Amsterdam based AI infrastructure provider Nebius is turning to the debt markets for another major capital push as the global race to build computing capacity intensifies. The company plans to raise up to $4.5 billion through convertible notes, giving it fresh financial firepower to expand data centres, strengthen its AI cloud business and purchase the high performance GPUs required to serve rapidly growing demand for artificial intelligence.
Two Notes With Long Term Maturities
Nebius said it is offering $2.75 billion of convertible senior notes due in 2030 and another $1.75 billion of convertible senior notes that will mature in 2034. The financing gives the company access to substantial capital while allowing it to continue investing in infrastructure at a time when AI computing requirements are increasing rapidly.
The proceeds are expected to support several areas of the business, including data centre construction, expansion of its AI cloud infrastructure and purchases of GPUs. These components are essential for companies developing and running increasingly sophisticated AI models, which require enormous amounts of computing power.
Building The AI Compute Backbone
Nebius operates in the rapidly expanding market often described as neocloud infrastructure. Instead of simply providing conventional cloud computing services, the company focuses heavily on infrastructure designed specifically for AI workloads.
Its model involves building and operating data centres equipped with large quantities of GPUs. Customers can then access this computing capacity through Nebius’s AI cloud platform. The company also provides specialised software and tools designed to help businesses deploy and operate AI applications more efficiently.
This approach places Nebius at the centre of the growing demand for specialised AI infrastructure. As AI companies develop larger models and enterprises increasingly integrate AI into their products and operations, demand for high performance computing infrastructure has risen sharply.
Big Tech Contracts Boost Demand
Nebius has already secured significant business from some of the world’s largest technology companies. The company has signed multi billion dollar agreements with Meta and Microsoft to provide AI infrastructure.
These contracts underline the scale of computing demand emerging from major AI players and provide Nebius with significant opportunities to expand its infrastructure footprint.
The company is competing in a market where access to GPUs, data centre capacity and reliable power has become strategically important. Building this infrastructure requires substantial upfront investment, making access to capital a critical part of the business model.
Expanding Beyond Infrastructure
Nebius has also been expanding its technology capabilities through acquisitions. In May, the company acquired US based startup Eigen for approximately $643 million in cash and stock.
Eigen specialises in improving the performance of leading open source AI models. The acquisition gives Nebius additional technology capabilities that complement its infrastructure business, potentially allowing customers to benefit from both computing resources and software designed to make AI models more efficient.
The move reflects a broader strategy in which AI infrastructure providers are increasingly looking beyond simply supplying computing capacity and are building tools that help customers make better use of that capacity.
A Capital Intensive AI Race
The planned $4.5 billion convertible note offering highlights just how capital intensive the AI infrastructure industry has become. Data centres require major spending on buildings, power systems, networking equipment and GPUs before they can generate returns.
For Nebius, the new financing could provide the resources needed to accelerate its expansion as AI demand continues to grow. With major technology companies already among its customers and additional investments being made in AI cloud infrastructure, the company is positioning itself to capture a larger share of the global market for specialised AI computing.
