Nebius Stock Is Up 170% in 2026, and Leopold Aschenbrenner Just Bought a 5.6% Stake. Here's Why

Nebius has rallied nearly 170% this year, but the article argues the stock could still benefit from explosive AI infrastructure demand, with revenue up 351% to $530 million and analysts projecting a 242% CAGR to $21.2 billion by 2028. Leopold Aschenbrenner’s Situational Awareness fund disclosed a 5.6% stake, reinforcing the investment case around neocloud infrastructure bottlenecks such as data centers, chips, and power. The piece is broadly supportive of Nebius’ long-term thesis, though it also flags valuation risk at 19x this year’s sales.

Analysis

The key signal is not the headline stake itself, but that a sophisticated AI-capex investor is validating neoclouds as a separate scarcity trade from software AI. If he is right, NBIS is less about near-term revenue and more about being a levered claim on the bottleneck layer: GPU availability, power, and deployment capacity. That also means the trade can rerate faster than fundamentals if hyperscaler/AI-lab capex stays elevated into the next earnings cycle, because the market will increasingly price NBIS on backlog visibility rather than current margins.

Second-order winners are the infrastructure enablers around the neocloud stack: NVDA should retain pricing power longer if dedicated AI clouds keep absorbing top-tier GPUs, while MSFT and META may be signaling that they prefer external capacity for bursts rather than owning every watt of their own buildout. The risk is that this becomes a crowded “scarcity premium” trade just as capacity starts catching up, which would compress multiple faster than revenue growth alone can offset. CRWV is the cleaner read-through on the sector than AMZN, because AWS is a slower-moving incumbent that can defend on breadth, not speed.

The contrarian miss is that the market may be overestimating how durable premium pricing is for neoclouds once customers standardize workloads and renegotiate after initial deployment. NBIS can still work, but the path is likely volatile: strong upside over 3-6 months if AI capex remains extreme, then sharper downside if there is any evidence of deferred orders, pricing pressure, or dilution to fund expansion. In this setup, the stock is more attractive as an event-driven momentum vehicle than a clean long-duration compounder at 19x sales.

Key Decisions for Investors

• Go long NBIS on weakness, but only with a 3-6 month horizon; use a staggered entry around post-earnings or post-rally pullbacks, since the stock is likely to trade on backlog/capacity headlines rather than current earnings quality.
• Pair long NBIS / short AMZN or a basket of legacy cloud names for a relative-value expression on AI infrastructure scarcity, with the thesis that specialized capacity should outgrow general-purpose cloud over the next 2-4 quarters.
• Overweight NVDA versus CRWV if you want the cleaner second-order beneficiary; dedicated AI clouds improve GPU utilization and reinforce NVDA’s ecosystem leverage without taking direct execution risk on neocloud margins.
• For risk-managed upside exposure, buy NBIS call spreads 3-6 months out rather than stock; the implied volatility should stay elevated, and defined-risk options are better suited to a name where multiple compression can be violent if capex sentiment turns.
• Trim or avoid adding after any sharp multi-week continuation move unless there is evidence of new contract wins; the main failure mode is not a bad quarter but the market deciding the scarcity premium is already fully priced.