Intel shares fell another 8% on Monday, extending a 45% collapse in October that stands as the steepest monthly decline on record. The drop followed a Thursday warning that fourth-quarter revenue will land $1 billion to $2 billion below consensus, as PC buyers balk at elevated memory and component costs. Intel still leans on that segment for roughly half its sales, and the upgrade cycle has effectively stalled.
The capital binge has not bought a marquee customer
Over the past five years Intel has deployed about $108 billion in capital expenditures to build a foundry operation that can rival Taiwan Semiconductor Manufacturing Co. The effort has yet to land a single high-profile outside customer for its most advanced process. The factory network still runs primarily on Intel’s own designs, leaving the economics of the model unproven.
A brief rally and a hard reversal
Sentiment flipped in August 2025 when the U.S. government took a 10% stake and Nvidia committed $5 billion. The stock multiplied sixfold over the next ten months, peaking in June on hopes that Intel’s server CPUs would power the agentic-AI wave and that its fabs would finally attract outside work. The October guidance cut unwound that narrative in a single session, sending shares down 29% on Friday, their worst day since the 1970s.
Capacity choices that undermine the pitch
Data-center revenue surged 78% year-over-year to $7.3 billion in the third quarter, and chief executive Lip-Bu Tan sees the total addressable market for server processors expanding fivefold to roughly $200 billion within five years. Yet Intel says it cannot secure enough wafers, substrates and memory to meet that demand. In response, it is redirecting some Oregon capacity from the 18A node, the process meant to woo foundry clients, to the older Intel 3 node used for its own data-center chips. David Yoffie, a Harvard Business School professor and former long-time board member, said the move signals to prospective customers that Intel may lack spare capacity, damaging its credibility as a merchant supplier. He warned that without a large-scale foundry win within a year, the strain on Intel’s economics will become more pronounced.
Execution as the only remaining argument
An Intel spokesperson said customer confidence rests on execution, citing progress on 18A and the upcoming 14A node along with continued customer engagement. The statement did not address the capacity reallocation. With the PC franchise shrinking and the foundry bet still unvalidated, the company is now asking the market to trust a roadmap that keeps moving the goalposts.