Microsoft’s latest AI chip goes head-to-head with Amazon and Google

Microsoft has launched the Maia 200, its second-generation AI inference accelerator built on TSMC’s 3 nm process and succeeding the 2023 Maia 100. The chip, containing over 100 billion transistors, is being deployed today in Azure’s US Central data-center region and is designed to run AI models faster and more efficiently, delivering roughly three times the FP4 performance of Amazon’s Trainium and surpassing Google’s TPU in FP8 benchmarks.

Rédigé à partir de l’ensemble des articles ci-dessous, et non d’un seul.