Microsoft Plans In-House AI Chips To Reduce Nvidia AMD Dependence

Microsoft is intensifying its efforts to develop proprietary AI chips, aiming to reduce reliance on industry leaders Nvidia and AMD for its data centers. The company has already integrated custom chips like the Azure Maia AI Accelerator and the Cobalt CPU while pioneering microfluidics cooling technology to address overheating issues. Microsoft's Chief Technology Officer, Kevin Scott, confirmed that the long-term goal is to predominantly use in-house chips and develop a comprehensive data center system including networking and cooling tailored for AI workloads. Despite this push for self-reliance, Microsoft continues to collaborate with Nvidia, exemplified by a recent strategic partnership involving the deployment of Nvidia systems with OpenAI, contributing to a strong financial performance with an 18% revenue growth in 2025. Nvidia remains the dominant player in AI semiconductors with robust financial health and strategic alliances, underscoring its critical role in AI infrastructure development. The competition among tech giants to enhance AI capabilities through custom chips and massive capital investments highlights the sector's rapid growth and the pressing demand for innovative computing solutions.