Intel CEO says CPU demand is so strong the company can serve only 50% of customers
Lip-Bu Tan says CEOs are calling to apologize for being unable to produce enough chips as Intel races to expand manufacturing capacity.
Intel is struggling to produce enough processors to meet surging demand, with CEO Lip-Bu Tan saying the company can currently supply only about half of what its customers are asking for.
“CPU demand is so high that we can only supply 50% of customers,” Tan said during a conversation with Cisco President Jeetu Patel at the Splunk conference in Denver.
The comment offers one of the clearest indications yet of the supply constraints facing Intel as demand for computing infrastructure accelerates with the expansion of artificial intelligence. Tan said that numerous CEOs have been calling him to apologize for not being able to produce enough products.
“Many CEOs are calling me apologizing for not being able to produce enough,” he said.
The shortage marks a striking change for a company that has spent years struggling with manufacturing delays, strategic missteps and intensifying competition in the semiconductor market. Under Tan, Intel has been attempting to rebuild both its product business and its manufacturing capabilities, while simplifying its organization and improving execution.
Intel’s second-quarter revenue rose 25% year over year to $16.1 billion, its strongest quarterly growth in 15 years, with the company saying demand for its products was outpacing its growing supply. Tan has increasingly argued that the expansion of AI beyond model training into inference and agent-based systems is creating a new source of demand for Intel’s traditional strength: CPUs.
In his conversation with Patel, Tan linked that demand to the growing number of AI agents expected to operate across computing systems.
“There are numerous agents, and in the future millions or trillions of agents will require resources,” he said. “Sufficient computing power and security are needed to respond to this.”
The comments point to a broader shift in the way Intel sees the AI market. While the first phase of the AI boom was dominated by demand for GPUs used to train large models, Tan has argued that the expansion of inference and agentic AI will require increasing amounts of general-purpose computing.
But meeting that demand is not simply a matter of producing more chips. Tan said Intel’s decision to remain in the manufacturing and foundry business is tied to the increasingly complicated architecture of modern computing systems, in which processors, memory, input/output components and other silicon must work together.
“Not only design, but the ability to perform manufacturing and advanced packaging is very important. That’s why we are participating [in the foundry business],” Tan said.
He described chip manufacturing as a particularly difficult business, requiring Intel to manage yields, defects, cycle times and variability well enough to make production volumes predictable.
Advanced packaging is becoming just as important, he said, because customers increasingly need different components integrated into a single system.
“Everything is increasingly changing to a system approach and packaging, and this is the future,” Tan said.
That strategy comes with significant financial and operational challenges. Tan said building the manufacturing capabilities required for the business demands substantial capital expenditure and careful execution. He also identified substrates, a critical component in chip packaging, as one of Intel’s biggest constraints.
“Japanese and Taiwanese companies are securing sufficient quantities through prepayment,” he said.
Tan also framed Intel’s foundry ambitions in strategic terms, arguing that dependence on a single supplier is becoming increasingly risky as computing systems grow more complex.
“Packaging that puts CPU, memory, I/O devices, and silicon together is not easy and requires a lot of technology,” he said. “Everything is increasingly changing to a system approach and packaging, and this is the future. That’s why I think relying 95% on one company, especially one based in Taiwan, is very dangerous.”
Intel’s manufacturing strategy has been one of the most closely watched parts of Tan’s turnaround. Intel has moved ahead with plans for high-volume production of its next-generation 14A process in 2028 after securing stronger customer demand, while the company’s foundry business reported second-quarter revenue of $5.77 billion.
At the same time, Intel has been restructuring aggressively. The company had cut its number of vice presidents from about 450 at its peak to roughly 200 and reduced management layers from 12 to six, as Tan seeks to accelerate decision-making and improve execution.
The immediate challenge, however, is less about convincing customers to use Intel products than about producing enough of them.
“It’s not easy and you have to approach it very carefully and it requires a lot of capital expenditure,” Tan said. “The good news is that the situation has improved over the past 18 months.”
Intel currently has a market capitalization of about $574 billion, with its shares up roughly 175% since the start of the year. The stock closed at $108.60 on Friday, however, still well below its $142.35 52-week high reached on June 30. The pullback from that level leaves the shares about 24% below their June peak, even after the extraordinary gains earlier this year.
