Two sources familiar with ongoing negotiations say that U.S. chip manufacturers Intel and Advanced Micro Devices — better known as AMD — are entering into extended purchase agreements with Chinese companies that buy server processors for data centers.
The development reflects a wider ripple effect from the artificial intelligence boom. Demand for AI-related hardware has expanded far beyond specialized AI chips, now driving up demand for memory, networking equipment, and the central processing units — or CPUs — that power servers. That shift is giving chip suppliers more leverage to push for long-term supply commitments from buyers.
AI data centers don’t just rely on graphics processors like those made by Nvidia. They also require large quantities of CPUs to handle tasks related to servers, storage, networking, and AI inference workloads.
According to the sources, the agreements being discussed generally lock in how much a customer will buy but leave pricing flexible. Most of these deals cover roughly one year of supply, though Intel and AMD have reportedly discussed commitments stretching two years or more with certain customers. The sources spoke on condition of anonymity because they were not authorized to discuss the matter publicly. Neither Intel nor AMD responded to requests for comment.
The trend mirrors what has already happened in the memory chip market, where AI-fueled shortages pushed customers to commit to longer supply deals.
This marks a notable change for server CPUs, which until recently were far easier to obtain than AI accelerators or memory chips. A tighter CPU supply could drive up costs and slow the pace at which Chinese cloud providers and internet companies roll out new AI services.
Server CPU prices in China are still on the rise, with some products seeing month-over-month price increases of more than 10%, according to one of the sources. Since the beginning of the year, some CPU products have gone up in price by more than 40% in China, that source added.
Earlier this year, Reuters reported that both Intel and AMD had warned Chinese customers of extended wait times for server CPUs, with Intel’s lead times stretching as long as six months for certain products.
The CPU supply crunch is expected to be a major topic when Intel presents its quarterly financial results on Thursday. In April, Intel’s CEO told analysts that demand “continues to run ahead of supply,” particularly for Xeon server CPUs. The CEO also mentioned a multi-year agreement with Google as one of several long-term contracts the company secured during the first quarter.
AMD, which is set to report its own results in early August, has already revised its server CPU market forecast upward to more than $120 billion by 2030, pointing to strong demand driven by so-called agentic AI workloads.
China ranks among the world’s biggest server markets, driven by rapid construction of data center facilities, AI computing clusters, and national computing infrastructure. That building boom has intensified competition for Intel and AMD chips, even as Chinese customers face separate U.S. government restrictions limiting their access to the most advanced AI graphics processors.
