Nebius Just Changed the AI Race—Here’s Why
Nebius is suddenly at the center of the AI infrastructure rush. The company delivered stronger-than-expected second-quarter revenue, signed several huge AI cloud contracts and raised its target for contracted computing power as customers scramble to secure capacity.
The bigger story, however, may not be the earnings beat itself. It is what those numbers reveal about the increasingly intense race for the infrastructure behind artificial intelligence.
Nebius Beats Expectations as AI Demand Accelerates
Nebius reported $582.3 million in revenue for the second quarter of 2026, up sharply from $105.1 million a year earlier. The result exceeded analyst expectations and highlighted how quickly demand for AI computing infrastructure is expanding.
The company’s AI cloud business was the main growth engine, with revenue from the unit rising nearly sixfold year over year.
That growth comes at a time when AI companies are competing for access to increasingly expensive computing resources. Training and running advanced AI models requires massive amounts of GPUs, data-center capacity, networking infrastructure and electricity.
For Nebius, that shortage is becoming a business opportunity.
The company is not simply selling cloud services. It is increasingly selling something customers urgently need: access to AI computing capacity.
The Numbers Behind the Surprise
Nebius’ second-quarter results showed just how quickly the business is scaling.
Revenue reached $582.3 million, while adjusted EBITDA rose to $236.2 million, compared with an adjusted EBITDA loss of $21 million in the same quarter last year.
That is a significant shift for a company spending aggressively to build out AI infrastructure.
Nebius also generated more than $2.2 billion in operating cash flow during the quarter, although its spending on property, equipment and intangible assets reached approximately $5.7 billion.
In other words, the company is simultaneously showing stronger financial performance and spending heavily to capture the next wave of AI demand.
That balance will be one of the most important things investors watch going forward.
Why Customers Are Racing for AI Capacity
The most revealing part of the update may be the size of the contracts Nebius is winning.
The company closed four deals averaging more than $1 billion in total contract value each during the quarter. Total contract value nearly quadrupled from the previous quarter, while the value of contracts with new customers increased more than ninefold.
That suggests AI infrastructure demand is moving beyond experimental projects.
Companies building AI models and AI-powered products need predictable access to computing resources. Waiting months for new capacity can become a serious competitive disadvantage.
For customers, locking in capacity today can therefore be a strategic decision rather than simply another cloud purchase.
And that is where Nebius believes it can benefit.
Nebius Raises Its AI Infrastructure Target
Nebius raised its target for contracted power at the end of 2026 to 5 gigawatts, up from a previous target of more than 4 gigawatts.
The company also said it expects to deploy more than 1 GW of capacity annually beginning in 2027.
Those numbers illustrate the scale of the AI infrastructure race.
A gigawatt is an enormous amount of power. Building infrastructure capable of supporting that level of computing requires data centers, electricity connections, cooling systems, networking equipment and huge quantities of advanced GPUs.
The challenge is not simply buying chips.
Companies must secure land, power and physical infrastructure and then bring that capacity online quickly enough to satisfy customers.
That is becoming one of the biggest bottlenecks in the AI industry.
The $40 Billion Number That Stands Out
Nebius said it has more than $40 billion in customer commitments and expects customer prepayments to exceed $9 billion this year.
Those commitments are important because they provide visibility into future demand.
They also show why AI infrastructure companies are willing to spend billions today.
If customers are prepared to commit significant amounts of money for future capacity, infrastructure providers have a stronger incentive to accelerate data-center construction and GPU deployment.
But commitments are not the same as guaranteed future revenue.
Nebius still has to build the capacity, secure the necessary power and equipment, and deliver the infrastructure on schedule.
That execution risk should not be ignored.
Nebius Says It Could Sell Out 2027 Capacity
Perhaps the most striking signal from the latest update is Nebius’ assessment of future demand.
The company said it could sell all of its planned 2027 capacity at current terms, but it is holding some capacity back to potentially capture future demand.
That suggests the company believes the AI computing shortage could remain significant even as more infrastructure comes online.
For the broader AI market, that is a notable development.
The conversation around artificial intelligence has increasingly shifted from simply asking which company will build the best model to asking a more fundamental question:
Who will have enough computing power to run those models at scale?
Nebius is positioning itself directly in that race.
Why the Market Reacted So Quickly
Investors appeared to respond strongly to the combination of stronger revenue, large customer contracts and increased infrastructure targets.
Nebius shares were reported more than 15% higher in premarket trading following the results.
The reaction reflects more than a single quarterly revenue beat.
Investors are increasingly looking for evidence that AI infrastructure companies can turn enormous demand into profitable, long-term contracts.
Nebius’ latest results provided several signals in that direction.
But the company’s aggressive spending also means expectations are becoming higher.
The more money Nebius commits to infrastructure, the more important execution becomes.
A delay in data-center construction, difficulty securing power, supply-chain problems or a slowdown in AI spending could create pressure on the company’s growth plans.
The Bigger AI Infrastructure Race
Nebius is not operating in isolation.
The AI cloud market has become increasingly competitive as specialized providers attempt to challenge the traditional cloud giants and capitalize on demand for high-performance computing.
The common problem is capacity.
AI developers want more GPUs. Data centers need more electricity. Infrastructure providers need more capital. And customers increasingly want guaranteed access to computing resources rather than uncertain availability.
That creates a powerful feedback loop.
More AI demand creates pressure for more computing capacity. More capacity requires massive infrastructure investment. And successful infrastructure providers can potentially secure long-term contracts before that capacity is even fully operational.
Nebius is betting that this cycle will continue.
Read More:- Steven Vandervelden Special Grand Jury: Why Did Pirro Make This Rare Move?
What Comes Next for Nebius?
The next stage could be more important than this quarter’s earnings beat.
Nebius now needs to turn its growing customer commitments into operational capacity while maintaining the economics that made those contracts attractive.
The company’s ability to secure power, deploy GPUs and expand data-center capacity could determine whether its current momentum becomes a long-term competitive advantage.
Investors will also be watching how quickly revenue grows relative to capital spending.
That is the central question surrounding Nebius now.
Can the company build enough AI infrastructure fast enough to satisfy customers without allowing the enormous cost of expansion to overwhelm its financial gains?
For now, the demand signal is difficult to ignore.
AI companies are still racing for computing power, and Nebius is positioning itself as one of the companies ready to supply it.
The latest earnings report suggests that the AI infrastructure race is no longer just about who has the most powerful technology.
It may increasingly be about who has the power, capacity and infrastructure to deliver it.