Nebius Up 125%: Why Burry's Short Sees a Bull Run Running Into a Wall of Spending

The 125% move shifted Nebius from concept trade to execution trade

After a sharp run earlier this year, Nebius is now being judged on delivery rather than possibility. A Microsoft-led agreement then sent shares soaring over 47% after the bell, forcing investors to confront a simpler question: does this demand translate into usable capacity, revenue, and margins?

Bulls see proof. A hyperscaler committing to a long-duration AI infrastructure deal is a much stronger signal than another speculative growth narrative. Bears, meanwhile, are focused on the harder part of the story: whether Nebius can absorb heavy spending without letting margins deteriorate before revenue fully ramps.

Microsoft's contract gives the bull case real substance

A long-duration demand signal matters

The core bullish argument is no longer abstract. The Microsoft agreement carries $17.4 billion over a five-year term, with potential added services pushing the total to about $19.4 billion. Nebius also said it will provide Microsoft with dedicated GPU infrastructure capacity from its New Jersey data center starting later this year. That gives the story more substance than a pure AI hype cycle.

Demand exists, but utilization is what matters

What bulls are really buying is not just headline contract value, but better visibility into whether Nebius can keep capacity occupied. In AI infrastructure, owning hardware is only part of the equation; the more important test is how quickly that hardware turns into revenue.

That context matters because Nebius already has signs of real customer absorption, including a long-term deal with Meta to provide up to $27 billion worth of computing capacity over five years. Together, those commitments suggest Nebius is not merely building compute for show. It is trying to fill it.

Revenue is showing up, but the timing debate remains

Bears still have a legitimate objection: heavy infrastructure spend can damage margins before revenue catches up. Nebius' first-quarter capex jumped to about $2.5 billion versus $544 million a year ago, and analysts have flagged pressure on margins despite strong revenue growth.

Still, revenue is already visible. First-quarter revenue reached $399 million, beating an estimate of $371.4 million. Pair that with management's view that the Microsoft agreement will accelerate growth, and the story looks less like a meme-driven spike and more like an aggressive infrastructure buildout with real customer demand behind it.

The remaining debate is timing. If utilization tracks those contracts, Nebius may earn a more durable valuation. If spending outruns occupancy, the multiple can compress quickly.

The short case is about margins, not demand

The risk is not a lack of customers

The bull case has a clear end product: $17.4 billion over a five-year term from Microsoft, potentially about $19.4 billion in total. The short case, though, lives in the middle mile between signed contracts and earnings conversion. Nebius already showed what that bridge can cost: first-quarter capex jumped to about $2.5 billion from $544 million a year ago.

In this model, the income statement does not wait for perfection. If deployment slips, utilization lags, or margin conversion weakens, the market can turn on the stock quickly.

Heavy spending can still overwhelm strong revenue growth

Reuters noted that analysts are flagging Nebius' heavy capital spending as a concern because it is putting pressure on margins despite strong revenue growth. The company also paused its share buyback program in late 2024 to redirect capital toward expanding its core AI infrastructure business. That is not a free pass. It is a bet that future utilization will outrun present strain.

What would weaken the bull or bear case

A clean bull confirmation would be stable or improving margins during the ramp, along with evidence that capacity tied to major customers is being used rather than simply ordered.

A clean bear confirmation would be another round of spending jumps without corresponding signs that contracted demand is becoming productive capacity. Until that bridge is clearer, the rally remains vulnerable to the oldest problem in infrastructure investing: paying today for profits that arrive too late.