity.

Now for cash burn, which basically just means how much cash a company uses up in a period. The 10-Q's cash flow statement only shows year-to-date totals, so to get the third quarter by itself, you subtract the six-month numbers in the June 10-Q from the nine-month numbers in this one.

Source: condensed consolidated statements of cash flows in each 10-Q.

That last line is a rough version of negative free cash flow, which is the cash left over after running the business and buying the equipment it needs. A lot of that equipment is trucks... Note 9 shows vehicles went from $42 million at the end of 2025 to $126 million at September 30, 2026.

I walk through the same napkin math in the video starting at 7:03.

2. The income statement versus the cash flow statement

Here's something that trips up a lot of people. Aurora's net loss and its cash burn are not the same number, and the biggest reason is stock-based compensation, which basically just means paying employees with shares instead of cash.

On the income statement, stock-based compensation is buried inside the expense lines. Note 6 breaks it out... $43 million of the quarter's $54 million sits in research and development, and $11 million sits in selling, general and administrative.

On the cash flow statement, that same $160 million for the nine months gets added back, because no cash left the building. That's normal accounting under GAAP, and it's why the cash loss looks smaller than the accounting loss.

But I wouldn't call it free. The cost just shows up in a different place, which is your share count, and we'll get to that next.

There's also a cash cost hiding down in the financing section. Aurora paid $67 million in the first nine months of 2026 for "taxes on net settlement of equity incentive awards," which means it held back some vested shares from employees and paid their tax bill in cash instead. That money is real, and it doesn't show up in operating cash flow at all.

And here's a tell I liked finding. MD&A says operating cash use rose $119 million from last year "primarily due to the settlement of the annual bonus in cash rather than Class A common stock." Said another way, a bonus that used to be paid in stock got paid in cash this year, which makes the burn look worse but means fewer new shares. When the mix between cash pay and stock pay shifts, you have to compare years carefully.

One more thing. The income statement shows a $13 million non-cash gain in the quarter from the "change in fair value of derivative liabilities" (Note 7), which made the net loss smaller and has nothing to do with hauling freight.

3. Share count and the at-the-market program

Dilution basically just means your slice of the company gets smaller because the company issued more shares. If you own the same number of shares and the total goes up, you own less of the business.

The balance sheet shows 1,943 million shares outstanding at the end of 2025 and 2,015 million at September 30, 2026. That's 72 million more shares, or about 3.7% in nine months, by my math. The cover page shows 2,020,447,857 shares (Class A plus Class B) as of October 21, 2026, so it kept ticking up after quarter end.

Source: condensed consolidated statements of stockholders' equity and Note 5.

An at-the-market program, or ATM, basically just means the company sells new shares into the market a little at a time through banks, instead of doing one big offering. Note 5 says Aurora sold about 6 million shares in the third quarter at an average price of $6.56, and about 39 million shares over nine months at an average of $7.13.

A pre-profit company using an ATM is normal. What I'd notice is the pace... only $41 million came in during the third quarter while the business used about $206 million.

Now here's the part I'd flag for you. Part II, Item 5 says that, as of September 30, 2026, Aurora had sold 189,522,984 shares through the ATM since it started in February 2025, for gross proceeds of $1,174,785,034. On October 28, 2026, the same day the 10-Q was filed, Aurora amended the program, added Canaccord Genuity, Oppenheimer and Craig-Hallum as sales agents, and raised the amount available to $1,000,000,000. That includes $245,835,654 that was left over, so roughly $754 million of it is new capacity, by my math.

There's also a pile of shares that could show up later. Note 10 lists 212 million potential shares that were left out of the per-share math because the company lost money, including 109 million unvested restricted stock units (shares employees get once they vest) and 77 million stock options. The 21 million public and private warrants in that total have an $11.50 exercise price and were set to expire on November 3, 2026, according to Note 7 and Item 5, so I don't expect much from them.

I talk about the ATM at 7:46 and the share count at 8:02 in the video.

4. Going-concern and liquidity language

A going-concern warning is when management, or the auditor, says there's "substantial doubt" the company can keep operating for the next twelve months. It's the closest thing accounting has to a fire alarm.

There isn't one in this filing. Under "Risks and Uncertainties" in Note 2, management says cash and investments "will be sufficient" to meet its needs "for a period of at least twelve months from the date of these financial statements." MD&A repeats the same idea under Liquidity and Capital Resources.

That's the normal, clean version of the language. But there are two phrases right next to it that I'd read twice.

