Amazon Web Services launched his chip artificial intelligence Trainium3 on Tuesday, in a significant effort to compete with Nvidia in the lucrative market for computing power for artificial intelligence.
The move intensifies competition in the AI chip market, where Nvidia currently dominates with an estimated 80% to 90% market share for products used in training language models that power applications such as ChatGPT.
Last week, Google caused a stir in the industry when it was reported that Meta, Facebook’s parent company, would use its AI chips in its data centers, a challenge to Nvidia, currently the most valuable company in the world on the stock market and a key thermometer of the AI investment fever.
This followed the launch of Google’s latest AI model last month, which was trained using the company’s own in-house chips, not Nvidia’s.
AWS, which will make the technology available to its cloud computing customers, said its new chip is lower cost than rivals and offers more than four times the computing performance of its predecessor, while using 40% less energy.
Energy consumption is one of the main concerns of the AI revolution. Big tech companies reduce or pause their net-zero emissions commitments as they race to keep up with technology.
LOOK: Burst of the bubble, employment and the future of the press: the challenges of artificial intelligence in 2026
AWS argues that its chip can reduce the cost of training and operating AI models by up to 50% compared to systems that use equivalent graphics processing units, or GPUs, primarily from Nvidia.
AWS also announced that it is already developing Trainium4, which is expected to offer at least three times the performance of Trainium3.
Amazon’s own chip development reflects a broader trend among cloud service providers.
Likewise, last week, Nvidia said in X that it is “delighted” with the competition and assured that “it is a generation ahead of the industry.”