AWS Trainium Chip Business Hits Multi-Billion Revenue, Challenging NVIDIA’s Pricing

According to Andy Jassy, AWS’s in-house AI compute chip business built around Trainium has already reached multi-billion-dollar revenues, signalling strong market endorsement of Amazon’s strategy to drive down the cost of AI computation, with momentum behind the Trainium2 line, production having surpassed one million units and more than 100,000 enterprises relying on Trainium2 primarily through the Amazon Bedrock platform.

Jassy noted that Trainium’s rise stems from its superior price-performance ratio compared with competing GPU offerings, offering a cheaper solution while delivering comparable—and for certain workloads, even superior—performance. AWS CEO Matt Garman confirmed that the company’s strategic partner Anthropic has played a pivotal role, with more than 500,000 Trainium2 chips deployed under Project Rainier to train and build Claude models.

At AWS re:Invent, Amazon unveiled Trainium3, boasting four times the performance of its predecessor and improved energy efficiency, and the upcoming Trainium4 is being designed to interoperate with NVIDIA GPUs via NVLink Fusion, enabling a hybrid architecture. This signals a shift away from a direct replacement strategy toward a hybrid compute model within AWS’s cloud, and judging by its revenue trajectory, Trainium is becoming a firmly established pillar of the cloud strategy.