Microsoft reveals Maia 300 AI chip, as it works to reduce dependence on Nvidia, AMD

Microsoft has unveiled the Azure Maia 300 AI chip to reduce its reliance on Nvidia and AMD, to be deployed at multigigawatt scale from next year alongside Maia 200 and Maia 100. Microsoft says Maia 300 offers 40% more tokens per megawatt than leading GPU-based systems, and a 96-chip rack delivers seven times the rack-scale token throughput of Maia 200. Maia 200 already powers workloads including Microsoft 365 Copilot and OpenAI models, with 1.7 times more tokens per watt than other GPU systems. Peers such as Amazon, Google, Meta, OpenAI and Anthropic are also building custom chips to avoid overdependence on Nvidia and AMD. The article notes that Nvidia regularly generates over 50% of its quarterly revenue from hyperscalers, and that for now hyperscalers still need to buy Nvidia's most in-demand chips.

Summary written by Upside AI from the original article. Read the source for full details.

Read the original on Yahoo Finance ↗

All news →