Intel Flags AI Server Chip Supply Crunch, Forecasts Weak Quarter; Shares Slide

Intel said on Thursday that it is struggling to meet rising demand for server processors used in artificial intelligence data centres, while issuing a revenue and profit forecast that fell short of market expectations. The outlook sent the company’s shares down about 13% in after-hours trading.

The cautious forecast highlights Intel’s ongoing challenge in navigating the fast-changing global semiconductor market, where product cycles are long and manufacturing decisions are often made years in advance. While Intel has recently introduced a new laptop processor aimed at regaining leadership in the personal computer segment, the company is now grappling with unexpected pressure on its data centre business.

Intel executives said they were caught off guard by a sharp surge in demand for server central processing units (CPUs) that work alongside AI accelerators in data centres. Despite operating its factories at full capacity, the company has been unable to produce enough chips to meet customer needs, resulting in missed high-margin sales opportunities.

“In the near term, I’m disappointed that we’re not able to fully satisfy demand in our markets,” Chief Executive Officer Lip-Bu Tan told analysts during a post-earnings conference call.

For the current quarter, Intel projected revenue in the range of $11.7 billion to $12.7 billion, compared with analysts’ average estimate of $12.51 billion, according to data compiled by LSEG.

Investors had been betting that massive investments by global technology companies in AI-driven data centres would fuel stronger demand for Intel’s traditional server CPUs, which are typically paired with market-leading graphics processors from Nvidia.

However, Intel’s Chief Financial Officer David Zinsner said even large cloud service providers were unprepared for how quickly AI workloads would strain existing infrastructure.

“They were all somewhat surprised by how fast demand ramped up,” Zinsner said, adding that customers were forced to upgrade older chip fleets due to declining networking performance.

Zinsner also noted that while Intel owns and operates its own factories, shifting production to different types of chips involves a time lag. The company had not anticipated such a rapid change in data centre demand when planning its manufacturing mix.

Turnaround under way, but pressure remains

After years of strategic missteps that weakened Intel’s position in the fast-growing AI chip market, Tan has been pursuing a turnaround strategy focused on cost reductions, fewer management layers, and a streamlined product roadmap.

Intel has also slowed heavy investment in its next-generation manufacturing process, known as 14A, while waiting to secure commitments from major external customers. Tan said two customers are currently evaluating the technical specifications of the 14A process, with decisions expected in the second half of the year.

Zinsner added that Intel’s capital expenditure may now remain steady, reversing earlier expectations of a decline.

Intel has begun shipping its new “Panther Lake” PC chips, the first products manufactured using its critical 18A process technology. Analysts have warned that the production ramp-up could weigh on margins, particularly as yields remain below ideal levels.

While Intel has said yields are improving month by month, Tan acknowledged during the earnings call that performance is “in line with internal expectations” but “still below where I want it to be.”

At the same time, a global shortage of memory chips has pushed up component prices, making personal computers more expensive. Zinsner said memory supply is likely to be tight in the first quarter before improving in the second.