Nebius plans to pay for $40 billion AI buildout with customer prepayments and debt
Summary
Nebius intends to finance a planned $40 billion expansion of its AI infrastructure by drawing on two sources of money: advance payments from customers and borrowed funds, TechCrunch AI reported on November 12, 2026.
The plan comes as a delayed Microsoft data center project has prompted scrutiny of how the company funds its growth. The report frames the spending as a test of the model used by neoclouds, the newer specialized AI cloud providers.
The source does not describe the delay, give a timeline for the spending, name lenders or customers, or break down how the $40 billion would be split between prepayments and loans.
Why it matters
Funding a buildout of this size through customer prepayments and loans ties capacity delivery directly to contract performance and credit markets, not just to demand. If one large data center project slips, the cash and lending terms behind a neocloud's expansion can come under pressure. Investors and lenders weighing neocloud exposure, and hyperscale customers relying on these providers for overflow GPU capacity, have reason to examine how sensitive that financing structure is to schedule risk.
Expert Lens
A prepayment-backed model works only as long as the capacity customers paid for arrives on time, so construction or energization delays may trigger renegotiation, refunds or tighter terms from buyers. Lenders are likely to price that ready-for-service risk into covenants, which could raise the cost of the debt half of the plan. This suggests neoclouds with concentrated customer books face sharper refinancing exposure than diversified operators. The next signal to watch is whether Nebius discloses new prepayment agreements or debt facilities and what delivery milestones they are tied to.
Sources
This brief was drafted with AI assistance from public sources; every number is attributed to the sources above. See the AI-assisted Disclosure and Methodology.