Microsoft hedges its bets: Deploying new racks from both AMD and NVIDIA

Glonghui, July 30th｜According to Kuai Technology, Microsoft CEO Satya Nadella confirmed in the latest earnings call that the Azure cloud platform will be among the first to deploy next-generation rack-scale AI infrastructure based on AMD Helios and NVIDIA Vera Rubin. According to previous reports, AMD Helios, as the brand’s first full-stack rack-mounted AI solution, is priced between $5 million and $5.5 million (approximately 36.24 million to 39.87 million RMB) per Helios rack. In comparison, the expected price of NVIDIA’s second-generation Rubin rack is only $3.5 million to $4 million (about 25.37 million to 28.99 million RMB). This means that for deployments of the same scale, the cost of the AMD solution is about 40% higher than NVIDIA. Industry analysis suggests that Microsoft is willing to pay an additional 40% premium to deploy AMD hardware mainly to replicate its diverse strategy on the AI software side. By simultaneously introducing top-tier solutions from both hardware giants, Microsoft can effectively reduce the supply chain risk of relying solely on NVIDIA and strengthen its bargaining power for Azure cloud services.