Amazon.com Inc.’s cloud division has secured a $38 billion agreement with OpenAI, marking one of the largest infrastructure partnerships in the artificial intelligence industry to date. Under the seven-year deal, Amazon Web Services (AWS) will provide the ChatGPT creator with access to hundreds of thousands of Nvidia graphics processing units, reinforcing OpenAI’s insatiable appetite for computing power. The companies announced the partnership on Monday, positioning it as a cornerstone of OpenAI’s next phase of large-scale AI development.
The arrangement underscores OpenAI’s dramatic transformation from a research organization into a full-fledged AI powerhouse, reshaping the global technology landscape. The company has reportedly committed to spending an unprecedented $1.4 trillion on infrastructure to develop and sustain its growing lineup of AI models — a scale of investment that has raised eyebrows across the financial world and sparked concerns about the formation of an AI bubble.
For Amazon, the deal marks a major comeback in the race to provide cloud infrastructure for the AI era. Despite being the world’s largest provider of rented computing power, AWS had been largely absent from OpenAI’s list of cloud partners, which includes Microsoft, Oracle, and Google. The partnership signals renewed confidence in Amazon’s technological capabilities and data center network. “As OpenAI continues to push the boundaries of what’s possible, AWS’s best-in-class infrastructure will serve as a backbone for their AI ambitions,” AWS Chief Executive Officer Matt Garman said in a statement.
Until now, Microsoft has been OpenAI’s primary cloud partner and largest investor, with the AI company recently pledging around $250 billion to Microsoft’s Azure platform. Oracle and Alphabet’s Google Cloud have also secured multibillion-dollar deals to provide capacity for OpenAI’s expanding operations, while the startup’s $22.4 billion agreement with CoreWeave has cemented that firm’s role among emerging “neo clouds” catering to AI developers.
Under the new deal, OpenAI will immediately begin using AWS’s computing resources, with all targeted capacity expected to be available by the end of 2026. Amazon plans to deploy clusters powered by Nvidia’s GB200 and GB300 AI accelerators to support both real-time ChatGPT responses and the training of next-generation language models. “Scaling frontier AI requires massive, reliable compute,” said OpenAI CEO Sam Altman. “Our partnership with AWS strengthens the broad compute ecosystem that will power this next era and bring advanced AI to everyone.”
Amazon’s relationship with the broader AI ecosystem continues to deepen. The company is one of the largest investors in Anthropic, a rival AI startup founded by former OpenAI researchers. AWS recently brought online a massive data center complex built for Anthropic, running on its proprietary Trainium2 AI chips. Meanwhile, Google announced last month that it would supply up to one million of its AI processors to Anthropic — a deal valued in the tens of billions.
The AWS–OpenAI partnership marks a significant moment in the escalating cloud wars, as the world’s biggest technology companies compete to power the most demanding AI workloads ever conceived.
