Nvidia Launches Usage-Based AI Revenue Model with AWS and Azure
NVDA•Nvidia is rolling out a usage-based AI revenue-sharing model with AWS and Microsoft Azure, enabling per-inference billing on its H100 and next-generation Blackwell GPUs. The program will align Nvidia’s data center income with cloud partner usage, targeting accelerated enterprise adoption and expanded GPU margin capture.
1. Launch Details
Nvidia’s new model allows AWS and Microsoft Azure customers to pay for AI inference workloads on a consumption basis, with Nvidia sharing a portion of GPU margin with its cloud partners. The program covers existing H100 accelerators and will extend to the upcoming Blackwell GPU family.
2. Data Center Revenue Impact
By tying Nvidia’s data center revenue directly to partner usage, the company aims to boost growth beyond hardware sales and capture recurring income streams. Analysts project the shift could raise data center revenue growth rates by several percentage points over the next two quarters.
3. Market Implications
The usage-based structure lowers upfront investment for enterprises, potentially accelerating large-scale AI deployments on Nvidia platforms. It also fortifies Nvidia’s competitive moat by deepening strategic ties with leading cloud providers and standardizing GPU consumption across services.
