Reports indicate that Nvidia plans to supply Amazon.com's cloud computing division with 1 million of its graphics processing unit chips and various other products by 2027.

This week, Nvidia and Amazon Web Services announced that AWS secured an agreement to purchase 1 million GPUs, though the specific timing of the transaction was not revealed. Ian Buck, vice president of high-performance computing and hyperscale at Nvidia, informed Reuters on Thursday that the sales would commence this year and continue until 2027.

During that same period, Nvidia CEO Jensen Huang stated that the company identifies a total sales potential of $1 trillion for its Rubin and Blackwell chip families.

Nvidia and Amazon did not reveal the financial details of their agreement. However, Buck informed Reuters that the deal includes a wide variety of Nvidia components in addition to the 1 million GPUs, such as Nvidia's Spectrum networking chips and the Groq chips that Nvidia launched this week following its $17 billion licensing agreement with an AI chip startup from late last year.

Also Read: AI & Tech: Visionary Pre-Budget Insights from Industry Leaders

Specifically, AWS intends to utilize a mix of Nvidia's Groq chips and six additional chips from Nvidia to enhance the efficiency of inference, which is the term for how AI systems produce responses and perform tasks for users.

"Drawing conclusions is difficult." It’s extremely challenging. To excel at inference, it isn't just a one-trick pony. "We indeed utilize all seven chips," Buck states.

Also Read: NVIDIA's Rubin Moment at CES: AI Supercomputing Breakthrough

The agreement also involves installing Nvidia's Connect X and Spectrum X networking equipment in AWS data centers. That decision is important because AWS data centers rely on specialized networking hardware that AWS has refined over several years.

Also Read: AI Autonomy Race: How Advanced are Top Countries' AI Strategies?

"They're definitely going to carry that out, as expected," Buck stated. "Currently, we are working together to implement Connect X and Spectrum X for significant workloads and our largest clients in AI with AWS."