As demand for AI applications accelerates, computing infrastructure is becoming a critical factor in delivering faster responses, improved efficiency and more scalable enterprise AI services. Microsoftβs Maia 200 is an in-house AI inference accelerator developed for its Azure cloud infrastructure, designed to improve the economics of running large AI models in production. Built using TSMCβs 3-nanometre process, the chip features more than 140 billion transistors and delivers over 10 petaFLOPS at FP4 precision and more than 5 petaFLOPS at FP8 precision, according to Microsoft. The company says Maia 200 offers 30% better performance per dollar than the latest-generation hardware in its existing fleet. It is intended to support workloads across Azure, Microsoft Foundry and Microsoft 365 Copilot, while Microsoftβs Maia SDK preview gives developers tools to optimise models for the platform. The launch highlights the growing importance of custom AI silicon, data-centre efficiency and integrated hardware-software design in scaling next-generation AI services.
