Samsung Foundry Poised for Inflection Point, Full Capacity and Profit Turnaround Seen in H2

According to industry sources cited by media outlets, Samsung Foundry is ramping up efforts to hit its target of 100% capacity utilization in the second half of this year. Its current overall utilization rate stands at roughly 70%–80%. Given the existing order backlog and progress in contract negotiations, internal and external stakeholders are optimistic about achieving the goal, which is expected to reverse its prolonged run of operating losses.

Dual Upgrades in Business and Capacity, Advanced Nodes Take Growing Share

On the business front, the company's high-end transition is yielding tangible results. Revenue from advanced processes is projected to exceed 50% of total revenue this year, while AI and high-performance computing (HPC) revenue is set to climb from over 10% last year to more than 30%, reinforcing its high-value-added business mix.

On the capacity front, Samsung is accelerating overseas expansion. Construction of its first fab in Taylor, Texas, remains on track to commence production this year, while work on the second Taylor fab will kick off within the year, with mass production targeted for 2030. Meanwhile, in response to rising market demand for the 1.4nm process, the company is evaluating plans for additional fab facilities.

On the process node front, mobile products built on the second-generation 2nm (SF2P) process will enter mass production in the second half of the year, further bolstering its competitiveness in the advanced-node market.

Demand Surge, AI and HBM as Core Recovery Drivers

The sharp uptick in Samsung Foundry's capacity utilization is driven mainly by strong demand for high-end chips from major U.S. tech firms. Industry insiders note that robust orders from U.S. tech giants for HBM base die and advanced-node chips have directly accelerated production ramp-up. With the 4nm process deployed for sixth-generation HBM (HBM4), the windfall from the global memory supercycle is spilling over into the foundry segment, forming a key pillar underpinning higher capacity utilization.

Optimized Order Structure, Diversified Clients Ease Concentration Risk

Samsung Foundry has long suffered from a concentrated customer base and heavy reliance on a limited number of key clients. Recently, however, leading tech companies including Qualcomm, AMD, and Google have entered into foundry partnership negotiations with Samsung, steadily expanding its customer portfolio.

Demand from cloud service providers (CSPs), AI, and HPC sectors continues to expand, while orders for the 2nm advanced process are growing steadily. Discussions on project cooperation with industry leaders such as Broadcom are also progressing, bringing incremental volume to fill capacity.

Furthermore, Tesla plans to adopt Samsung's 2nm process for its next-generation proprietary chip, which could secure stable medium- to long-term order pipelines for Samsung. However, it is said that the conversion of new orders still hinges on a critical hurdle. Volume production partnerships with new customers like Tesla depend on yield improvements for Samsung's 2nm process. The current yield stands at around 60%, which needs to rise steadily to approximately 70% to convert tentative orders into binding mass-production contracts.