REDMOND (dpa-AFX) - Microsoft (MSFT) has unveiled Maia 300, its third-generation artificial intelligence accelerator designed to improve the performance and efficiency of large-scale AI inference workloads across its Azure infrastructure.
The company said Maia 300 delivers up to seven times greater rack-scale token throughput and more than 2.5 times the performance per watt compared with Maia 200.
Each accelerator includes 288 GB of HBM4 memory with 18 TB per second of bandwidth, providing 2.5 times greater memory bandwidth and 33 percent more memory capacity than its predecessor.
Maia 300 racks contain 96 accelerators, double the density of Maia 200, while a two-tier Ethernet-based network can connect as many as 12,288 accelerators in a superpod.
Microsoft said the architecture is designed specifically for long-context, reasoning-intensive and agentic AI inference. The company plans to use the technology across its AI portfolio, including
Microsoft Copilot, GitHub Copilot and Microsoft Foundry, with additional details expected at Microsoft Ignite 2026 in November.
Copyright(c) 2026 RTTNews.com. All Rights Reserved
Copyright RTT News/dpa-AFX
© 2026 AFX News
