Microsoft appears to be facing challenges in deploying artificial intelligence infrastructure at the pace suggested by its public statements and financial investments, according to an investigation that examined discrepancies between company announcements and actual chip installations.
The technology company targeted having 1.8 million advanced AI chips operational in its datacenters worldwide by the end of 2024. Internal documents reviewed indicate the company currently has 2.2 million such chips installed globally. This figure falls considerably below what some industry experts anticipated given Microsoft’s stated $280 billion expansion initiative. The gap between public statements about datacentre capacity additions and actual operational chip deployments suggests either that newly built facilities are not yet fully operational or that they lack sufficient hardware inventory.
Microsoft has publicly committed to doubling its global datacentre footprint by mid-2027 and claims to have added 5 gigawatts of capacity over the past two years. However, the actual operational status of these facilities remains uncertain. Academic researchers who have analyzed Microsoft’s sustainability reports, which are independently audited, suggest the company’s AI capacity may be substantially lower than its own public claims indicate. One expert noted that while announcing power capacity increases may occur within a single quarter, actually bringing that capacity online and deploying it for computing proves far more difficult in practice.
The opacity surrounding chip procurement and deployment complicates efforts to assess progress in artificial intelligence development broadly. Nvidia, which manufactures these specialized processors, closely guards information about production volumes and customer identities. Major technology companies similarly do not publicly disclose their chip inventories, making independent verification nearly impossible. Microsoft disputed the Guardian’s calculations but did not specify which figures were inaccurate or provide clarification regarding discrepancies between announced capacity and verified installations.
Several factors may contribute to the apparent shortfall. Some of Microsoft’s largest datacentre projects appear only partially operational, and the company’s commercial arrangements with OpenAI may account for portions of its infrastructure deployments not reflected in analyzed documents. Industry sources suggest the total number of AI chips at Microsoft has shown minimal growth over the past year.
Article Attribution | Read More at Article Source
Article summary produced by Claude AI