Key Highlights
- AWS plans to acquire 2 million more Nvidia GPUs for rollout between 2027 and 2028
- This order supplements the 1 million Nvidia processors Amazon secured in March
- The procurement encompasses Nvidia Blackwell Ultra, Rubin, and Rubin Ultra processors
- Nvidia’s CFO announced Amazon will implement Nvidia’s complete physical AI infrastructure for robotics operations
- After-hours trading saw Nvidia shares climb 4%; Amazon stock increased approximately 0.7%
Amazon has committed to installing 2 million more Nvidia GPUs throughout its AWS infrastructure during the coming two years, as both corporations announced Wednesday. The revelation sent Nvidia shares surging 4% during extended trading hours, with Amazon stock advancing roughly 0.7%.
Deployment of these processors is scheduled for 2027 through 2028. This expansion builds upon Amazon’s March commitment of 1 million Nvidia GPUs, currently being integrated into AWS facilities throughout this year.
The new agreement encompasses Nvidia’s cutting-edge Blackwell Ultra, Rubin, and Rubin Ultra GPU models. These premium processors are engineered specifically for AI model training and inference, with individual units commanding list prices reaching tens of thousands of dollarsāthough enterprise-scale buyers generally secure volume discounts.
Neither Amazon nor Nvidia disclosed specific monetary details regarding this arrangement.
During Wednesday afternoon’s earnings call, Nvidia CFO Colette Kress validated the expanded collaboration. She indicated that Amazon’s AWS business “is deploying an additional two million GPUs starting this quarter through the second of FY29.”
Nvidia’s Physical AI Technology Comes to Amazon Warehouses
Kress further disclosed that Amazon intends to implement Nvidia’s comprehensive physical AI platform across its warehouse robotics operations. This development creates a secondary dimension to their collaboration extending beyond cloud computing infrastructure.
Nvidia CEO Jensen Huang discussed the arrangement during a Wednesday evening CNBC interview. “This is going to be a great partnership,” he commented.
Amazon has simultaneously invested in proprietary chip development via Annapurna Labs, producing its own processors including Trainium, an AI accelerator positioned as an alternative to Nvidia’s offerings. Industry reports suggest the upcoming Trainium generation will incorporate Nvidia networking components.
Hyperscalers Maintain Nvidia Dependency Despite Custom Chip Development
While Microsoft and Alphabet’s Google are similarly pursuing proprietary chip solutions, both continue as significant Nvidia clients. This trend mirrors Amazon‘s approach.
The three dominant cloud providers are wagering that tailored processors can optimize performance for their particular use cases. Nevertheless, Nvidia’s GPU technology continues dominating the AI model development landscape.
Combining March’s agreement with Wednesday’s announcement, Amazon’s aggregate Nvidia GPU procurement totals 3 million units.
Wednesday’s Nvidia earnings call provided the setting for this disclosure. The 4% after-hours stock appreciation demonstrates investor attention to these cloud provider partnership developments.
Jensen Huang’s post-earnings CNBC interview emphasized the significance of the Amazon alliance, with leadership from both organizations publicly endorsing the broadened agreement.
Get 3 Free Stock Ebooks
Discover top-performing stocks in AI, Crypto, and Technology with expert analysis.
- Top 10 AI Stocks - Leading AI companies
- Top 10 Crypto Stocks - Blockchain leaders
- Top 10 Tech Stocks - Tech giants
