A Guardian investigation reveals a gap between Microsoft’s public claims about its AI capacity and the actual number of advanced chips deployed in its data centers. The problem stems from data center construction falling behind schedule, leaving thousands of chips sitting in inventory. Microsoft has ordered over 300,000 next-generation AI chips from TSMC for delivery in 2027, while developing its own Maia chips. This situation creates constraints for its commercial products like Copilot and Azure services using OpenAI models.
Source: Read the original article
