Samsung drops 4 nm and ramps 2 nm GAA at its Taylor, Texas fab, scaling to 50–100k wafers, with Exynos 2600 and a $16.5B Tesla AI6 deal to rival TSMC.
© A. Krivonosov
Samsung is moving to strengthen its standing in the global chip race with a sweeping reset of its U.S. manufacturing strategy. The company is dropping plans for a 4 nm process in favor of a more advanced 2 nm GAA node, a shift that could bring it closer to TSMC, the current industry leader. At the heart of the plan is the Taylor, Texas fab, already outfitted with EUV lithography tools needed for high-density designs.
The site was initially slated for 4 nm production, but the course has changed. Samsung is now prioritizing mass production of 2 nm chips despite the technical hurdles. Output is being scaled up—from an initial 20,000 to 50,000 wafers per month—with the potential to reach 100,000, positioning the Texas plant as one of the company’s core manufacturing hubs.
The first chip on the new node is expected to be the Exynos 2600. Early yield figures were not especially high, but they have been improving, making the effort increasingly viable. In parallel, Samsung is accelerating work on the next iteration, SF2P+, which could surface within the next couple of years. The pivot reads as an attempt to skip an interim step and meet rivals at the leading edge.
Momentum is also fueled by a major deal with Tesla worth about $16.5 billion: Samsung will produce AI6 chips for autonomous driving systems. There are additional orders from Chinese companies in the crypto-hardware space, though some of those contracts will likely be fulfilled outside the U.S.
Taken together, these moves suggest Samsung aims not merely to catch up to TSMC but to go head-to-head in advanced process technology. If the pace holds, the American facility could become a cornerstone of its global strategy and nudge the balance of power in semiconductors over the next few years.