Neocloud Providers Report Triple-Digit Growth as AI Infrastructure Demand Surges

CoreWeave and Nebius delivered Q2 earnings that beat expectations, with revenue climbing over 100% year-over-year as specialized GPU cloud providers capture share from constrained hyperscalers.

Specialized AI cloud infrastructure providers reported exceptionally strong second-quarter results this week, signaling continued demand for GPU computing capacity that traditional cloud giants cannot meet quickly enough. CoreWeave posted Q2 revenue of $2.58 billion, up 112% from a year earlier, while beating analyst expectations on both revenue and adjusted loss per share. The company raised its full-year revenue guidance to between $12.4 billion and $13.2 billion and increased its capital expenditure forecast to $35 billion to $39 billion. Its revenue backlog stands at $104 billion, excluding more than $25 billion in new third-quarter commitments. Nebius reported Q2 revenue of $582 million, up 454% year-over-year, with its AI cloud business accounting for 98% of total revenue. The company swung to adjusted EBITDA of $236 million from a $21 million loss in the prior-year period. Both stocks surged following the announcements, with CoreWeave shares jumping 14% in extended trading and Nebius rising 28%. Neoclouds are GPU-focused infrastructure providers built specifically for AI workloads, differentiating themselves from hyperscalers like AWS and Azure through faster deployment, purpose-built architecture, and multi-year take-or-pay contracts. The sector currently accounts for roughly 17% of AI infrastructure investment and is expected to grow to over 30% within a decade, according to industry estimates. Major customers include Microsoft, Meta, OpenAI, and Anthropic, with some hyperscalers themselves renting neocloud capacity to fill their own infrastructure gaps. CoreWeave is targeting more than 8 gigawatts of active power by 2030, while Nebius has raised its contracted power target to 5 gigawatts by year-end 2026. However, the capital-intensive model has drawn scrutiny: CoreWeave carries $35 billion in debt, and concerns about vendor financing loops and rising long-term interest rates caused neocloud stocks to pull back earlier this week despite strong fundamentals.

How could this affect you

Traced by who's actually in the path of this — not everyone is.

Likely

If you invest in cloud infrastructure or AI hardware supply chains

Neocloud buildouts could shift market share from traditional hyperscalers and increase demand volatility for GPU manufacturers like Nvidia, whose chips anchor most neocloud deployments and whose financing increasingly backstops neocloud expansion.

Direct

If your company is evaluating AI compute options

Neoclouds offer GPU capacity on shorter timelines than hyperscalers can currently deliver, but their multi-year take-or-pay contracts lock you into fixed capacity commitments whether or not you use every hour, creating financial risk if your workload forecasts prove wrong.

Likely

If you hold positions in traditional cloud providers

Hyperscalers becoming neocloud customers suggests they cannot build fast enough to meet their own AI infrastructure needs, which could compress their margins on AI services or delay their ability to capture the full economic value of the AI compute boom.

Off-Lens

Who is covering this — and who is not.

1 source reviewed

🇺🇸United States

100%

🇺🇸fool.com

Chanakya's Move

Whose move this was, what they're betting on, what could counter it.

This is a land-grab play dressed as a technology story. The neoclouds are betting that securing power, space, and GPU supply before 2029 will matter more than software differentiation or unit economics once hyperscaler capacity finally comes online. CoreWeave and Nebius are spending tens of billions to lock in multi-year contracts with customers who have no alternative supplier—because the alternative supplier (AWS, Azure, GCP) is often the customer itself. The strategic exposure is circularity: Nvidia finances the buildout, neoclouds buy Nvidia chips, customers pre-pay for capacity, and that capital flows back through the same small network of players. If demand holds, these companies become essential infrastructure. If hyperscaler capacity arrives faster than expected or AI workloads plateau, the neoclouds are left holding expensive, low-utilization assets financed at rising rates. The counter-move is already visible: Meta is reportedly building its own cloud service to sell excess capacity, turning a major neocloud customer into a potential competitor. The race is whether neoclouds can lock enough contracted revenue before their customers become their rivals.