Musk and Palihapitiya Warn of Potential Slowdown in AI Infrastructure Spending

Elon Musk and venture capitalist Chamath Palihapitiya said the rapid expansion of artificial-intelligence (AI) data-center infrastructure could encounter a slowdown as early as 2027. Both highlighted concerns about tighter credit markets, delays in data-center construction and the risk that expensive computing equipment might remain idle.

Palihapitiya, posting on X, argued that specialized cloud providers—often called "neoclouds" and exemplified by firms such as CoreWeave and Nebius—are poorly capitalized and may struggle to fund their expansion. He warned that these challenges could leave "a lot of partially stranded compute" in inventory next year, potentially prompting larger hyperscalers to prioritize internal AI workloads over third-party customers.

Musk echoed the assessment, noting that lenders are reluctant to finance companies without "substantial revenue independent of AI data centers." He cited SpaceX’s recent $40 billion debt financing for GPUs, explaining that the company’s investment-grade credit profile and diversified revenue streams from Starlink and launch services give it a financing advantage that many neoclouds lack.

Analysts note that Nvidia’s business model relies on continuous GPU purchases by both hyperscalers and neoclouds. If financing constraints or construction delays force operators to postpone or cancel projects, Nvidia could see a slowdown in orders even if overall AI demand stays strong. While major hyperscalers such as Microsoft, Amazon and Alphabet have stronger balance sheets, their spending is not unlimited, and infrastructure bottlenecks could make additional GPU purchases less attractive.

The commentary suggests that investors should consider the possibility that AI demand and infrastructure spending may diverge, potentially affecting Nvidia’s valuation if GPU orders weaken.