Samsung Foundry Capacity Utilization Rate Expected to Reach 100% This Year, Countdown to Profitability Begins

Samsung Electronics' foundry business is approaching a critical turning point. Driven by demand for HBM base chips and major US tech clients, the division's Capacity Utilization Rate is expected to touch 100% in the second half of this year. The chronic loss structure that has plagued it for over a year may thus reach an inflection point.

According to a report by the Korean media outlet Chosun Ilbo, Samsung Foundry's current Capacity Utilization Rate is estimated to be in the 70% to 80% range. Given the existing order backlog and contract progress, achieving full production within the year is "almost a foregone conclusion." During its second-quarter earnings briefing, Samsung Electronics stated that utilization rates across all process nodes had improved compared to the same period last year, with advanced processes below 8nm having "reached peak levels." The company also projected that order volume for its 2nm projects this year would more than double compared to the previous year. While noting the difficulty in providing an exact timeline for returning to profitability, the company stated that it is "achievable in the near future."

Market analysts expect that non-memory businesses, including foundry and System LSI, will return to profitability as early as the third quarter or at the latest by the fourth quarter. If the normalization of capacity utilization combines with the effect of rising wafer prices, the pace of profit improvement could exceed market expectations.

Dual Drive from HBM and AI Demand Accelerates Utilization Climb

The core drivers behind the rebound in capacity utilization come from two main lines: First, the adoption of the 4nm process for sixth-generation High Bandwidth Memory (HBM4) directly transmits the memory boom cycle to the foundry business; second, Cloud Service Providers (CSPs) and customers in AI and High-Performance Computing (HPC) are continuously expanding their demand for the 2nm process, with project negotiations underway with major clients such as Broadcom.

This background is particularly noteworthy—since 2024, Samsung Foundry's capacity utilization rate has long remained below 50%, meaning high fixed costs could not be effectively covered, leading to sustained losses. The semiconductor foundry business requires massive equipment investment; once production lines sit idle, losses accumulate rapidly over time. Conversely, once utilization returns to normal, new revenue translates directly into profit.

Diversified Order Sources, but 2nm Yield Remains a Key Threshold

In terms of customer structure, Samsung Foundry is accelerating efforts to move away from excessive reliance on a single major client. Tech giants such as Qualcomm, AMD, and Google have successively entered negotiations, and Tesla is reportedly planning to have its next-generation chips produced using the 2nm process, significantly thickening the medium-to-long-term order pipeline.

However, the industry generally believes that the key to converting these potential clients into large-scale mass production contracts depends on whether the yield rate of the 2nm process can stabilize above 70%. Currently, this yield rate is estimated to be around 60%, still lagging behind the threshold for mass production. An industry insider stated, "Normalization of utilization is a leading indicator for the foundry business to cross the break-even point, but to confirm a substantive improvement in the business's fundamental health, we must wait for the stabilization of 2nm yields and the actual transition of major clients to mass production."

Rising Share of Advanced Processes, Parallel Capacity Expansion

Business structure adjustments are also accelerating in parallel. The share of revenue from advanced processes is expected to exceed 50% this year, while the share of revenue related to AI and HPC is poised to jump significantly from the low end of approximately 10% last year to over 30%. In the second half of the year, mass production of mobile products based on the second-generation 2nm (SF2P) process will commence.

On the capacity front, preparations for the launch of Fab 1 in Taylor, Texas, USA, are proceeding as scheduled this year, while groundbreaking for Fab 2 in Taylor is planned within the year, targeting mass production by 2030. As inquiries regarding the 1.4nm process increase, plans for additional fabs have also been placed on the research agenda.

Brokerage analysts pointed out that if the effect of rising wafer prices synergizes with the normalization of utilization, the pace of profit improvement for Samsung Foundry could be faster than previously expected by the market. However, there are also cautious voices in the market: while 100% capacity utilization is important in itself, whether the 2nm yield can break through 70% is the true watershed moment for flagship clients like Qualcomm and AMD to initiate large-scale outsourcing. In other words, the normalization of utilization marks the starting point for the end of losses, but the deep restructuring of the business still relies on continuous breakthroughs in advanced process technology and the substantive conversion of major clients to mass production.