Skip to content
Market News

AI Compute Demand Rises as Cloud Providers Face Debt Costs, Says Grayscale

Grayscale reports rising AI computing demand, 50% higher H100 rental rates and mounting financing costs for cloud providers.

Grayscale CFN
  • H100 rental rates climbed approximately 50% year over year as GPU shortages tightened computing capacity.
  • CoreWeave faces qualified demand 10 times its available capacity, while Nebius reported a quarterly GPU-hour price increase exceeding 20%.
  • Financing expenses account for 26% of revenue among major AI cloud providers, compared with just 1% for leading hyperscalers.

AI computing prices are climbing as demand for GPUs exceeds available capacity, but rising borrowing costs are pressuring providers. Grayscale reports that H100 rental rates increased roughly 50% over the past year, while Nebius and CoreWeave reported strong demand. However, financing expenses now account for 26% of revenue among the five largest AI cloud providers.

GPU Rental Prices Rise as Capacity Tightens

According to Grayscale, H100 rental rates have increased approximately 50% over the past year. Nebius (NBIS) also reported that average GPU-hour prices rose more than 20% quarter over quarter in Q3 2026.

Meanwhile, CoreWeave (CRWV) reported qualified demand roughly 10 times its available capacity through 2027. Data from CBRE also shows tightening supply across North America. Primary data center vacancy fell to 1.4% during the first half of 2026. Additionally, 80.4% of capacity under construction had already secured tenants.

Nebius and CoreWeave Face Higher Growth Expectations

The supply shortage coincides with substantially higher revenue growth forecasts for AI cloud providers. Consensus estimates project CoreWeave’s revenue will grow 207% over the next year.

Nebius has an even higher projected growth rate of 484%. By comparison, analysts expect revenue across the five largest hyperscalers to increase 24%.

However, expanding computing capacity requires substantial investment. Providers must finance infrastructure before converting additional GPU capacity into billable computing hours.

Financing Costs Put Pressure on AI Cloud Providers

Grayscale identified borrowing expenses as a major challenge facing the sector. Yields on broad high-yield debt benchmarks have increased more than 100 basis points over the past year.

Over the last four quarters, interest expenses represented 26% of revenue for the five largest AI cloud providers. In comparison, the five largest hyperscalers recorded just 1%.

The figures highlight the financing gap between the two groups. Consequently, Grayscale identifies the speed of deploying new computing capacity as central to whether revenue growth can outpace borrowing costs.