Michael Burry's Nebius Short Is Underwater After a 454% Revenue Quarter

On Aug. 6, Michael Burry disclosed a short position in Nebius Group (NASDAQ: NBIS) at around $212 a share, a bet he described as larger than the one he placed against Oracle at the same time. Six days later, the artificial intelligence (AI) cloud company reported second-quarter results before the market opened on Wednesday.

The report did not go his way. Revenue rose 454% year over year to $582.3 million, and management reaffirmed its full-year guidance. The stock jumped about 34% to around $259 as of this writing -- about 22% above the price he disclosed shorting at.

But a losing first week doesn't settle whether he's right. The more useful question, I'd argue, is which of Burry's concerns the quarter actually answered, and which one it left standing.

Image source: Getty Images.

A quarter with little to argue with

The demand side of the short case looks weaker today than it did on Aug. 6.

Nebius' AI cloud business grew revenue 514% year over year to $575 million. And its annualized run rate (the last month of the quarter's AI cloud revenue multiplied by 12) reached $3.0 billion. That figure was $1.9 billion at the end of March and $1.25 billion at the end of December, so the dollar gains are getting bigger each quarter, not smaller.

The economics improved alongside the growth, too. The AI cloud unit's adjusted earnings before interest, taxes, depreciation, and amortization (EBITDA) margin reached 50%, up from 45% in the first quarter and 24% in the fourth quarter of 2025.

What's more, the deals behind the growth are getting stronger. Nebius closed four contracts in the quarter averaging more than $1 billion each in total value, and about 70% of the deals it closed in the quarter included customer prepayments covering 50% to 60% of the related capital spending.

"We could sell our entire 2027 capacity on these terms today," founder and CEO Arkady Volozh wrote in his quarterly letter to shareholders.

And the funding picture, a core worry for any company building data centers this fast, held up as well. Operating cash flow came in at $2.2 billion, and the company was sitting on $8 billion of cash at the end of June.

Burry's sharpest argument is about depreciation

Burry's case against AI infrastructure stocks was never really about demand. In a widely covered post last November, he argued that companies in the space are depreciating chips too slowly, stretching the accounting life of hardware that turns over on a two-to-three-year product cycle and overstating profits as a result. He put Oracle's potential overstatement at about 27% by 2028.

His Aug. 6 disclosure pointed at a related worry: obligations, such as leases that haven't started yet and purchase commitments, that sit off the balance sheet.

Nebius' own report shows why that argument is hard to dismiss. Depreciation and amortization came in at $259.7 million in the second quarter -- 45% of revenue, and more than the company's entire adjusted EBITDA of $236.2 million. On a GAAP basis, Nebius lost $190.4 million in the quarter. And the company depreciates its servers and network equipment over five years -- a life it extended from four years at the start of 2026, citing usage patterns and utilization commitments. Extending useful lives is precisely the move Burry says flatters the industry's profits.

The bill is still growing. Nebius spent $5.7 billion on capital expenditures in the second quarter alone. If all of that spending were depreciated over five years, it would imply more than $1.1 billion of annual depreciation once the equipment is in service -- more than the company's current adjusted EBITDA, annualized.

An expensive stock either way

None of this is to say the short is a good bet. Demand is contracted, often prepaid, and growing, and the company keeps hitting its own targets. Betting against a business executing this well could prove expensive, whatever the accounting eventually shows.

But the price asks a lot. At around $259, Nebius trades at about 20 times this year's guided revenue of $3.0 billion to $3.4 billion -- a steep multiple even if the growth lands as planned. And shareholders are funding the build-out in dilution as well as dollars. The company sold 12.7 million shares through June 30 at an average price of $223.60, raising about $2.8 billion, with 12.3 million shares still available under the program.

To me, the quarter refuted the easy version of the short case and left the hard version intact. Nebius likely proved the demand, and it arguably proved the funding. What it hasn't proved is that the growth makes money once the full cost of the hardware lands on the income statement.

Daniel Sparks and his clients do not have positions in any of the stocks mentioned. The Motley Fool has positions in and recommends Oracle. The Motley Fool has a disclosure policy.