The scramble for artificial-intelligence computing power is pushing the industry's biggest players to make their own silicon rather than rely solely on Nvidia.
Amazon CEO Andy Jassy said the company's chip business already has $225 billion in commitments, according to Yahoo Finance. That figure signals how much demand Amazon believes it can capture by designing processors in-house instead of buying every chip from an outside supplier.
The move puts Amazon in direct competition with Google, which has also invested heavily in custom AI silicon. Yahoo Finance, 24/7 Wall St., and AOL.com all framed the story the same way: a race to beat Nvidia, with the open question being whether Google or Amazon has the better in-house design.
Nvidia remains the company to beat. Its graphics processors have become the default hardware for training and running large AI models, giving it enormous pricing power and long waiting lists. By building their own chips, Amazon and Google aim to cut costs, reduce dependence on a single vendor, and tailor hardware to the specific AI workloads running in their cloud data centers.
The sources do not settle which company's silicon is superior, presenting it instead as an ongoing contest. What is clear is that two of the largest cloud providers now see custom chips as central to their AI strategies rather than a side project.
Why it matters: if Amazon and Google succeed in fielding competitive in-house chips, they could chip away at Nvidia's dominance and reshape the economics of AI — potentially affecting what businesses and consumers ultimately pay for AI services.