Intel and others press on TSMC dominance
TSMC still leads advanced foundry capacity, but governments and customers are funding alternatives and Intel is expanding partnerships — including with Google — to advance next‑generation AI and cloud infrastructure. The market is valuing geographic and strategic optionality as much as pure process leadership. (reuters.com, eenewseurope.com)
Taiwan Semiconductor Manufacturing Co. still makes most of the world’s most advanced chips, but Intel and its customers are putting money behind backup options. (reuters.com) Intel said on April 9 that Google will keep deploying Intel Xeon processors in Google Cloud and expand joint work on custom infrastructure processing units for artificial intelligence systems. Intel said the deal is a multiyear collaboration tied to Google Cloud’s C4 and N4 instances, which use Intel Xeon 6 chips. (intel.com, newsroom.intel.com) That partnership comes as Taiwan Semiconductor Manufacturing Co., or TSMC, pushes its A16 process toward production in the second half of 2026 and expands in Arizona. TSMC said on March 4, 2025 that it would raise its planned United States investment to $165 billion, up from $65 billion, for advanced manufacturing in Phoenix. (tsmc.com, tsmc.com) A foundry is a chip factory that makes processors designed by other companies, and TSMC has dominated that business at the leading edge for years. Reuters Breakingviews wrote on April 15 that governments and big customers are now paying for geographic spread and supply-chain insurance, not just the smallest transistor size. (reuters.com) Washington has already put federal money behind that shift. The U.S. Department of Commerce said in November 2024 that TSMC Arizona won up to $6.6 billion in direct funding under the CHIPS Incentives Program. (nist.gov) Intel is trying to turn that policy tailwind into foundry business and product proof at the same time. Intel said Panther Lake, its first client processor built on Intel 18A, is already in production, and it said Clearwater Forest, a server chip on the same process, is being manufactured at Fab 52 in Chandler, Arizona. (intel.com, newsroom.intel.com) TSMC still has the scale advantage. Its Arizona site says the third Phoenix fab is slated for N2 and A16 technologies, but volume production is targeted for the end of the decade, not this year. (tsmc.com) The pressure on TSMC is less about a single rival overtaking it in 2026 than about customers refusing to rely on one geography and one supplier. Google’s deeper Intel tie-up shows how cloud companies are spreading bets across chip designs, factories and regions as artificial intelligence spending rises. (intel.com, reuters.com)
