Nvidia Corp. said Thursday that rental prices for its H100 GPUs continue to rise amid strong demand for AI infrastructure [1].
This trend directly impacts the profitability of AI cloud-infrastructure providers. Companies that lease these chips to other businesses can realize higher margins as the cost of accessing the hardware increases.
Nebius Group, a provider of AI cloud infrastructure, saw its share price increase by 15% [2] following the announcement. The surge reflects investor confidence in the company's ability to capitalize on the scarcity and high demand for Nvidia's high-end GPUs.
Other industry players also saw gains on Wall Street. Shares for CoreWeave and Iren each rose by four% or more [2] after the disclosure regarding rental rates.
Nvidia's H100 GPUs remain the industry standard for training large-scale artificial intelligence models. The continued price climb suggests that the appetite for compute power is outstripping the available supply of hardware in the cloud market [3].
The financial impact is felt most acutely by the specialized cloud providers who act as intermediaries between Nvidia and the end-users. By securing large fleets of H100s, these firms can leverage the rising market rates to increase their own revenue streams [4].
Market analysts said that the revelation from Nvidia's leadership highlights a persistent gap in the AI supply chain. As long as the demand for generative AI remains high, the rental market for these specific chips is expected to remain competitive [5].
“Rental prices for its H100 GPUs continue to rise amid strong demand for AI infrastructure”
The rising cost of H100 rentals indicates that AI compute remains a scarce commodity despite the rollout of newer hardware. For the broader market, this suggests that the 'picks and shovels' phase of the AI boom—where infrastructure providers profit from the physical requirements of AI—is still yielding high returns, shifting the financial advantage toward those who already own and operate large-scale GPU clusters.
