Intel is projecting stronger-than-expected revenue and profits for the third quarter, crediting a wave of demand for server central processing unit chips as the artificial intelligence industry drives a massive expansion of computing infrastructure.
The Santa Clara, California-based company expects third-quarter revenue to land between $15.8 billion and $16.8 billion — well above the analyst consensus estimate of $15.10 billion, according to data from LSEG. Adjusted earnings are expected to reach 38 cents per share, topping the analyst estimate of 27 cents.
For the second quarter ending June 27, Intel reported sales of $16.13 billion — a 25.4% increase — with adjusted earnings of 42 cents per share. Both figures significantly outpaced analyst expectations of $14.42 billion in revenue and 21 cents per share. Adjusted gross margin came in at 41.8%, beating the estimated 38.8%.
A major driver behind Intel’s improved performance is the rise of so-called agentic AI — technology in which autonomous software agents perform tasks like computer coding on behalf of users. This shift has sparked a resurgence in data center CPU demand, which Intel’s own leadership admitted earlier this year caught them off guard, with orders outpacing the company’s production capacity.
Intel shares have slipped more than 25% from a record closing high on June 22, part of a broader pullback in chip stocks. Despite that recent slide, shares are still up more than 170% for the year.
Chief Financial Officer David Zinsner told Reuters in an interview that the surge in demand has led Intel to increase its capital expenditure forecast for this year from $18 billion to $20 billion, with plans to spend “up meaningfully next year” as well.
“That’s signaling the confidence around the growth opportunities for the business,” Zinsner said.
Zinsner also revealed that Intel has entered into a series of long-term agreements with customers covering both data center CPUs and specialized chips called XPUs. Those contracts span three to five years and include a mix of pricing and volume commitments. Still, he urged caution about relying too heavily on such agreements.
“You can’t completely hang your hat on (long-term agreements) because when things change, a lot of times things get renegotiated,” he said. However, “they’re not signing those unless they have real confidence around what they’re going to invest.” He added: “It gives us a pretty good confidence around what we should be planning in terms of output.”
On the financial side, Zinsner said Intel currently holds about $30 billion in cash along with a $10 billion line of credit. He did not rule out a future stock offering, though no specific plans are in place.
“I wouldn’t miss the possibility that we would do that. But no specific plans at this point,” he said.
Intel CEO Lip-Bu Tan has been steering a turnaround effort at the company, which lost ground to Nvidia during the first wave of the AI boom, when graphics processors became the dominant tool for AI workloads.
A key pillar of Intel’s recovery plan is its contract manufacturing — known as its foundry business. The division recently secured Elon Musk’s Tesla as a client for its next-generation 14A manufacturing process, to be used in Tesla’s “Terafab” AI chip project. In April, U.S. President Donald Trump announced that Apple had agreed to have processors made by Intel, though neither company has officially confirmed that arrangement.
Meanwhile, competition in the CPU space is intensifying. Nvidia, which dominates the AI accelerator market, is moving into CPUs with its “Vera” processor, while major tech companies including Amazon and Alphabet continue building their own in-house processors based on Arm technology.
