Intel’s Terafab shifts narrative

Intel’s April 7 decision to join Elon Musk’s Terafab chip project gave its foundry push a marquee customer and a fresh burst of credibility. (usnews.com) Reuters reported Intel will work with SpaceX and Tesla on processors for robotics and data centers at Terafab, a planned Austin, Texas, complex Musk said could target 1 terawatt a year of compute. Intel shares rose more than 2% after the announcement. (usnews.com) Yahoo Finance said the deal landed as Intel stock had climbed nearly 50% in April, adding to a run of announcements that investors read as evidence the company’s contract-manufacturing business was becoming easier to believe. (finance.yahoo.com) A foundry is a chip factory for outside customers, not just for its own designs. Intel has spent the past five years trying to become that kind of supplier after Taiwan Semiconductor Manufacturing Co. pulled far ahead in advanced production. (intel.com) (tsmc.com) Intel’s current sales pitch centers on a manufacturing step called Intel 18A, which the company says is ready for customer projects and uses two new tricks: RibbonFET transistors and PowerVia backside power delivery. Intel says 18A offers up to 15% better performance per watt and 30% higher chip density than Intel 3. (intel.com) Intel has also been using its own products to prove the process can run at scale. In October 2025, it said Panther Lake personal-computer chips were already in production on 18A and that Arizona’s Fab 52 would reach high-volume 18A manufacturing later that year. (intel.com) The customer list has widened beyond Intel’s own products. At Intel Foundry Direct Connect in April 2025, the company put Microsoft, MediaTek and Qualcomm executives on stage and said lead customers were already working with its next process, Intel 14A. (intel.com) Reuters Breakingviews said those moves give chip buyers more optionality, especially as artificial-intelligence demand strains the supply chain. But the column also said Taiwan Semiconductor still holds the scale advantage, the customer depth and the process lead that new challengers must close. (reuters.com) Taiwan Semiconductor’s own roadmap shows why that gap is hard to erase. The company says its 2-nanometer process entered volume production in the fourth quarter of 2025 and that N2P volume production is scheduled for the second half of 2026, while A16 is positioned above that for high-performance chips. (tsmc.com) That leaves Intel in a more credible position than it occupied a year ago, but still in a catch-up race where one deal does not equal manufacturing maturity. The next test is whether 18A can move from internal proof points and headline customers to steady, repeatable output for outsiders. (intel.com) (reuters.com)