Microsoft today announces the Maia 200 AI accelerator, its successor to Maia 100, built on TSMC’s 3nm process with more than 100 billion transistors. Microsoft says Maia 200 delivers three times the FP4 performance of Amazon Trainium Gen3 and stronger FP8 than Google’s TPU v7, will host OpenAI’s GPT-5.2, improves inference cost-efficiency by about 30 percent, and begins deployment in Azure US Central.
Key Points
- 1Introduces Maia 200 chip with over 100 billion transistors on TSMC 3nm, deployed in Azure
- 2Claims 3× FP4 performance versus Amazon Trainium Gen3 and superior FP8 over Google TPU v7
- 3Enables hosting of OpenAI’s GPT-5.2 and offers about 30 percent better inference performance per dollar
Scoring Rationale
Official announcement of a high-performance 3nm in-house accelerator with cloud deployment; independent benchmarks remain limited.
Sources
Public references used for this report.
Practice with real Logistics & Shipping data
90 SQL & Python problems · 15 industry datasets
Used by DS/ML engineers at top companies
High-Value Overnight OrdersEasyDelivered International ShipmentsMediumOn-Time Delivery Rate by CarrierHard
250 free problems · No credit card
See all Logistics & Shipping problems
