In a massive expansion of their partnership, Amazon is set to add 2 million Nvidia GPUs to its AWS data centers over the next two years. This deal covers cutting-edge Blackwell and Rubin architectures, signaling a massive bet on the future of generative AI.
- Amazon is adding 2 million additional Nvidia GPUs to its AWS infrastructure by 2028.
- The deal includes next-gen hardware: Nvidia Blackwell Ultra and Rubin series.
- The partnership extends to robotics, networking, and enterprise AI software.
- Nvidia's data center revenue surged 117% year-over-year to $89 billion.
The alliance between Amazon and Nvidia has reached a new, monumental scale. On Wednesday, the two tech giants announced an expanded partnership that will see Amazon adding another 2 million Nvidia GPU chips to its Amazon Web Services (AWS) data centers over the next two years.
This massive order is a direct response to the 'surging demand' from startups, enterprises, and even governments for high-performance computing. The chips slated for deployment in 2027 and 2028 include Nvidia's most advanced architectures, specifically the Blackwell Ultra, Rubin, and Rubin Ultra GPUs, which are engineered to handle the most intensive AI training workloads.
Why This Matters
BozokMedia analysis shows that this move highlights a complex paradox in the tech industry. While Amazon is aggressively developing its own custom silicon—such as the Trainium AI chips and Graviton CPUs—to reduce dependency on third-party vendors, it is simultaneously doubling down on Nvidia to ensure it remains at the cutting edge of the AI arms race. This 'co-opetition' is a defining characteristic of the current semiconductor landscape.
The sheer scale of this deal underscores that compute power has become the most valuable commodity in the modern digital economy.
Beyond mere hardware, the partnership is deeply integrated. Nvidia's networking hardware, CPUs, and robotics platforms will be woven into the AWS fabric. Notably, Amazon plans to adopt Nvidia's full 'physical AI' stack to power its massive fleet of warehouse robots, utilizing platforms like Omniverse, Isaac, and Jetson.
Historical Background: The Silicon Arms Race
The race for AI dominance has fundamentally shifted the power dynamics of Silicon Valley. A few years ago, general-purpose CPUs were the standard; today, specialized GPUs are the engine of progress. Amazon's custom chip business has already crossed a $25 billion annualized revenue run rate, yet Nvidia remains the undisputed 'GOAT' (Greatest of All Time) in the AI hardware space, with data center revenues skyrocketing by 117% to $89 billion in the last quarter alone.
| Feature | Nvidia (Partner) | Amazon (Internal) |
|---|---|---|
| Primary Hardware | Blackwell, Rubin GPUs | Trainium, Graviton CPUs |
| Market Role | Infrastructure Provider | Cloud Service & Custom Silicon |
| Strategic Focus | AI Dominance | Cost Optimization & Scale |
Nvidia CEO Jensen Huang emphasized during the earnings call that AI is moving into a productive phase where it generates 'profitable tokens,' justifying the hundreds of billions of dollars being poured into infrastructure by companies like Amazon.
Frequently Asked Questions
1. Which Nvidia chips will Amazon be using?
Amazon will integrate advanced models including the Blackwell Ultra and the upcoming Rubin and Rubin Ultra series.
2. Is Amazon trying to compete with Nvidia?
Yes, Amazon is developing its own Trainium and Graviton chips to provide alternatives to Nvidia's high-cost hardware for specific workloads.
