A significant shift in Voltage Regulator Module (VRM) technology, crucial components for AI servers, is causing power shortages and extending lead times for these parts to six months. This change is driven by the increasing power demands of advanced AI chips, requiring more sophisticated and robust power delivery systems within servers.
This development matters because VRMs are essential for converting and regulating power to AI processors, ensuring stable operation. The current shortages and extended lead times signal potential bottlenecks in the production and deployment of new AI servers and data centers, which could slow down the expansion of AI infrastructure.
The mechanism involves AI chips demanding higher power, necessitating new VRM designs capable of handling increased current and efficiency. Manufacturers are struggling to ramp up production of these new VRM types quickly enough to meet the surging demand, leading to the reported supply constraints and longer delivery schedules.
This situation primarily impacts companies involved in AI server manufacturing, such as Nvidia (NVDA) and AMD (AMD), who rely on these servers for their chips. Data center operators like Microsoft (MSFT), Amazon (AMZN), and Google (GOOGL) could also face delays in expanding their AI capabilities, potentially affecting their cloud and AI service offerings.
An AI breakdown of exactly what changed and who it moves.