Amazon Web Services (AWS) is poised to significantly deepen its involvement in Nvidia’s market, a strategic move that could represent one of the most substantial challenges yet to Nvidia’s long-standing dominance in AI chips.
Peter DeSantis, Amazon’s AI chief, informed Bloomberg that AWS is currently in discussions regarding the sale of its proprietary AI chip, Trainium, to other companies for deployment in their data centers. DeSantis refrained from disclosing the identities of these potential chip buyers.
The company confirmed to TechCrunch that these discussions about external chip sales are in their nascent stages. This initiative stems from Amazon CEO Andy Jassy’s annual shareholder letter in early April, where he hinted at the possibility of selling the company’s highly sought-after homegrown AI chips due to overwhelming demand.
Jassy articulated this demand, stating, “If our chips business was a standalone business, and sold chips produced this year to AWS and other third parties (as other leading chips companies do), our annual run rate would be ~$50 billion. There’s so much demand for our chips that it’s quite possible we’ll sell racks of them to third parties in the future.”
Assessing the potential impact on Nvidia, a competitor generating an estimated $50 billion annually would not necessarily cripple Nvidia, which currently boasts a $326 billion revenue run rate, especially if it maintains its recent strong quarterly performance. However, this projected revenue for AWS’s chip business is comparable to Intel’s annual revenues, signaling a formidable new player.
AWS has historically been reluctant to sell its AI chips directly, primarily due to the "waterfall effect" of its revenue model. While customers are charged directly for the AI tokens processed by these chips on its cloud platform, AWS also generates substantial revenue from a multitude of other essential services required for AI applications, including storage, security, networking, and monitoring.
Crucially, Amazon has consistently highlighted that the capacity for its chips sells out faster than it can produce them. In the same April shareholder letter, Jassy noted that the current Trainium chip capacity had sold out almost instantly. He added that the capacity for its successor, Trainium4, which is over a year away from availability, had also been fully allocated. This acute demand existed even before AWS formally incorporated OpenAI into the models it supports.
Consequently, selling chips to external entities would likely necessitate placing existing AWS customers on waiting lists, unless the company can significantly ramp up surplus manufacturing through partners like TSMC. Achieving this with TSMC, however, would present a considerable challenge, as Nvidia has recently surpassed Apple to become the foundry’s largest customer.
Doron Aronson, an AWS spokesperson, further corroborated the potential for AWS to sell these chips, remarking, “While we’ve historically declined requests to sell chips directly, Andy noted it’s quite possible we’ll sell racks of them to third parties in the future.”
Thus, while Nvidia’s founder and CEO Jensen Huang recently announced a new $200 billion market opportunity for Nvidia in selling CPUs for AI, thereby venturing into Intel and AMD’s territory, Jassy clearly harbors his own ambitious chip strategy: a $50 billion market that would directly embed AWS deeper into Nvidia’s competitive landscape.
The Editorial Staff at AIChief is a team of professional content writers with extensive experience in AI and marketing. Founded in 2025, AIChief has quickly grown into the largest free AI resource hub in the industry.