Intel CEO Lip-Bu Tan forecasts over 6x growth in server CPU market

For the past few years, the AI boom has belonged to the GPU. Intel CEO Lip-Bu Tan thinks the humble server CPU is about to get its turn.

Tan estimates that agentic AI could drive the server CPU market to more than $200 billion within five years. For a chip category many had written off as the supporting cast, that is a big promotion.

The forecast lands as Intel says it can’t make enough of these processors to meet demand it already has.

From sidekick to co-star

According to the research findings, the server CPU market sits at approximately $30-35 billion in 2025. Tan’s projection has it climbing past $200 billion by 2030.

That path implies a compound annual growth rate of more than 45%.

The engine behind the forecast is agentic AI. These are systems that don’t just answer a prompt. They plan, coordinate and carry out multistep tasks, which leans heavily on orchestration and inference.

That shows up in hardware ratios. Tan described a move away from the old setup of roughly one CPU for every 7-8 GPUs. The emerging balance is closer to 1:1, and in some cases tilts in the CPU’s favor.

Intel’s numbers are already moving

Intel’s own results suggest the trend is not purely theoretical. Its Data Center and AI segment posted revenue of $6.3 billion in Q2 2026, a 59% jump year over year.

The company credits rapid adoption of its Xeon 6 processors for the surge. It also points to a deliberate shift toward chips with higher core counts, which fit the parallel, coordination-heavy work that agents generate.

The catch is supply. Tan revealed that Intel can currently fulfill only about half of the demand for CPUs aimed at these advanced workloads, as of September 2026.

He also said numerous CEOs have been reaching out to him directly in search of more chips.

Intel is not alone in seeing this. The research findings note similar growth trends at AMD, along with rising market share for Arm-based designs.

The comeback story behind the forecast

Its strategy now rests on a few pillars. The 18A process node is central to its manufacturing ambitions. Custom ASIC work is another, with the company eyeing a custom AI chip opportunity it pegs at more than $100 billion.

Outside voices are starting to buy in. New Street Research recently upgraded Intel and projected the company might achieve up to $50 billion in server CPU revenue by 2030.

That estimate assumes Intel captures a 20-25% share of a market New Street anticipates will exceed $220 billion.

What this means

For chipmakers, constrained supply cuts both ways. Analysts project that semiconductor bottlenecks could last several years. That environment tends to support higher average selling prices.

The risk is that Intel can’t capture the demand it is forecasting. Filling only about half of current orders leaves room for AMD and Arm-based rivals to win customers who simply can’t wait.

Investors should also treat the $200 billion figure as a forecast, not a booking. It depends on agentic AI scaling as fast as its backers hope. A projection from a company CEO naturally reflects that company’s view of its own opportunity.

Still, the signals line up in an interesting way. A CEO forecast, a 59% revenue jump, a supply shortfall and a third-party upgrade all point in the same direction. What to watch next: whether rivals report similar shortages, whether CPU prices firm up, and whether Intel can close the gap between the orders it receives and the chips it ships.