At its annual re:Invent conference, Amazon.com Inc. (AMZN) unveiled its latest self-designed artificial intelligence chip, Trainium3, marking a significant step in its efforts to scale its AI hardware business and directly challenge the market leader, Nvidia Corp. (NVDA).
Amazon Web Services (AWS) CEO Matt Garman announced the new chip, stating that the Trainium business line already represents “a multibillion-dollar business today and continues to grow really rapidly.”
A primary driver for Amazon’s in-house chip development is cost reduction. While Nvidia’s powerful GPUs are the industry standard for training and running advanced AI applications, they come at a premium price. Reports indicate Nvidia’s latest Blackwell chips cost between $30,000 and $40,000 each, with AI workloads often requiring clusters of thousands in data centers.
AWS Vice President Dave Brown told Yahoo Finance that developers can save “30% to 40% by using Amazon chips instead of Nvidia’s.” This offers a more cost-effective alternative for AWS customers and reduces Amazon’s own reliance on Nvidia’s supply. According to Bloomberg, Amazon accounts for approximately 7.5% of Nvidia’s revenue.
Analysts suggest Amazon’s move is unlikely to dethrone Nvidia in the near term. Amazon is not expected to cease Nvidia purchases entirely, and Nvidia boasts a deep and diverse customer base to offset potential shifts in demand.
Nvidia CEO Jensen Huang has previously stated the company’s cloud GPUs are sold out and Blackwell sales are “off the charts.” For its fiscal third quarter 2026 (ended October 26, 2025), Nvidia reported revenue of $57 billion, a 62% year-over-year increase, with data center revenue soaring 66% to $51.2 billion. The company forecasts full-year revenue of $212.8 billion, projecting growth to $316 billion in fiscal 2027 with its next-generation Rubin architecture.
The chipmaker has recently secured major deals with OpenAI, Anthropic, Intel, Palantir Technologies, Alphabet, Microsoft, Oracle, and xAI, underscoring robust demand.
However, competitive pressures are rising. The market’s reaction to Amazon’s chip and reported negotiations between Meta Platforms and Alphabet’s Google for data center chips signal that Nvidia’s largest customers are actively seeking alternatives. Nvidia’s leadership now faces the task of continually proving the superior value of its GPUs to justify their premium.
Amazon’s foray into AI semiconductors is underpinned by two core strengths:
The AI boom presents a major growth vector for AWS. Garman emphasized that AWS was a strong business before AI’s emergence, and this new technology is now providing an additional boost. By developing its own AI chips, Amazon not only strengthens its infrastructure competitiveness but also positions AWS to capture more value from the ongoing AI wave, potentially unlocking a new cycle of growth for its cloud division.
