Intel and AMD Brewing Q3 New Round of CPU Price Increases

The demand for AI computing power continues to explode, and the CPU market is experiencing a new round of price increases.

According to a report by Taiwan's "Business Times" on April 22, industry insiders from the ODM (Original Design Manufacturer) sector have stated that since March of this year, the prices of consumer CPUs have risen by 5% to 10%, while server CPUs have seen an even greater increase of 10% to 20%. Supply chain sources have revealed that major international manufacturers are contemplating initiating a new round of price hikes in the third quarter.

The core logic driving this round of price increases involves two main factors: firstly, the rapid warming of AI server demand, which is driving the procurement of core computing components; and secondly, the highly concentrated advanced process capacity, making it difficult for the supply side to respond to the increased demand in real time.

Intel, AMD Take Turns, Price Increase Becomes Industry Consensus

Intel made the earliest move. It has been reported that Intel increased the prices of PC CPUs in March this year, and further raised the prices of server CPUs on April 1, leading to a recovery in gross margins in the second quarter. The market anticipates that there will be room for another increase of about 8% to 10% in the second half of the year.

On AMD's side, market news indicates that their server CPU product line is expected to experience price increases in both the second and third quarters, with a total cumulative increase expected to reach 16% to 17%.

The actions of these two major CPU giants signify that the price increase is not an isolated phenomenon, but rather a result of industry-level supply and demand imbalances.

Capacity Bottleneck is the Core Contradiction

The root cause of the CPU price increase lies in the production capacity, not in any unusual fluctuations in demand.

As 2-nanometer and 3-nanometer processes enter the mass production and expansion phase, AI chips, GPUs, TPUs, and CPUs are competing for limited wafer foundry production capacity on the same production line. Supply chain insiders have stated straightforwardly: "Currently, CPUs are still in a severe supply shortage, and the end to price increases is not yet in sight."

The direction of TSMC's capacity allocation is seen as a significant guiding indicator. Industry insiders in wafer foundry have noted that TSMC's continued investment in 3-nanometer capacity is driven by the concurrent surge in demand for CPUs and AI ASICs. Currently, Intel and AMD's mainstream generation CPUs, as well as Nvidia's upcoming Vera CPU, all utilize the 3-nanometer process - indicating extremely fierce competition at the same node.

Intel Invests $14.2 Billion to Regain Control Over Production Capacity

Faced with limited external production capacity, Intel has chosen to build its own capacity.

Reports indicate that Intel recently announced it would purchase 49% of the shares of Fab 34, a wafer factory in Ireland, for $14.2 billion, thereby regaining control over the factory's production capacity. Fab 34 is a key production base for Intel's 4 and 3 processes and is currently one of the most productive advanced process wafer factories.

The market interprets this move as: Intel is preparing in advance for the continuous surge in CPU demand in the AI inference era, in order to reduce reliance on external foundries.

Supply-Demand Gap May Persist Until 2027

Analysts from institutions believe that between 2026 and 2027, the CPU market will maintain a highly scarce situation, with the key constraint being production capacity, not demand.

Reports point out that as long as AI infrastructure continues to expand and advanced process and packaging bottlenecks are not effectively alleviated, the trend of rising CPU prices will continue, leading to comprehensive benefits throughout the wafer foundry, packaging testing, and equipment material supply chains.

For the market, the duration and magnitude of this round of price increases will largely depend on the pace of production capacity expansion by wafer foundry companies such as TSMC, as well as the actual speed of AI computing power investment implementation.