Building AI infrastructure has become one of the most capital-intensive undertakings in technology. Oracle Corp (NYSE:ORCL) is raising debt to fund its aggressive data center expansion, CoreWeave, Inc. (NASDAQ:CRWV) has turned to equity and debt markets to finance growth, and hyperscale cloud providers are deploying billions of dollars from their own balance sheets. Nebius Group N.V. (NASDAQ:NBIS), however, is pursuing a different approach: asking customers to help pay for the infrastructure before it is built.
The AI cloud provider used its second quarter earnings call to reveal that upfront customer payments are becoming an increasingly important source of funding for its expansion. If the model proves sustainable, it could offer a new way for AI infrastructure companies to finance soaring capital expenditures while reducing—though not eliminating—their reliance on traditional funding sources.
Customer Prepayments Are Funding AI Infrastructure
Nebius said customer prepayments are now embedded in a majority of its large commercial agreements.
Chief Revenue Officer Marc Boroditsky said “roughly 70% of the deals we closed in Q2 included an upfront prepayment,” while Founder and CEO Arkady Volozh said those agreements “represent a yield of $20 million to $25 million per megawatt with upfront payments that cover 50% to 60% of the associated capex.”
Rather than waiting years to recover its investment after a data center becomes operational, Nebius is collecting a meaningful portion of project costs before infrastructure is deployed. The company also expects customer prepayments to exceed $9 billion this year, providing a substantial source of capital alongside more traditional financing options.
A Different Answer to AI’s Financing Challenge
The financing question has become one of the defining issues facing AI infrastructure providers. Building GPU clusters and modern data centers requires billions of dollars in upfront investment, often well before those assets begin generating meaningful revenue.
Different companies are taking different approaches. Oracle has leaned on debt markets to support its expanding AI infrastructure footprint, while CoreWeave has relied on a combination of equity, debt financing and customer commitments to fund rapid growth. Hyperscale cloud providers such as Microsoft Corp (NASDAQ:MSFT), Amazon.com Inc (NASDAQ:AMZN) and Alphabet Inc (NASDAQ:GOOG) (NASDAQ:GOOGL) Google have largely financed AI investments from their own balance sheets.
Nebius is attempting to shift part of that burden to customers that are eager to secure scarce computing capacity, effectively using committed demand to help finance future expansion.
Why the Model Matters for Investors
Customer prepayments do not eliminate Nebius’ need for debt or equity financing, nor has the company suggested they will replace traditional capital raising. They do, however, have the potential to reduce external funding requirements by providing cash before projects are completed, improving capital efficiency and shortening the time needed to recover investments.
The model also offers another important signal: customers are willing to commit significant sums well before capacity becomes available, suggesting confidence that demand for AI computing power will remain strong.
Whether other AI infrastructure providers adopt similar financing structures remains to be seen. But as the cost of building AI data centers continues to climb, Nebius’ approach points to a broader shift that could reshape how the industry’s next wave of expansion is funded—not only by investors and lenders, but increasingly by customers themselves.
Image Via Shutterstock
Login to comment