Kevin Scott, Microsoft's Chief Technology Officer, disclosed that the company intends to predominantly utilize self-developed chips in its data centers moving forward, in a bid to curtail its reliance on suppliers like NVIDIA and AMD. At present, Microsoft's data centers are heavily reliant on chips from NVIDIA and AMD, as they seek the 'optimal cost-effectiveness' for individual chips. Nevertheless, they have commenced a gradual shift towards incorporating self-developed chips. For instance, they launched the Azure Maia AI accelerator and Cobalt CPU in 2023, and are presently in the process of developing next-generation semiconductor offerings. Scott emphasized that giving precedence to chips is an integral facet of Microsoft's strategy to construct a comprehensive data center system. This is because self-developed chips are more adept at catering to their specific requirements and boosting operational efficiency. Furthermore, with tech behemoths pouring over $300 billion into the artificial intelligence sector this year, the problem of a computing power shortage has come to the fore. Despite Microsoft's efforts to expand its data centers, it continues to grapple with meeting the burgeoning market demand.