Microsoft Unveils In-House AI Chip to Reduce Nvidia Dependence

Microsoft has unveiled the Maia 300, its third-generation, in-house designed artificial intelligence chip. The new chip is intended to serve as the primary AI inference system for Microsoft's own products and services. The Maia 300 is engineered to enhance the cost-efficiency and performance of AI workloads within cloud environments. According to Microsoft, the chip is capable of more effectively handling the latest reasoning models and agentic tasks. Scott Guthrie, executive vice president of Microsoft's cloud group, estimates that Maia 300 will deliver 40 percent more tokens per megawatt compared to existing GPU-based systems. Deployment of the Maia 300 chip into production is scheduled to commence in 2027, with its design prioritizing energy efficiency. This development signifies a move by Microsoft to lessen its reliance on external graphics processing unit (GPU) providers, such as Nvidia, whose chips have been widely utilized for AI computations. The Maia 300 is expected to yield cost savings and performance improvements across Microsoft's cloud offerings, including Azure. Its development forms part of a broader strategy to optimize the company's AI infrastructure and maintain its competitive edge in the rapidly evolving AI market.