
According to South Korean media outlet Chosun Biz, citing industry sources, Samsung Electronics has further accelerated the break-even timeline for its semiconductor foundry business. The division, which was previously expected to turn profitable between late 2026 and 2027, is now projected to reach profitability as early as the third quarter of 2026. If achieved, this would mark a full recovery roughly four years after recording losses of around one trillion KRW in 2022.
Industry observers attribute this improved outlook primarily to rapid progress in Samsung’s 2nm Gate-All-Around (GAA) process technology and its High Bandwidth Memory (HBM) base die production. Demand for HBM4 base dies is reportedly nearly fully booked, while execution in advanced process nodes has exceeded earlier market expectations. By the first quarter of 2026, Samsung’s 2nm GAA yield is expected to surpass 60%. Although still below the commonly cited 70% threshold for full-scale economic efficiency, analysts suggest the current level is sufficient to support initial mass production and attract additional customers. Samsung also indicated that utilization rates at its advanced process lines have reached peak levels, delivering double-digit year-on-year revenue growth even during seasonal slowdowns.
On the customer expansion front, contracts secured since 2025 are expected to contribute more significantly to revenue starting in the second half of 2026. A key milestone includes a US$16.5 billion long-term supply agreement signed in July 2025 with Tesla, Inc. for its AI6 platform. In addition, Samsung is reportedly involved in manufacturing efforts related to AI accelerator and LPU-related chips for NVIDIA and Groq. Internal projections suggest that Samsung’s 2nm-related orders in 2026 could grow by more than 130% compared with 2025. Beyond these major clients, collaboration scope with technology leaders such as Apple Inc. and Nintendo is also reported to be expanding.
Another structural improvement comes from Samsung’s new fab in Taylor, Texas, operated by Samsung Electronics. The facility is expected to begin full-scale production in 2026, helping to alleviate the burden of fixed costs that previously weighed heavily on the foundry segment’s profitability. As utilization ramps up, depreciation and amortization pressure is expected to be diluted, significantly improving the overall cost structure.
At the same time, broader market conditions are also working in Samsung’s favor. Surging demand for AI accelerators has pushed leading-edge capacity at TSMC close to full utilization. This tightening supply environment is prompting some integrated circuit designers to diversify their manufacturing strategies, with reports indicating that Advanced Micro Devices is evaluating a dual-foundry approach for certain next-generation GPU products, potentially including Samsung as an alternative production partner.