AMD revamps server CPU market size to 120 billion dollars

On May 5th, after the US stock market closed, AMD announced its financial results for the first quarter of fiscal year 2026, with a record-breaking revenue of $10.3 billion, a year-on-year increase of 38%, exceeding the upper limit of the guidance. The gross margin rose to 55%, and the EPS grew by 43% year-on-year to $1.37, while the free cash flow more than tripled, reaching a record $2.6 billion.

Revenue for the second quarter is expected to be around $11.2 billion, a year-on-year increase of about 46%, higher than market expectations. AMD Chairman and CEO Lisa Su set the tone for this strong data set at the beginning:

These results mark a clear inflection point in our growth trajectory and a structural transformation of our business. The data center is now the main driving force of our revenue and profit growth.

Intelligent Entity AI Ignites Demand, Server CPU Market Expected to Double

The most market-attention increment information comes from AMD’s significant upward revision of the Total Addressable Market (TAM).

Last November, AMD predicted the server CPU market to have a Compound Annual Growth Rate (CAGR) of about 18% for the next three to five years. But in this earnings call, this forecast was directly doubled.

AMD's Chairman and CEO Lisa Su expects the Total Addressable Market (TAM) for server CPUs to grow at an annual rate of more than 35%, surpassing $120 billion by 2030. Su noted:

The expansion of inference and agent-based AI workloads requires CPU computing power far beyond previous expectations — these tasks rely not only on GPUs and accelerators but also require a significant amount of CPU for task orchestration, data movement, and parallel execution.

At the earnings call, Su stated that the company categorized the demand for CPUs into three categories: general computing, CPUs as GPU head nodes, and CPUs servicing agent-based AI workflows. Among these, the fastest growth is in the third category.

She pointed out that in AI infrastructure deployment, the ratio of CPU to GPU is shifting from the past 1 to 4 or 1 to 8 towards almost 1 to 1, and in some high-density agent scenarios, the number of CPUs may even exceed GPUs. Su stated:

Previously, CPUs predominantly served as host nodes for GPUs. Now, while customers plan accelerator purchases, they are also planning the long-term capacity for CPUs in parallel.

During Intel's Q1 earnings call, CEO Pat Gelsinger also indicated that the past ratio of CPU to GPU was 1 to 8, now it's 1 to 4, and it's moving toward 1 to 1, or even better.

This structural change is directly reflected in AMD's order growth rate. In the first quarter, AMD's server CPU revenue increased by more than 50% year-over-year.

The company expects that the year-on-year growth rate of server CPU revenue in the second quarter will exceed 70%, and this strong growth momentum will continue into the second half of 2026 and even into 2027.

Data Center AI Business Accelerates, MI450 Shipments to Start in the Second Half

Regarding Instinct GPUs, AMD disclosed some impactful customer progress. The company not only continues deep cooperation with OpenAI but also secured a mega order from Meta.

Currently, AMD has begun to provide samples of the MI450 series GPUs to core customers and plans to initiate preliminary mass production in the third quarter and significantly ramp up in the fourth quarter. Su revealed:

A key example is our expanded strategic partnership with Meta, deploying up to 6 exaFLOPs of AMD Instinct GPUs, across multiple product generations.

The agreement includes custom GPUs based on the MI450 architecture, co