Intel’s Rally Needs a Real Foundry Customer

Intel’s stock was trading near $97 on September 16, down from the $102.94 level cited in the latest Wall Street debate, but the retreat does little to deflate the central question: when does the company’s foundry ambition become a business rather than a capital-intensive promise?

The answer is not another technology demonstration. It is a binding, high-volume commitment from a major outside chip designer to use Intel’s leading-edge process, particularly 18A or the coming 14A node.

Intel has made genuine progress. In its second-quarter report, the company said Intel Foundry revenue rose 31% to $5.8 billion, while Xeon 6+ became its first server processor built on 18A. Intel also said a subset of its Core Ultra Series 3 processors had entered high-volume manufacturing using ASML’s High-NA EUV technology. The company and ASML separately disclosed that Intel had processed more than one million wafers on High-NA systems, a milestone that underscores its manufacturing lead.

But impressive engineering statistics do not automatically produce attractive foundry economics. Intel Foundry still generated only $293 million of external revenue in the quarter and posted a $2.1 billion operating loss, according to reporting based on the company’s results. Fortinet’s Security Processor 6 is a useful first customer win, though it is being built on Intel 4 rather than the newest 18A process.

That distinction matters because Intel is spending more than $20 billion on capital expenditures in 2026 while the stock has already climbed more than 300% over the past year, according to 24/7 Wall St. Investors are no longer paying for a distressed manufacturer with optionality. They are paying for a semiconductor platform that can challenge Taiwan Semiconductor Manufacturing in contract production.

Nvidia’s investment in Intel adds strategic validation, but it does not commit Nvidia to manufacture chips there. The company still needs a marquee customer willing to reserve meaningful capacity, accept Intel’s yields and design ecosystem, and sign a multiyear commercial agreement.

Until that happens, Intel’s rally remains a bet on execution. A disclosed 18A or 14A anchor order would change the argument quickly. Without one, the stock’s valuation leaves little room for another delay.