2026 is shaping up to be a pivotal year for Samsung’s chip foundry ambitions, thanks to fresh claims that its 2nm GAA (gate-all-around) manufacturing yields are finally stabilizing. If that momentum holds, Samsung could be positioned to attract more high-profile customers at a time when TSMC’s production capacity is increasingly tight and many chip designers are looking for a dependable second source.
After a difficult stretch that reportedly included an estimated 7 trillion won (around $4.85 billion) loss in 2025, Samsung is now said to be on a recovery path. The improved outlook isn’t only about next-generation 2nm. The company is also emphasizing its proven 4nm and 8nm nodes—processes that are widely viewed as more mature and profitable—while it continues to refine 2nm GAA for larger-volume orders.
One market watcher, Kim Dong-won of KB Securities, estimates Samsung’s foundry business could achieve profitability in 2027, as the company stabilizes advanced-node output and increases utilization across its manufacturing lines. The same outlook also points to a strong rebound period ahead, highlighting how Samsung managed to navigate broader semiconductor headwinds that pressured earnings beginning in 2022.
A major part of Samsung’s next step reportedly depends on the progress of its Taylor, Texas facility. The site is believed to represent more than $37 billion in investment, and it is expected to begin EUV-related testing operations in March as Samsung prepares to transition from 4nm production toward 2nm GAA. If Taylor ramps smoothly, it could strengthen Samsung’s case as a serious alternative manufacturing partner for global chip companies.
Interest in Samsung’s advanced lithography is said to include big names such as AMD, while Qualcomm is also being discussed as a possible customer—particularly as work progresses on upcoming Snapdragon flagship platforms. Even if Qualcomm doesn’t commit immediately in early 2026, future product plans could shift depending on performance, yields, cost, and capacity availability across the industry.
One scenario being discussed is a split approach: Qualcomm could choose TSMC’s 2nm N2P node for a higher-end Snapdragon 8 Elite Gen 6 Pro, while using Samsung’s 2nm GAA for the standard Snapdragon 8 Elite Gen 6. That kind of dual-sourcing strategy could help control costs and reduce supply risk—two factors that matter more than ever as cutting-edge nodes become more expensive and in higher demand.
For Samsung, the takeaway is clear: if 2nm GAA yield stability continues and Taylor execution stays on track, 2026 could mark the turning point where its foundry business shifts from heavy losses to a more credible, competitive growth story—powered by both next-gen 2nm and the steady profitability of its established 4nm and 8nm production.
