Intel и AMD подписывают многолетние контракты на серверные CPU с дата-центрами Китая

Intel and AMD have signed multi-year agreements to supply server processors to data centers in China — this was reported in July 2026 by the tech outlet TheNextWeb. The contracts lock in supply volumes amid rising prices for server CPUs, which during the AI boom were considered cheap and interchangeable commodities.

Why server CPUs stopped being "boring"

Server processors were long seen as a solved problem: cheap, abundantly available, and easily replaceable. The big AI-boom money flowed into graphics accelerators, while the central processors next to them were treated as a background expense item. This assumption, as TheNextWeb notes, began to fall apart specifically in China, where server CPU prices have been climbing.

The reason is that every cluster of AI accelerators still relies on host central processors: they manage computation, feed data to the GPUs, and hold the rest of the infrastructure together. The explosive growth in the number of AI data centers has driven up demand for "ordinary" x86 chips from Intel and AMD, not just for expensive GPUs.

What multi-year contracts provide

Multi-year supply commitments protect Chinese data center operators from shortages and price spikes. Instead of one-off purchases at spot prices, operators are locking in volumes and terms years in advance — a signal that both sides expect further pressure on the server hardware market.

For Intel and AMD, such deals are also beneficial: guaranteed demand for their server lineup in a major region at a time when both companies are competing for a share of the data center market.

• Players: Intel and AMD — the largest manufacturers of x86 processors for servers
• Region: data centers in China
• Format: multi-year supply commitments instead of one-off purchases
• Deal trigger: rising prices for server processors
• Context: demand for CPU hosts for AI clusters

Why China specifically

China is a special case because the country's access to top-tier AI chips is limited by US export rules. Intel's and AMD's server x86 processors do not fall under these strict restrictions the way flagship AI accelerators do, so stable CPU supply becomes a separate strategic task for Chinese data centers — and a reason to contract volumes in advance.

Despite the development of domestic processors, a significant share of China's server infrastructure remains tied to the x86 architecture of Intel and AMD. Long-term contracts cement this dependency while also giving operators predictability on price and availability.

Why it matters

Server processors were supposed to remain the boring corner of the AI boom — but in

China, that assumption is falling apart.

— TheNextWeb

The AI infrastructure market is no longer splitting neatly into "expensive GPUs" and "cheap CPUs." Shortages and price pressure are spreading to components that hardly anyone used to worry about. When server processors get more expensive and get locked up under multi-year contracts, it's a sign that the AI boom's bottlenecks are spreading across the entire supply chain.

What this means

Intel's and AMD's multi-year deals show that the bottleneck in AI infrastructure is becoming not just the GPUs, but all the hardware around them. If even "boring" server processors are getting more expensive and moving under long-term commitments, it means the AI boom is putting pressure on the entire tech stack, not just on accelerators.