Amazon CEO Andy Jassy revealed in April that the company's current Trainium chip capacity had sold out almost instantly, even before external sales began. This rapid absorption, including Trainium4 capacity (a chip slated for availability in the future), confirms strong market confidence, according to TechCrunch.
Amazon has historically been a massive buyer of AI chips for its cloud infrastructure, but it is now entering the market as a direct seller of its own advanced AI chips to external data centers. This strategic shift makes Amazon a direct competitor to its former key suppliers, challenging established industry dynamics.
The AI chip market faces a significant shake-up. Amazon's entry will likely accelerate innovation, diversify the supply chain, and potentially lead to more competitive pricing for high-performance AI hardware. This development reshapes the Amazon AI chip market entry vs Nvidia 2026 narrative.
Trainium's Proven Demand and Core Capabilities
Trainium's immediate sell-out, even for future models like Trainium4, confirms its market relevance. Trainium3 features eight large cores, each with four specialized engines—Tensor, Vector, Scalar, and GPSIMD—running simultaneously, according to AWS. This architecture delivers up to 2x MXFP8 compute throughput compared to Trainium2 and includes 144 GB HBM3e per chip, providing 4.9 TB/s bandwidth, according to AWS. The specifications underscore Trainium's high performance for demanding AI workloads.
Advanced Performance and Competitive Landscape
Trainium3 significantly boosts performance with 1.7x higher bandwidth than Trainium2. Trn1 instances also support up to 800 Gbps of second-generation Elastic Fabric Adapter (EFAv2) network bandwidth, according to AWS. These enhancements enable more efficient data processing and high-speed chip communication for large AI models. Meanwhile, Nvidia's founder and CEO Jensen Huang declared a new $200 billion market for Nvidia in selling CPUs for AI, beyond GPUs, according to TechCrunch. The pivot signals early recognition of increased competition in the core AI training GPU market. Trainium3's capabilities make it a strong contender, while Nvidia's expansion into AI CPUs highlights intensifying competition across the entire AI hardware market.
