Why Nebius Is Raising Billions to Expand Its AI Cloud

Nebius turned to the convertible-debt market to finance the expensive next phase of its AI-cloud expansion. On August 19, 2026, the Amsterdam-based company announced a proposed private offering of $4.5 billion in convertible senior notes, split between $2.75 billion due in 2030 and $1.75 billion due in 2034.

Later the same day, Nebius announced pricing for an upsized $5 billion offering: $3 billion of 0.50% notes due in 2030 and $2 billion of 4.50% notes due in 2034. Buyers also received options to purchase up to an additional $750 million of notes.

What Nebius said the money would fund

Nebius said it would use the proceeds for continued growth, including:

• constructing and fitting out data centres;
• expanding its data-centre footprint;
• developing its full-stack AI-cloud offering;
• purchasing critical components, including GPUs; and
• general corporate purposes.

The financing reflects the cost of supplying AI computing at scale. A provider must build and equip data-centre capacity and secure large GPU deployments before it can deliver the computing resources required for model training and production workloads. Nebius’s stated investment plan is therefore closely tied to increasing the amount of AI capacity it can offer customers.

What a “neocloud” provider does

Nebius is an AI-focused cloud company rather than a broad, general-purpose hyperscale cloud. Its business combines GPU-based infrastructure with software and platform services for developing, training and deploying AI systems. The company describes its broader goal as a full-stack platform spanning data, model training and production deployment.

That specialization allows a neocloud to focus its capital and operating model on AI workloads. In practice, Nebius supplies computing capacity and related services to technology companies, including dedicated infrastructure for major customers.

Why customer commitments matter

The debt offering came as Nebius was preparing to expand capacity for substantial customer commitments.

Meta agreed to purchase $12 billion of dedicated AI capacity from Nebius over five years, with the potential to buy up to another $15 billion of capacity. That gives the agreement a possible total value of $27 billion, although the additional amount depends on capacity becoming available.

Nebius also has a multiyear agreement to deliver dedicated AI-infrastructure capacity to Microsoft from a new data centre in New Jersey.

These agreements help explain why Nebius is raising capital before all of the associated capacity is deployed: the company needs to build the infrastructure that supports contracted and prospective demand. They do not, by themselves, guarantee that every potential option will be exercised or that construction will proceed without execution risks.

The Eigen AI acquisition adds a software layer

Nebius’s expansion is not limited to physical infrastructure. In May 2026, it agreed to acquire Eigen AI in a cash-and-stock transaction valued at approximately $643 million. Eigen specializes in inference and model optimization, technologies intended to improve how efficiently AI workloads run.

Nebius announced that the acquisition closed in June and said Eigen’s technology would strengthen Token Factory, its managed inference platform.

The combination points to a two-part strategy: add more GPU-rich capacity while improving the software layer used to run AI models in production. The debt raise provides funding for the first part, while the Eigen deal supports the second.

What the financing means for Nebius

The core trade-off is straightforward. Convertible notes give Nebius access to substantial capital for expansion, but they also create future debt obligations and may result in share dilution if the notes convert into equity. The final $5 billion transaction was larger than the $4.5 billion proposal, underscoring both the scale of Nebius’s capital needs and the demand available for AI-infrastructure financing.

Nebius is effectively trying to convert customer demand into deployed AI capacity: data centres, GPUs, networking and a more capable cloud platform. Whether that strategy produces attractive returns will depend on how quickly the company can build the infrastructure, integrate Eigen’s technology and turn large commercial commitments into operating revenue.