AI Is Fueling Demand Intel Didn’t Expect

Intel walked into its latest earnings call with a problem most chipmakers would love to have… demand is running ahead of supply.

Data center demand is pushing Intel ahead of expectations

The company reported fourth-quarter revenue of $13.7 billion, slightly ahead of expectations, with earnings also beating forecasts. But the bigger story was how much demand, especially from AI-driven data centers, has caught Intel off guard.

CEO Lip-Bu Tan acknowledged that the company is still struggling to keep up. 

“We delivered these results despite supply constraints, which meaningfully limited our ability to capture all of the strengths in our underlying markets,” he said. “We are working aggressively to address this and better support our customers’ needs.”

That pressure is coming almost entirely from the data center side. Intel’s Data Center and AI group grew 9% year over year to $4.7 billion, which is its fastest sequential growth in years. According to CFO David Zinsner, that growth could have been even higher if Intel had more chips to ship.

AI workloads are changing demand patterns

What’s different this time is the type of demand Intel is seeing. It’s not just traditional server refresh cycles. It’s AI workloads that rely heavily on CPUs to manage data movement, orchestration, and inference.

“The continuing proliferation and diversification of AI workloads is placing significant capacity constraints on traditional and new hardware infrastructure,” Tan said, adding that CPUs remain very central to how AI systems actually function.

Zinsner went a step further, pointing out that AI agents are driving a shift in how computing demand shows up. “The world is shifting from human-prompted requests to persistent and recursive commands driven by computer-to-computer interactions,” he said. “The CPU’s central function coordinating this traffic will drive not only traditional server refresh, but new demand that grows the installed base.”

The result is a supply crunch that Intel admits it didn’t fully anticipate. Zinsner said the company had expected higher core counts, not a surge in overall unit demand, especially from hyperscalers. “It has rapidly increased over the third and fourth quarters,” he said, noting that the trend now looks like it could last for years.

PCs take a back seat in favor of data centers

That surge is forcing tradeoffs. Intel is funneling as much supply as possible toward higher-margin data center chips, even if it means tighter availability on the PC side.

The company’s Client Computing Group saw revenue fall by 7 percent, and Intel acknowledged that lower-end PC processors are receiving lower priority. 

“We’re shifting as much as we can over to data center to meet the high demand, but we can’t completely vacate the client market,” Zinsner said. “So we’re trying to support both as best we can and obviously work our way out of this supply issue. I do believe that the first quarter is the trough. We will improve supply in the second quarter.” 

At the same time, memory and component shortages are making things harder. As Zinsner put it, “industry-wide supply for key components like DRAM, NAND and substrates have come under increasing pressure.”

That combination helps explain why Intel expects first-quarter revenue to dip before improving later in the year. The company said it expects supply conditions to gradually ease after Q1.

A cautious optimism on Wall Street

Despite the near-term “oops,” investor sentiment has improved. Intel shares shot up in 2025, and analysts point to renewed confidence in Tan’s turnaround strategy.

“It’s the most optimistic I think people have felt about the company in a long time,” said Gabelli Funds analyst Ryuta Makino. “That’s really the big Intel bull case here – I think there will be at least a double-digit server CPU (central processing unit) price hike in 2026.”

Intel still has plenty to prove, especially as it works through manufacturing challenges and ramps its next-generation process technology. But for the first time in a while, the issue isn’t whether customers want Intel’s chips. It’s whether Intel can make enough of them fast enough.

And in the current AI cycle, that’s a very different problem to have.

Last year’s push around edge AI signaled that demand was about to accelerate beyond the data center, and today that prediction is playing out in real time. As AI workloads continue to spread across environments, Intel is now working through the reality of supporting that growth at scale, from centralized infrastructure to the edge.