The artificial intelligence boom has created a new kind of gold rush. Companies are racing to build models, train algorithms, and deploy AI tools across industries. But behind that race sits a less visible battle: the scramble for computing power.
One company suddenly sitting near the center of that story is Nebius (NASDAQ:NBIS). The firm operates large-scale AI cloud infrastructure designed to train and run advanced AI models. While the headlines often focus on chip makers or AI labs, Nebius is positioning itself as a supplier of the infrastructure those companies depend on.
That role became clearer after the company secured a five-year agreement worth roughly $27 billion with Meta to supply AI computing capacity. The deal includes about $12 billion of dedicated capacity, with Meta retaining the option to purchase an additional $15 billion over the contract period.
Meta’s agreement is not happening in isolation. Nebius previously struck a $17.4 billion infrastructure deal with Microsoft, and Nvidia has invested $2 billion in the company. Together these developments point to a larger shift in the AI ecosystem. The focus is expanding beyond chips and models toward the companies capable of building the infrastructure that powers them.
The story unfolding around Nebius highlights four forces shaping the next phase of the AI economy.
Exploding AI Demand & A Global Shortage Of Compute Capacity
One of the most striking signals from Nebius’ recent earnings call is how quickly demand for AI computing is accelerating. Management stated that the company sold out its available capacity in both the third and fourth quarters of 2025. The same situation continued into early 2026.
In many cases, capacity is already committed before the infrastructure is fully deployed. That reflects a broader imbalance between demand for AI computing and the pace at which data-center infrastructure can be built.
Another important trend is the scale at which customers are expanding. AI startups that once required hundreds of GPUs are now requesting thousands or even tens of thousands. These workloads were historically associated with the largest frontier-model developers. The shift suggests that AI capabilities are spreading quickly across industries and companies of different sizes.
Large enterprises are also becoming heavier users of AI infrastructure. Businesses are increasingly integrating AI into routine processes such as software development, research, content production, and operational analytics. As those workloads expand, they require greater amounts of computing power and longer-term contracts with infrastructure providers.
This dynamic has created an environment where compute capacity has become a scarce resource. Companies seeking to train and deploy AI systems are increasingly willing to sign multi-year agreements to secure access to that capacity.
Nebius Is Building Massive Infrastructure To Capture The AI Wave
To meet this surge in demand, Nebius is undertaking a rapid expansion of its global infrastructure footprint. The company recently announced plans for nine new data centers and is accelerating its power procurement strategy to support future growth.
Power capacity has become a key metric in the AI infrastructure race. Nebius reported that it has already secured more than two gigawatts of contracted power and is targeting over three gigawatts by the end of 2026. That level of capacity can support very large GPU clusters designed to train and run AI models.
The company also expects to deploy 800 megawatts to one gigawatt of operational data-center capacity over the same period. These facilities will host the computing hardware used by customers ranging from AI startups to large technology firms.
The scale of investment required for this expansion is significant. Nebius expects capital expenditures between $16 billion and $20 billion in 2026 alone. Those funds will be used to secure power infrastructure, construct data centers, and deploy GPUs.
A large portion of this spending will be funded by long-term contracts already signed with customers. Management noted that upfront payments from those agreements are providing meaningful cash flow visibility. The company ended 2025 with $3.7 billion in cash and equivalents, providing additional financial flexibility.
Taken together, these investments illustrate how AI infrastructure is evolving into one of the largest capital-intensive sectors in technology.
Strategic Partnerships Are Turning Nebius Into An AI Infrastructure Hub
Nebius’ recent partnerships highlight how important infrastructure providers have become within the AI ecosystem. The company is not only selling computing capacity but also integrating itself into the supply chain of several major technology companies.
The $27 billion Meta agreement is one of the most prominent examples. Under the deal, Nebius will deploy large-scale AI infrastructure using Nvidia’s next-generation Vera Rubin platform, expected to become available later this year. The computing capacity will begin delivering to Meta starting in 2026.
This agreement follows the earlier $17.4 billion contract with Microsoft, which will ramp throughout 2026 and is expected to contribute fully to revenue beginning in 2027. These deals illustrate how hyperscalers are increasingly securing infrastructure capacity through partnerships rather than building everything internally.
Nvidia’s $2 billion investment in Nebius further reinforces the company’s role within the AI infrastructure landscape. The collaboration allows Nebius to gain early access to new computing technologies while working with Nvidia on large-scale data-center deployments.
Beyond infrastructure, Nebius is also expanding its AI platform capabilities. The company recently acquired Tavily, an agentic search technology that connects AI agents to web data. The platform already serves many enterprise clients and brings hundreds of thousands of developers into Nebius’ ecosystem.
These partnerships and acquisitions suggest that Nebius is attempting to build a broader AI cloud platform rather than operating purely as a data-center provider.
Rapid Financial Growth & A Market Valuation Reflecting High Expectations
Nebius’ financial performance reflects the pace at which demand for AI infrastructure has expanded. The company reported fourth-quarter revenue of $228 million, representing year-over-year growth of 547%. Its core AI cloud business grew even faster, posting 830% annual growth.
Annualized run-rate revenue reached $1.2 billion at the end of 2025, exceeding previous company guidance. Management expects full-year revenue between $3 billion and $3.4 billion in 2026, with an annualized run rate projected to reach $7 billion to $9 billion by the end of that year.
Margins are also expected to improve as the infrastructure scales. Nebius reported positive adjusted EBITDA in the fourth quarter and guided toward approximately 40% adjusted EBITDA margins in 2026. The company still expects operating losses as it continues investing heavily in infrastructure expansion and research.
Despite this growth, the company’s valuation multiples reflect how early the business remains in its expansion cycle. Based on recent data, Nebius trades around 64.3× LTM enterprise value to revenue and roughly 62× price-to-sales. Profit-based multiples remain difficult to interpret because the company is still scaling its infrastructure and generating limited earnings.
Forward metrics are somewhat lower, with the market pricing Nebius at roughly 10.2× next-twelve-month EV-to-revenue and about 24× forward EV-to-EBITDA based on consensus expectations.
These valuation levels indicate that investors are already assigning significant value to the company’s potential role in the AI infrastructure market.
Final Thoughts
The rapid rise of Nebius highlights how the AI boom is expanding beyond algorithms and semiconductor chips. The infrastructure required to train and operate AI systems is becoming a major industry in its own right.
Contracts with companies such as Meta and Microsoft illustrate how large technology firms are securing computing capacity years in advance. At the same time, demand from AI startups and enterprises continues to increase as AI adoption spreads across sectors.
Nebius’ aggressive infrastructure investments and strategic partnerships suggest it aims to position itself as a key supplier of AI computing resources. The company is scaling rapidly, but it remains in an early phase of development relative to the size of the market opportunity.
From a valuation perspective, Nebius trades at elevated trailing revenue multiples that reflect high expectations around future growth. Forward valuation metrics appear more moderate as projected revenue expands, though profitability remains tied to the pace of infrastructure deployment and capital investment.
The company’s trajectory illustrates how critical computing infrastructure has become to the AI ecosystem. Whether Nebius ultimately becomes a long-term cornerstone of that ecosystem will depend on its ability to scale capacity efficiently while maintaining demand from large technology customers and emerging AI developers.
Disclaimer: We do not hold any positions in the above stock(s). Read our full disclaimer here.
