Layer 2.5 · head to head
CoreWeave vs Nebius
Two public companies at this layer, so for once you can read both sets of numbers before signing.
| CoreWeave | Nebius | |
|---|---|---|
| Can you leave | Portable. The workload is standard enough to move to another vendor without a rewrite. | Movable with effort. Expect to redo images, storage wiring and networking. |
| Counterparty | Publicly reporting, which is the point — you can read the filings. Carries substantial GPU-collateralised debt with maturities in the 2026-2028 window, and revenue is heavily concentrated in a few customers. Concentration cuts both ways: it funds the buildout and it is the risk. | One of the few at this layer you can actually diligence — quarterly results are public. Revenue grew roughly 4.5x year-over-year in Q2 2026 off a small base. Same 2026-2028 GPU-debt maturity picture as its peers. |
| What it is | Purpose-built GPU cloud. Owns or leases its own datacentre capacity and sells it directly. | Purpose-built GPU cloud. Owns or leases its own datacentre capacity and sells it directly. |
| How you reach it | You get a cluster. Assumes you already run Kubernetes at scale. | You get virtual machines. The familiar cloud model. |
| Accelerators | H100, H200, GB200 NVL72, B200, A100, L40S | H100, H200, B200, L40S |
| Regions | US, EU, UK | EU (Finland), US |
| Pricing model | On-demand and reserved. Reserved is the real product; on-demand list price is high enough that it reads as a discouragement. | On-demand and reserved, published openly. |
| Getting started | No formal minimum on-demand, but the commercial motion is contracts, not credit cards. | Self-serve on-demand. |
| Capacity | Large and contracted years ahead. Most capacity is spoken for by a small number of very large customers. | Expanding hard; Finland datacentre is the anchor. |
| Ownership | Public (IPO completed). NVIDIA is an investor. | Public, files with the SEC as a foreign private issuer. |
CoreWeave
For: Teams that already run Kubernetes at scale and are buying committed capacity rather than experimenting.
The catch: List on-demand pricing is among the highest of the neoclouds — widely reported around $6/hr for an H100 where specialists sit near $2. You are not meant to pay list; if you are paying list, you are using it wrong.
Economics: Priced 30-40% under the hyperscalers on comparable committed terms, which is the entire pitch. Against the cheaper specialists it is not a price play at all — it is a scale, network and support play.
Nebius
For: European workloads where data residency matters, and anyone who wants a supplier whose numbers they can read before signing.
The catch: Smaller footprint than CoreWeave, and the European centre of gravity is a feature or a latency problem depending entirely on where your users are.
Economics: Published pricing sits below CoreWeave's list on comparable parts. Publishing at all is a differentiator at this layer.
Neither table row is a price
Deliberately. Published on-demand rates at this layer move weekly, and essentially nobody signing a real contract pays them — every serious buyer pays less than every list figure either of these companies publishes. Quoting one here would date this page within a month.
The GPU rental price index carries dated, sourced figures instead, and the durable finding there is the spread: the identical H100 rents from roughly $1.38 to $12.29 an hour depending only on who you rent it from.
The layers underneath both
Whichever you pick is renting you chips in a building that needs power. In 2026 that is the constraint that binds: Microsoft has disclosed an Azure backlog it cannot fill for want of megawatts rather than accelerators, and the US interconnection queue exceeds 2,600 GW with roughly 80% of projects withdrawing before they energise.
Layer 2 — Silicon · Layer 1 — Energy · The interconnection queue