Nebius Group has rapidly become a cornerstone of AI infrastructure, locking in $24.4 billion in capacity agreements with Microsoft and Meta and securing a $2 billion strategic investment from Nvidia. With revenue projected to surge from $90 million to $9 billion in two years, this neocloud specialist offers a rare blend of high-growth potential and valuation discount versus peers. But can it execute at scale?
The artificial intelligence arms race is no longer just about algorithms and models; it’s about the raw compute power needed to train and run them. Enter Nebius Group, a neocloud provider that has catapulted from obscurity to a central role in AI infrastructure, inking deals that total tens of billions of dollars with the world’s leading AI developers.
Over the past 18 months, Nebius has achieved what many infrastructure startups can only dream of: validated partnerships with three of the most influential players in tech. A $19.4 billion capacity agreement with Microsoft for Azure, a $3 billion deal with Meta Platforms, and a $2 billion equity investment from Nvidia itself. This trifecta of validation signals a structural shift in how AI hyperscalers source their computing needs.
What Is a Neocloud, and Why Does It Matter?
Traditional cloud providers like AWS, Azure, and GCP built their platforms for general-purpose computing. Neoclouds like Nebius are architected from the ground up for the unique demands of AI training and inference. This means optimized server racks, custom-built data centers, and cost-efficient access to GPU clusters paired with integrated software services.
As AI models grow exponentially more complex, the need for specialized infrastructure has become acute. Hyperscalers are outsourcing massive chunks of their capacity needs to players like Nebius because the pace of internal data center construction cannot keep up with insatiable demand for compute.
The Hyperscaler Deals: A Validation of the Model
The sheer scale of Nebius’s recent wins is staggering. The Microsoft deal, running through 2031, involves providing GPU infrastructure capacity at Nebius’s Vineland, New Jersey facility. This is not a minor pilot; it’s a long-term capacity agreement that locks in future revenue and demonstrates trust at the highest levels.
Shortly after, Meta signed a five-year, $3 billion AI infrastructure pact. For context, these are companies that typically build their own data centers. Their decision to partner externally underscores a critical industry bottleneck: even with virtually unlimited capital, scaling physical infrastructure is slow, complex, and capital-intensive.
By outsourcing to Nebius, Microsoft and Meta can reallocate substantial capital expenditure budgets away from construction and chip procurement while still guaranteeing capacity for their AI models. This allows them to focus resources on software and model development, where their core competitive advantages lie.
Nvidia’s $2 Billion Bet: More Than Just Capital
The most strategic signal came from Nvidia. The chipmaker’s $2 billion investment in Nebius is a clear endorsement of the neocloud model and complements its existing stake in another leader, CoreWeave. This partnership likely grants Nebius preferred access to Nvidia’s next-generation Rubin and Blackwell Ultra GPU architectures, ensuring its ability to offer cutting-edge capacity to customers.
This relationship is symbiotic. Nvidia needs guaranteed, high-volume buyers for its GPUs to sustain its growth, and Nebius provides that channel. The investment was reported by AOL Finance, confirming the strategic alignment between the two companies.
Financial Trajectory: From $90 Million to $9 Billion
Nebius’s financials tell a story of hypergrowth. Two years ago, the company’s annual recurring revenue (ARR) was a modest $90 million. For 2026, Nebius is guiding for an ARR run rate between $7 billion and $9 billion. This represents a near-100x increase in just 24 months, a growth curve rarely seen in the capital-intensive infrastructure space.
The stock market has taken notice. Over the past 12 months, Nebius shares have surged 326%, reflecting investor bets on the AI infrastructure spend wave. However, this momentum raises a key question: does the valuation justify the growth?
Valuation: A Discount to CoreWeave?
Using the midpoint of its 2026 ARR guidance ($8 billion), Nebius trades at an implied market cap-to-ARR multiple of approximately 3.5. This serves as a proxy for the price-to-sales ratio.
Why does this matter? Because its primary competitor, CoreWeave, trades at a price-to-sales multiple of 6.4, despite being a significantly larger company. While CoreWeave has its own strengths, including an earlier start and broader customer base, Nebius’s valuation gap seems wide, especially given its landmark contracts with Microsoft and Meta provide substantial revenue visibility.
Investors are pricing in execution risk—the challenge of building out data centers at the promised scale. Yet, the long-term, multibillion-dollar agreements with hyperscalers suggest Nebius has already cleared the biggest hurdle: proving it can win business from the very companies that once might have been considered competitors.
The Investment Thesis for 2026 and Beyond
The bull case for Nebius hinges on several interconnected factors:
- Unprecedented Partnerships: Deals with Microsoft, Meta, and Nvidia create a powerful moat and revenue foundation.
- Structural Industry Shift: Outsourcing of AI infrastructure is a secular trend, not a temporary fix.
- Valuation Dislocation: Trading at a discount to CoreWeave on similar growth metrics presents a potential arbitrage opportunity if Nebius executes.
- Full-Stack Control: Unlike pure-play GPU lessors, Nebius’s integrated approach from facilities to software may yield higher margins over time.
For investors seeking exposure to the AI hardware boom without betting on a single chip designer or cloud giant, Nebius offers a diversified play on the entire infrastructure stack.
Key Risks Investors Must Consider
The opportunity is compelling, but risks are significant:
- Execution Risk: Scaling from $90 million to $9 billion in ARR requires flawless capital allocation and construction management. Delays or cost overruns could derail margins.
- Concentration: Reliance on a few large hyperscalers means losing one could have an outsized impact. However, the long-term contracts mitigate this somewhat.
- Competition: CoreWeave is a formidable rival, and giants like AWS and Google could eventually compete more aggressively in specialized AI clouds.
- Technology Obsolescence: Rapid GPU iteration means Nebius must continually reinvest to stay current, a capital-intensive endeavor.
Why Nebius Could Be a Buy for 2026
Despite the 326% run-up in the stock, Nebius may still be in the early innings of its growth story. The company’s ability to secure capacity agreements with entities that have their own massive data center ambitions is the highest form of validation.
For investors with a multi-year horizon, Nebius represents a rare chance to own a company at the epicenter of AI infrastructure spending—a trend that shows no sign of slowing as models become more compute-intensive. The valuation discount to CoreWeave, combined with the strategic Nvidia partnership, creates a compelling risk-reward profile for those who believe in the neocloud thesis.
Bottom line: Nebius is not without risk, but its contracts transform it from a speculative startup into a company with tangible, contracted revenue visibility. In an AI landscape where access to compute is the ultimate bottleneck, Nebius is building the pipes. For investors looking beyond the headline names in AI, this stock warrants a deep dive.
Stay tuned to onlytrustedinfo.com for the fastest, most authoritative analysis on breaking financial news and market-moving developments. Our team delivers actionable insights that cut through the noise, helping you navigate complex investment landscapes with confidence. Explore our latest coverage to stay ahead of the curve.
