Intel shares fell 45% in October, their steepest monthly decline in FactSet records. The stock dropped 29% in a single session after the company’s revenue outlook for the current quarter came in below analysts’ expectations, its worst daily performance since the 1970s, and fell about 8% again in Monday trading. A weak PC market has added to the pressure, while investors are reassessing whether Intel can turn its foundry operation into a meaningful source of outside revenue.
Intel has spent heavily over the past five years to catch up with Taiwan Semiconductor Manufacturing Co. (TSMC), with cumulative capital spending of $108 billion. But its most advanced manufacturing process still lacks a major external customer, and its fabs primarily serve Intel’s own products. The shares rallied after the U.S. government took a 10% stake in the company in August 2025 and Nvidia later invested $5 billion. They rose about sixfold over 10 months, peaking in June this year. Since then, optimism about Intel’s AI business and foundry plans has been tested by the latest revenue forecast.
Higher component costs weigh on PC outlook
Intel expects current-quarter revenue to come in $1 billion to $2 billion below analysts’ estimates. Rising prices for memory and other components are weighing on the PC market, which accounts for about half of Intel’s revenue. Consumers are also delaying computer upgrades, adding to demand pressures. The shares closed down 29% at $59.97 on Friday. They were trading around $60.13 on Monday, down 8.8% from the previous close.
Data-center revenue has been moving in the opposite direction. It rose 78% year over year in the third quarter to $7.3 billion. Chief Executive Lip-Bu Tan said he expects the global server-processor market to grow as much as fivefold over the next five years, reaching about $200 billion. Intel, however, says it does not have enough capacity to meet demand, citing constraints on wafers, substrates and memory. Those shortages are also affecting PC sales.
Intel 3 capacity shift puts 18A plans in focus
To ease the supply squeeze, Intel plans to shift some capacity at its Oregon factory from its advanced 18A process to the more mature Intel 3 process, increasing wafer supply for data-center products. The move puts the balance between meeting demand for Intel’s own chips and serving foundry customers under scrutiny. The 18A process is a key part of Intel’s effort to attract outside manufacturing business.
David Yoffie, a Harvard Business School professor who served on Intel’s board for nearly 30 years, said prospective customers may question whether the company has enough spare capacity beyond its own needs to be a dependable supplier. He also said the financial strain of the foundry business could become more apparent if Intel has not landed a large customer within a year. Tan said on Thursday that the company would announce “several customers” over the next two quarters.
Intel’s own 18A chips have entered high-volume production in Arizona, where another fab is under construction. In Ohio, the first structural beam was recently installed at a new plant, though the project is nearly five years behind its original schedule. Former CEO Pat Gelsinger championed heavy investment in the foundry business. Tan has taken a more cautious approach to spending, preferring to expand investment once customer demand is clearer.
Potential customers weigh capacity and competition
Intel designs chips as well as offering advanced manufacturing and packaging. Mizuho Securities analyst Jordan Klein said in-house manufacturing gives Intel greater control over supply, but limited capacity could also cost the company revenue. Some chipmakers, including Broadcom and Qualcomm, may be reluctant to outsource production to a company that competes with them in product markets.
Yoffie said separating the foundry business could help address conflicts of interest. Klein said AMD may be the only major chipmaker clearly unwilling to use Intel’s manufacturing services, given their direct competition in CPUs and concerns involving competition, confidentiality and privacy. AMD was scheduled to report third-quarter results after the market close on Tuesday.
Changes in AI workloads have also brought renewed attention to Intel’s CPUs. GPUs from Nvidia and AMD play a prominent role in large-scale computing tasks such as model training and chatbot inference. Intel CPUs can also run the virtual machines used by agent applications, including Meta’s Muse. Whether Intel can turn rising server-processor demand into steady orders will depend on its ability to supply customers and on their willingness to choose the company.
Yoffie also noted that U.S. capacity may account for only 20% to 25% of TSMC’s global production, and that building it will take years. For Intel, progress in winning external customers, allocating 18A capacity and completing factory projects will provide concrete tests of whether its foundry plans can be delivered.