With TSMC's advanced capacity booked out on AI demand, Samsung is winning more foundry orders — reportedly from Google, AMD and BYD per Nikkei — as customers diversify; its 2nm yields and rising orders could return the loss-making foundry to profit, though it trails TSMC badly.
Original market reporting from the FXMARE News Desk, produced under the FXMARE editorial policy. It reports facts only and is not investment advice.
Samsung Electronics is reeling in a wave of fresh chip-manufacturing business as the industry's reliance on a single dominant foundry runs up against its limits. With Taiwan's TSMC unable to keep pace with surging demand for artificial-intelligence processors, a growing roster of major technology customers is turning to the South Korean firm as a backup source, according to a report cited in the headline from Nikkei, which named Google, AMD and Chinese automaker BYD among those steering more orders Samsung's way.
The shift stems from a simple bottleneck. The world's leading contract chipmaker has effectively sold out its most advanced production lines, with reporting suggesting its cutting-edge capacity is spoken for well into the latter part of the decade as AI accelerators, cloud-data-center silicon and high-end mobile chips compete for the same scarce slots. One large chip designer has publicly described the leader's output as a chokepoint for the entire supply chain, and that scarcity has forced even its biggest clients to hunt for additional manufacturing partners.
That dynamic has handed Samsung, the only other company capable of producing chips at the most advanced two-nanometer level, a long-sought opening. AMD has reportedly placed a two-nanometer order with the Korean firm and is weighing a split-supplier approach for part of its next generation of graphics processors. Google, meanwhile, is said to be in discussions for Samsung to build a memory-interface component of an upcoming custom AI processor, with the Taiwanese rival still slated to make the main computing core. The company has also been linked to work for Nvidia and Tesla, marking a notable expansion among precisely the kind of marquee customers it had long struggled to land.
For Samsung's foundry arm, the momentum could mark a turning point. The unit has been losing money since 2023, and management is targeting a return to profitability this year, helped by improving yields on its two-nanometer process and rising utilization as more orders flow in. Reports point to yields around the halfway mark ahead of full-scale production slated for the second half of the year, with advanced-node order volumes expected to climb sharply from a year earlier. The company's chip chief recently met Nvidia's Jensen Huang in Seoul to deepen their foundry ties.
Even so, the gap with the market leader remains vast. Industry data put the Taiwanese firm's share of the pure-play foundry business at roughly three-quarters in the first quarter, against single digits for Samsung, underscoring how far the challenger has to climb. Analysts also caution that winning trial orders is not the same as locking in sustained business: customers will keep a close eye on whether Samsung can deliver consistent yields at scale, since splitting production across multiple manufacturers adds cost and complexity and can weigh on output.
For the broader market, the trend signals a gradual loosening of one company's grip on the most strategically important corner of the technology supply chain. As AI demand keeps straining capacity, the willingness of giants like Google and AMD to diversify their suppliers points to a structural shift that could reshape the competitive landscape in advanced chipmaking. For Samsung, each new contract is a chance to prove it belongs in that conversation; for its customers, it is increasingly a matter of securing enough supply to feed their AI ambitions.
Disclaimer. This is an editorially-reviewed FXMARE news report for informational purposes only. It is not investment advice or a recommendation to trade. Markets can move quickly — always do your own research before trading.