If you've been following the stock market this year or the artificial intelligence (AI) boom, you had to have heard of Nebius (NASDAQ: NBIS) by now. The stock is on an absolute tear, up more than 200% this year. The stock jumped almost 50% last week alone.

Nebius' revenue is on track to roughly quadruple this year, but the stock still sits in a gray zone where some investors see a future AI infrastructure giant and others see a capital-hungry niche player priced for perfection. The debate comes down to what Nebius is actually building, how durable that revenue is, and whether the current valuation already assumes most of the upside.

Missed Nvidia in 2009? This Rare Signal Is Flashing Again. In 2009, a "Double Down" signal flashed for a little-known chipmaker called Nvidia. For the first time in years, that same "Total Conviction" signal is flashing for a company 1/100th the size of Nvidia. Continue »

Explosive revenue growth

Nebius' growth is coming in huge steps. In Q1 2026, revenue hit $399 million, up 684% versus about $51 million a year earlier, and up 75% from Q4. Its AI cloud unit produced $389.7 million in Q1, an 841% year-over-year increase, and by Q2, total revenue reached $582.3 million as AI cloud sales rose nearly sixfold, beating analyst expectations. Management projects $3 billion to $3.4 billion of revenue this year, which would be roughly quadruple 2025 revenue and put Nebius into midtier hyperscaler territory based on the top line alone.

What Nebius actually sells

Nebius offers a specialized AI native cloud, not a general-purpose cloud like Amazon Web Services or Microsoft's Azure. It offers non-virtualized GPU clusters, fast networking, and a full stack of tools for training, fine-tuning, and inference, built for large models rather than traditional enterprise workloads. Customers can spin up GPU capacity in minutes, scale clusters elastically, and pay under flexible consumption models tuned to AI training and inference rather than generic computing hours. That focus lets Nebius pitch better performance and a lower total cost of ownership to AI natives and enterprises that prioritize raw training throughput over deep integration with a broader cloud ecosystem.

Capacity, capital and contracts

The revenue ramp-up is not happening in a vacuum. Nebius has signed multibillion-dollar AI cloud contracts, including at least four core infrastructure deals averaging more than $1 billion each, and has raised its contracted power target to 5 gigawatts (GW) to support those workloads. It's spending heavily to keep up. The company anticipates $20 billion to $25 billion of capital expenditure (capex) in 2026 alone, mostly on GPUs, data center build-outs, and related hardware, on top of earlier plans in the $16 billion to $20 billion range. To relieve some of that pressure, Nebius has rolled out an infrastructure partnership model where third-party operators finance and own AI data centers, while the company provides its architecture, hardware design, and software stack, and then sells the resulting capacity through its global marketing organization.