This commitment, announced Wednesday, follows a previous agreement from March to install 1 million Nvidia chips within Amazon Web Services infrastructure starting this year.
The deal secures a massive supply of specialized hardware designed specifically for training and running complex artificial intelligence models.
The expansion signals that Amazon remains reliant on Nvidia’s market-leading hardware to power its cloud computing services despite developing its own internal competitive chips.
By securing millions of these processors, Amazon ensures its data centers can handle the increasing intensity of AI workloads.
This massive capital expenditure highlights the ongoing race among technology giants to build the physical infrastructure necessary to support next-generation software.
The rollout will primarily affect Amazon Web Services, the company’s cloud division, as it scales up to meet customer demand for AI processing power.
These chips, known as GPUs (chips that handle intensive mathematical tasks simultaneously), are the industry standard for modern machine learning.
While Amazon continues to refine its own chipmaking efforts, this deal confirms that Nvidia's technology will remain a cornerstone of Amazon’s technical architecture for the remainder of the decade.