Microsoft announces Maia 200, a custom AI inference accelerator built on TSMC's 3nm process with native FP8/FP4 tensor cores, 216GB HBM3e memory, and 272MB on-chip SRAM. The chip delivers over 10 petaFLOPS in FP4 and 5 petaFLOPS in FP8, offering 30% better performance per dollar than current hardware. Maia 200 features a
Table of contents
Engineered for AI inferenceOptimized AI systemsA cloud-native development approachSign up for the Maia SDK previewSort: