Microsoft has announced the launch of its second-generation Maia 200 artificial intelligence chip, delivering up to three times higher performance to support advanced AI workloads, including future OpenAI models such as GPT-5.

Built on a 3-nanometre architecture, the chip features 140 billion transistors, 216GB of memory, and a custom liquid-cooling system to ensure stable performance.

Microsoft said the move is part of its strategy to strengthen cloud infrastructure efficiency, cut operating costs by up to 30 percent, and reduce reliance on external AI chips in running generative AI applications and Copilot services.