AI Infrastructure2026-11-13
TechCrunch AI
Nebius plans $40B AI infrastructure buildout
Nebius is testing the neocloud playbook with a planned $40 billion expansion of AI infrastructure. The company intends to fund the buildout through customer prepayments and loans, a financing model that could let it grow quickly but also exposes it to risk if demand or construction timelines shift. Questions about that approach have grown as delays at a Microsoft data center raise doubts about how reliably such projects can be delivered. Nebius is one of several GPU cloud providers trying to meet surging demand for AI compute while competing with much larger hyperscalers. Its plan shows how AI infrastructure spending has become as much a capital-markets story as a technology story. Building large data centers requires enormous upfront investment, long-term customer commitments and access to power, chips and construction capacity. Customer prepayments can help finance expansion, but they also tie the company's future to the willingness of AI developers to pay ahead of time. Loans add leverage and increase the cost of failure. The delays at Microsoft's data center are notable because they suggest that even the largest and best-resourced operators can run into scheduling and construction problems. For Nebius, the challenge will be proving that its financing model can support rapid growth without creating a fragile balance sheet. If the plan works, it could establish the company as a serious alternative for AI compute capacity. If it stumbles, it could become an example of how the race to build AI infrastructure can outpace the ability to finance and operate it. The scale of the $40 billion plan makes Nebius a company to watch in the neocloud sector.