• Note 2 says the company "will need to opportunistically raise additional capital" to commercialize the Aurora Driver at scale. Need is a strong word.
• Both Note 2 and MD&A say management "will continue to evaluate the timing and nature of discretionary operating expenses." Said another way, if raising money gets hard, spending is the lever they'd pull.

Put that next to my 14-to-15-month runway math and the twelve-month test passes, but not by a huge margin if spending rises, which is why the ATM refill matters.

5. Risk factors, MD&A and controls

Part II, Item 1A says there have been "no material changes" to the risk factors in the 2025 10-K. That's normal for a 10-Q, and it means the full risk list lives in the annual report.

Item 4 says the CEO and CFO concluded disclosure controls were effective and there were no changes in internal control over financial reporting during the quarter. That's also normal, and it's what I'd expect to see.

In MD&A, two things stood out to me.

• Revenue is still tiny next to costs. Revenue was $3 million in the quarter and cost of revenue was $8 million, so Aurora spent more delivering freight than it earned for it.
• Costs got reshuffled. MD&A says some research and development costs now sit in selling, general and administrative "due to a realignment of resources." SG&A rose 29% in the quarter to $49 million. When costs move between lines, I'd be careful comparing growth rates year over year.

And here's one I'd actually ask about if I were still on the audit team. Note 7 shows the risk-free rate used to value the earnout shares rose to 5.1% at September 30, 2026, from 3.8% at the end of 2025, while MD&A says other income fell to $10 million from $16 million in the quarter because of "lower market rates." Those can both be true, because they're measured over different terms, but it's the kind of thing I'd want management to walk me through. The dollar amount is small, a $23 million liability, so this is a question, not an accusation.

6. The footnotes an auditor reads first

When I open a 10-Q, there are four footnotes I go looking for before anything else. Here's what Aurora's say.

Revenue recognition (Note 3). Revenue is what a company earns from selling its product, and the footnote tells you when it gets counted. Aurora books revenue "over time" as freight moves from origin to destination, priced per mile, and it uses the amount it has the right to invoice. That's a simple, normal policy, and with only $6 million of revenue in nine months, there isn't much room for games yet.

Commitments and contingencies (Note 11). This note only covers legal claims, and it says no material loss contingencies were recorded. What I'd notice is what's not there... Note 2 says Aurora plans to rely on a single supplier, AUMOVIO, for its next-generation hardware, but this 10-Q doesn't put a dollar amount on purchase commitments. I'd go back to the 10-K for that.

Related parties. There's no related-party footnote in this 10-Q. That's not unusual for a quarterly report, but it means you'd check the 10-K and the proxy statement if that matters to you. The only insider item is in Item 5, where one director adopted a trading plan on September 30, 2026 to sell up to 50,000 shares, which is small next to two billion shares outstanding.

Subsequent events. These are things that happened after the quarter ended but before the filing. Aurora doesn't have a separate note for them, but Item 5 covers two: the warrants expiring on November 3, 2026, and the October 28 ATM amendment. The ATM one is the bigger deal, for the reasons in section 3.

One small thing I'd tick and tie, which is auditor-speak for matching a number in one place to the same number somewhere else. The segment table in Note 12 shows nine-month cost of revenue of $20 million, while the income statement shows $21 million. The filing says amounts may not sum due to rounding, and that's almost certainly what this is, but it's exactly the kind of difference I was trained to circle.

7. What I'd watch next quarter

So now that we've been through the filing, here's my short list for the next 10-K, which covers the fourth quarter and the full year.

• Cash and investments versus $1,034 million. If the company really uses around $250 million in the fourth quarter, you'd expect something under $800 million before any new share sales.
• How much of the refilled ATM gets used, and at what average price compared with $6.56 in the third quarter. A lower price means more shares for the same dollars.
• The share count versus 2,015 million. I'd want to know how much growth came from the ATM and how much came from employee stock.
• Whether the twelve-month liquidity sentence changes at all. Watch for any new words in that paragraph, because auditors and lawyers argue over every one of them.
• Purchase commitments and the AUMOVIO relationship in the annual footnotes, which have more room for detail than a 10-Q.

The honest caveat

A 10-Q only shows you the money side. Aurora's whole bet is that driverless trucks eventually earn a lot more per mile than they cost, and nothing in this filing can prove or disprove that yet.

This page is education, not advice. I'm a CPA, but I'm not your CPA, and nothing here is a recommendation to buy or sell anything. You can read more in the disclaimer.

If you read the filing yourself and see something I missed, let me know in the comments on the video. Thanks for reading all the way to the end.