Samsung Electronics' foundry division is accelerating its efforts to improve 2nm process yield, using Tesla orders as a strategic springboard to break into major technology customers.

On September 28, according to industry sources cited by South Korean media ChosunBiz, the yield of Samsung Electronics Co., Ltd. (SSNLF) 2nm process has improved further from the previously widely recognized level of around 55%, and is now approaching 60%.

With yield stabilizing, Samsung is concentrating engineering resources on optimizing power consumption and thermal performance, aiming to build a foundation of trust in technical negotiations with potential major customers.

Meanwhile, the second-generation 2nm process SF2P, designed to compensate for the performance shortcomings of the existing process, is also being accelerated, with a target of achieving mass production within this year.

Analysts view this move as having a clear strategic direction: using the opportunity of supplying 2nm chips to Tesla next year to fully open a breakthrough in orders from major technology companies.

Yield Nears 60%, Power Efficiency Becomes Key Variable

According to industry sources cited in the report, the improvement trajectory of Samsung's SF2 process yield is clear: approximately 30% in mid-last year, rising to the 50% range in January this year, reaching around 55% by August, and now approaching 60%.

SF2 is Samsung's first 2nm mass-production process adopting a gate-all-around (GAA) architecture. Its first mass-produced product is the self-developed mobile application processor Exynos 2600, which has been equipped in some models of the Galaxy S26 series launched earlier this year and is manufactured at the S3 line in Hwaseong, Gyeonggi Province.

Samsung adopted a copper heat dissipation module in the chip packaging, reducing thermal resistance by 30%.

However, after the Exynos 2600 was launched, some overseas tests showed that it consumed more power than Qualcomm's latest Snapdragon product when performing the same tasks, raising external doubts about the process's energy efficiency. Some views suggest this reflects inherent limitations of the SF2 process itself.

In response to these issues, Samsung is preparing the second-generation 2nm process SF2P, targeting mass production within this year. It is reported that SF2P delivers a 12% performance improvement, 25% lower power consumption, and 8% smaller chip area compared to SF2.

Samsung's current engineering focus is on ensuring consistency in product performance during mass production. Specifically, it aims to keep the deviation in heat dissipation and power consumption within an acceptable range after the chip is installed in customer products. This capability will serve as Samsung's core technical endorsement in negotiations with major fabless chip design companies.

TSMC Dominates as Samsung and Intel Vie for Second Place

The global advanced process foundry market currently presents a clear "one superpower, two strong contenders" pattern.

Taiwan Semiconductor Manufacturing (TSM), backed by core customers such as Apple, AMD, and Nvidia, recorded a 72.5% market share in the second quarter of this year, while Samsung Foundry's share was only 5.9%, with the gap between the two further widening.

TSMC's 2nm wafer price has stabilized at approximately $30,000, 50% higher than 3nm (around $20,000), and it maintains a "no discount" principle even for major customers. Industry insiders generally believe that TSMC's 2nm capacity has been booked through 2028.

On Intel's side, its 2nm-class process 18A has entered mass production at Fab 52 in Arizona, USA. The Panther Lake processors produced there have helped Intel capture 56% of the AI PC market, achieving certain results on its own product line.

However, 18A has yet to make significant progress in recruiting external customers. It is reported that Intel internally tends to pin its hopes for external customer breakthroughs on the next-generation 14A process, which is planned for mass production after 2027.

Against this backdrop, the industry widely expects the market landscape to feature TSMC in the lead while Samsung and Intel compete for the second tier. Regarding the break-even point for Samsung's foundry business, the mainstream market view holds that it must be observed in conjunction with the capacity utilization rate of the Taylor wafer fab in Texas.

A semiconductor industry insider stated:

The yield approaching 60%, along with the initiation of power optimization efforts, are meaningful signals of recovery.

However, the person also noted:

The market share gap with TSMC is too large, and achieving a true shift in the market landscape will still require more time.