Nvidia Confirms 1 Million GPU Delivery to AWS Through 2027
Nvidia Corporation (NASDAQ:NVDA) announced that it will supply 1 million graphics processing units (GPUs) to Amazon Web Services between 2026 and 2027. This marks the first official timeline for a major partnership between the two tech giants, covering a wide range of chips and networking equipment.
Multi-Chip Strategy for AI Workloads
Ian Buck, Nvidia’s vice president of hyperscale and high-performance computing, confirmed to Reuters that GPU deliveries will start this year and continue through 2027. The deal goes beyond GPUs—AWS will also acquire Nvidia’s Spectrum networking chips and newly released Groq chips, which Nvidia licensed in a $17 billion deal with AI startup Groq in late 2025.
AWS plans to deploy Groq chips alongside six other Nvidia chip types to optimize AI inference workloads—the process of generating responses and executing tasks in AI systems. Buck emphasized, “Inference is hard. It’s wickedly hard… To be the best at inference, it is not a one-chip pony. We actually use all seven chips.”
Networking Equipment Integration
The partnership also includes deploying Nvidia’s Connect X and Spectrum X networking equipment in AWS data centers, a notable shift as AWS has traditionally relied on its proprietary networking hardware. Buck noted that AWS will continue using its own custom gear but will collaborate on deploying Nvidia equipment for critical AI workloads and major clients.
Financial terms of the agreement were not disclosed.
Trillion-Dollar Opportunity for Nvidia
The AWS deal aligns with CEO Jensen Huang’s projection that Nvidia could access a $1 trillion sales opportunity for its Rubin and Blackwell chip families by 2027. This estimate excludes CPUs, networking chips, Groq products, and the Rubin Ultra variant, suggesting that the total addressable market could be even larger.
Huang also highlighted that Groq integration could unlock $300 billion in annual revenue per gigawatt, with roughly 25% of GPU workloads expected to leverage Groq chips. The Nvidia-Groq system, called the LPX, is designed as an optional integration with the Vera Rubin platform but has not yet been scaled for broad deployment.
Key Developments to Watch
Investors should monitor several aspects of this partnership:
- GPU delivery pace: Whether Nvidia meets the 1 million-unit multi-year commitment and how shipments are distributed by chip generation
- Groq deployment timeline: Scaling of inference-optimized Groq chips beyond pilot programs
- Networking adoption: AWS’s use of Connect X and Spectrum X versus its proprietary equipment
- Fiscal 2027 results: Financial impact of the AWS partnership expected to be reported in early 2027
- Competitive response: How rivals like AMD react to Nvidia’s growing footprint in AWS infrastructure
This collaboration represents a significant expansion of Nvidia’s presence in cloud AI infrastructure and positions the company to capitalize on the growing AI computing market.
