Nebius Stock: $40 Billion in Contracts, 20% Short Interest, and the Capex Question in Between
Published on 10/25/2026 at 20:11 | Editorial boerse-global.de
Short sellers have been building their positions in Nebius, and that growing skepticism may say more about the company's prospects than its latest expansion headlines. Reuters reported Thursday that bearish bets now account for 20% of the free float, up from 15.4% at the start of 2026 — a climb the news agency placed within a broader sector selloff driven by doubts over the sustainability of debt-financed AI infrastructure.
That doesn't prove the skeptics are right. But it does move the goalposts: fresh capacity and new partnerships on their own are unlikely to be enough to quiet the economic concerns. What matters is what Nebius does with the opportunities it has bought.
Contracts Are a Foundation, Not a Return Guarantee
The company said Thursday it has signed long-term contracts worth more than $40 billion with investment-grade customers. That is a weighty counterargument to the idea of capacity being built without matching demand — long-term agreements deserve more trust than mere expectations of a growing AI market.
Even so, investors should not confuse contract volume with economic success. A contract is, first and foremost, a promise of future business. In my view, any valuation must also weigh the capital required to fulfill those contracts and the returns that capital can generate. That is precisely where Nebius faces its decisive test.
The company widened its financial room for maneuver on Thursday with a $3.5 billion senior secured financing facility from Citi. It runs for 364 days, and the proceeds are earmarked for data center development, GPU procurement and general corporate purposes. I read the facility neither as an automatic seal of quality nor as evidence of problematic expansion. It creates flexibility, but it also makes the financing terms a key part of the growth equation.
Should investors sell immediately? Or is it worth buying Nebius?
Pricing starts at Term SOFR plus 2.75 percentage points a year, with the spread rising to a maximum of 3.50 points. Alongside operational execution, then, financing belongs among the yardsticks investors should use to judge the buildout. Availability of capital and economic quality are two different things — and with a maturity this short, that distinction deserves more attention than the size of the credit line.
Applications Must Justify the Buildout
Nebius isn't betting on infrastructure alone. On Wednesday the company, together with NVIDIA, unveiled the open-source Physical AI Workbench, a robotics development framework spanning data generation, simulation, training and evaluation. To my mind this approach is strategically more interesting than another pure capacity announcement: it addresses the question of what customers actually do with the compute.
According to the company, customers report development time cut by up to 40%. That describes a potential user benefit, not Nebius's profitability — yet it is relevant, because a demonstrable productivity gain is a better argument for AI spending than general hopes for technological upheaval. The collaboration announced Wednesday by Sereact adds another concrete use case: together with NVIDIA, Nebius, CEVA Logistics and Zalando, a closed training loop is planned for Sereact's Cortex robot control system, aimed at robot-assisted returns processing at CEVA Logistics sites for Zalando.
Such projects make the demand story more tangible, but they are not yet proof that the entire infrastructure buildout pays off economically. I would treat them as operational progress, not as financial success booked in advance. The same caution applies to the preliminary framework agreement announced Wednesday with HUMAIN and DataVolt, covering up to 50 MW at the planned Oxagon AI campus in Saudi Arabia, with capacity to become available in stages from the fourth quarter of 2027. The word "preliminary" is what counts here: it opens up prospects rather than delivering available capacity, making it an expansion option rather than an argument that dispels today's return concerns.
The Analyst Critique Hits the Real Issue
The analyst side also shifted its yardstick on Wednesday. According to media reports, Piper Sandler cut its Nebius price target to $185 from $237 while keeping a "Neutral" rating. Analyst James Fish cited sector-wide concerns about AI demand, return on invested capital, free cash flow and financing clarity, and the firm lowered its long-term valuation multiples across the peer group.
That reasoning matters more than the target itself. It isn't a mere mood swing but an economic test: growth only earns a high valuation if the expected return justifies the capital required to produce it, and the size of an expansion program is no substitute for that. The same logic sits behind the Reuters report on worries that debt-financed AI infrastructure investment may not be durably sustainable.
My verdict is therefore sober. Nebius has solid arguments for its expansion in long-term customer contracts and concrete applications. A more attractive growth story alone is not enough. For a convincing valuation, the economic benefit of the buildout has to move to the foreground. More capacity is an intermediate step — return on capital remains the decisive touchstone.
Ad
Nebius Stock: New Analysis - 25 October
Fresh Nebius information released. What's the impact for investors? Our latest independent report examines recent figures and market trends.
