Intel is reportedly preparing for another round of CPU price increases scheduled for May, marking the third such adjustment this year. This move follows previous price hikes implemented in February and March, as the company grapples with massive demand and ongoing supply challenges.
The upcoming May price increase is expected to affect Intel’s overall CPU portfolio. This includes processors from the consumer-focused Core Ultra family to the high-performance Xeon server processors. Prices are anticipated to rise by a few more percent, with specific adjustments varying by CPU sector.
This latest adjustment builds upon earlier increases. In February, Intel initiated a first round of CPU price increases, which ranged from 10% to 15%, depending on the specific segment and SKU. Just a month later, in March, the company introduced another hike of approximately 15%. Earlier reports indicated that the consumer CPU sector, including the Core Ultra family, saw a 10% increase during that period.
When considering these consecutive adjustments, the cumulative goal for the price hike is projected to be around 30% higher compared to 2025 pricing. This significant increase underscores the current market dynamics Intel is navigating.
A primary factor contributing to these price adjustments is Intel’s struggle with CPU supply. The company is facing a notable problem with its CPU supply that it cannot address immediately, despite a large portion of its production being handled internally by Intel Foundry.
The manufacturing process itself presents complexities. While Intel Foundry manages the bulk of orders, certain CPUs necessitate silicon from TSMC. This is especially true for multi-die packaging designs, where components are manufactured on both Intel’s proprietary nodes and TSMC’s nodes. Shipping these advanced CPUs becomes impossible until every required part arrives and Intel can complete the assembly using its advanced packaging techniques.
Driving this unprecedented demand for CPUs are AI data centers. These facilities are currently purchasing an extraordinary volume of CPUs, leading to a significant scarcity in the market. Barely any CPUs are left available due to this intensive acquisition.
Historically, in the initial phase of AI data center buildouts, server infrastructure spending heavily favored GPUs. At that time, GPUs outnumbered CPUs by more than 12 units per single CPU on average.
However, the landscape has evolved dramatically. Due to increased computational needs and the sophisticated ways modern AI is being utilized, the CPU-to-GPU ratio has shifted considerably. This ratio has risen significantly, now standing at approximately 1:8, meaning one CPU is typically paired with eight GPUs. Some estimates even suggest this ratio could climb as high as 1:4, indicating four GPUs per single CPU socket.
This fundamental shift in the CPU-to-GPU dynamic translates directly into a multi-fold surge in CPU demand. The current market situation is one where major players, including AMD, Intel, and various Arm-based manufacturers, are finding it difficult to keep pace with the overwhelming requirements.
In response to these challenges, Intel is implementing pricing adjustments. These changes reflect the increasing difficulty in accessing CPUs and serve as a measure while the company’s manufacturing capabilities slowly pick up pace to meet the heightened global demand.