Intel, AMD lock in long-term China CPU deals as server processor prices surge

Intel and Advanced Micro Devices are negotiating longer-term purchase commitments with Chinese server customers for data-centre processors, according to two people familiar with the discussions. The agreements typically lock in purchase volumes but not prices, with most covering about a year of supply; however, Intel and AMD have discussed commitments of two years or longer from some customers. This marks a significant shift for server CPUs, which historically have been easier to obtain than AI accelerators or memory chips. The move underscores a broader consequence of the AI boom: demand has extended beyond Nvidia-style graphics processors to encompass CPUs, memory, networking gear, and other infrastructure components, giving suppliers greater leverage to seek long-term deals. AI data centres require large numbers of CPUs to support servers, storage, networking, and inference workloads alongside GPU accelerators. Server CPU prices are climbing steeply in China, with month-on-month increases topping 10% for some products; some CPU products have risen more than 40% since the start of the year. This price surge follows reports earlier in 2024 that Intel and AMD had notified Chinese customers of lengthy lead times for server CPUs, with Intel lead times reaching as long as six months for some products. Tighter CPU supply could raise costs and slow deployment for Chinese cloud providers and internet companies expanding AI services. The trend echoes the memory-chip market, where AI-driven shortage has similarly pushed buyers toward longer-term supply commitments. Intel CEO Lip-Bu Tan told analysts in April that demand "continues to run ahead of supply," especially for Xeon server CPUs, and cited a multi-year deal with Google among several long-term contracts signed in the first quarter. AMD has already raised its server CPU market forecast to more than $120 billion（約19兆円） by 2030, citing strong demand related to agentic AI workloads. China remains one of the world's largest server markets, fuelled by rapid data-centre construction and AI computing infrastructure buildout, which has intensified competition for Intel and AMD processors even as Chinese buyers face separate U.S. restrictions on access to the most advanced AI GPUs. Intel reports quarterly results on Thursday and is expected to discuss CPU supply challenges; AMD is due to report in early August.

The AI boom has created a secondary wave of component scarcity beyond GPUs. While Nvidia's graphics processors have dominated headlines, data centres require far more than accelerators: they need large numbers of CPUs to handle servers, storage, networking, and inference workloads. Intel and AMD, long commodity suppliers in the server space, now find themselves in a seller's market. This mirrors earlier dynamics in memory chips, where AI-driven shortage pushed buyers toward longer-term supply commitments. The shift is particularly acute in China, one of the world's largest server markets, where rapid data-centre construction and AI infrastructure buildout have intensified competition for processors. Chinese customers, already constrained by U.S. restrictions on access to advanced AI GPUs, now face tightening CPU availability—a dual pressure that could slow their AI service expansion.