Leopold Aschenbrenner is not an ordinary hedge fund manager. The former OpenAI researcher has built a fund, Situational Awareness, that now manages roughly $13.7 billion per its latest 13F filing.
According to a recent 13G filing with the Securities and Exchange Commission (SEC), Situational Awareness recently bought 12.4 million shares of Nebius Group (NBIS +2.10%) -- representing a 5.6% ownership stake in the neocloud company.
Aschenbrenner's Nebius position is not a random allocation. Rather, it fits a meticulously constructed artificial intelligence (AI) infrastructure thesis that's on display across his fund's entire book.
Image source: The Motley Fool.
Why did Leopold Aschenbrenner just buy Nebius stock?
Situational Awareness is aggressively long the physical infrastructure layer of AI data center build-outs. Specifically, Aschenbrenner's portfolio includes meaningful positions in competing neoclouds Iren (IREN 0.87%) and CoreWeave (CRWV +2.57%).
Iren is former Bitcoin miner pivoting its infrastructure to help manage AI workloads. Back in November, Iren secured a $9.7 billion capacity deal with Microsoft. More recently, the company also inked $5.5 billion worth of deals with Nvidia.
Meanwhile, CoreWeave boasts a backlog exceeding $99 billion and works with hyperscalers including OpenAI, Anthropic, and Meta Platforms.
Adding Nebius completes the pure-play neocloud trifecta, each riding the same secular infrastructure demand tailwinds.
What are neoclouds and why do they matter for AI?
A neocloud is a cloud infrastructure provider serving GPU-intensive AI workloads, including model training, inference deployments, and agentic systems at scale. Unlike hyperscalers such as Amazon Web Services (AWS), Microsoft Azure, or Google Cloud Platform (GCP), which offer bundled services across compute, database management, data analytics, and storage, neoclouds focus purely on leasing access to GPU hardware designed specifically for managing AI applications.
The neocloud business model is increasingly important because hyperscalers are struggling to provision enough GPU capacity to meet demand from premier AI labs and large enterprises. Given some of the deals referenced above, it's clear that even the hyperscalers are starting to outsource GPU capacity to neoclouds rather than compete with them.
NASDAQ: NBIS
Key Data Points
Should you buy Nebius stock right now?
Nebius reported revenue of $399 million during the first quarter, up 684% year over year. According to management, the company is on pace to achieve between $7 billion and $9 billion in annualized run rate revenue by year-end.
Given the company's current market capitalization of $58 billion, this would imply a forward price-to-sales (P/S) multiple of roughly 6.4 at the high end of Nebius' ARR guidance. I do not see this as unreasonable for an AI infrastructure business growing at this pace with contracted, multiyear agreements from some of the largest AI hyperscalers.
With that said, Nebius' stock has already skyrocketed by more than 176% in 2026. If management delivers on its ARR target, investors could be buying into a hypergrowth story at a fairly reasonable price point relative to the frothy valuations seen during prior years of the AI revolution.
The risk, however, is that current momentum could be overextended. Investors following Aschenbrenner into Nebius are arguably paying a premium for a story the market already knows -- all at a price that leaves little room for execution error. While I like Nebius as a long-term AI infrastructure opportunity, I think there will be better entry points for the stock.
