Amazon Web Services (AWS), the cloud computing arm of Amazon, is reportedly in early-stage discussions to sell its custom-designed AI chips, known as Trainium, to third-party data centres. This strategic shift could see Amazon directly challenging Nvidia, the current leader in the artificial intelligence chip market, and marks a significant evolution in AWS's business model.
The potential move stems from comments made by Amazon CEO Andy Jassy in an April shareholder letter, where he highlighted the high demand for Amazon's proprietary AI chips. Jassy suggested that if the chip operation were a standalone entity, supplying both AWS and external customers, it could achieve an annual run rate of around $50 billion. Peter DeSantis, Amazon's AI chief, confirmed to Bloomberg that talks are underway regarding the sale of Trainium chips, though specific potential buyers were not disclosed.
Historically, AWS has primarily used its custom chips internally to power its cloud services, offering customers access to AI processing capabilities without selling the physical hardware. The revenue generated from these chips has largely been indirect, coming from the extensive suite of associated cloud services like storage, security, and networking that companies require for their AI applications. Shifting to direct chip sales would represent a substantial departure from this approach.
The decision to explore external sales comes despite Amazon's current struggles to meet internal demand for its AI chips. CEO Jassy previously stated that the capacity for both current Trainium chips and the upcoming Trainium4, due in over a year, has sold out almost instantly. This suggests that for Amazon to significantly penetrate the external market, it would likely need to dramatically scale up its manufacturing capabilities, potentially competing with Nvidia for foundry capacity from partners like TSMC.
For the UK, this development could have several implications. Increased competition in the AI chip market could lead to a more diverse supply chain and potentially lower costs for businesses developing AI solutions, from startups to large enterprises. This could foster greater innovation and accelerate the adoption of AI across various sectors, from finance to healthcare. However, it also highlights the UK's reliance on global supply chains for critical technology components, underscoring the importance of resilient infrastructure and strategic partnerships.