Samsung may have led the 3nm race with its Gate-All-Around (GAA) process, but the production yields from this process have yet to meet satisfactory levels, if a new leak coming out of Korea is to be believed.
A report from Naver reveals that Samsung initially set an ambitious target of 70% yield for both its first- and second-generation 3nm GAA processes. However, this goal now appears far out of reach.
Reportedly, the first iteration, dubbed “SF3E-3GAE,” is faring comparatively better, with yields hovering around 50-60%. While this is closer to the target than its younger sibling, it still falls short of the level needed to make the technology commercially attractive.
This shortfall seems to be a key reason behind the lack of new clients adopting Samsung’s 3nm GAA process. Qualcomm’s decision to manufacture the Snapdragon 8 Elite exclusively on TSMC’s 3nm “N3E” architecture is a case in point.
It keeps getting worse…
The outlook for Samsung’s second-generation 3nm process (dubbed SF3-3GAP), however, is even grimmer. The report suggests its yield rates are just 20%, less than a third of the initial target.
This lack of progress could further undermine confidence in Samsung’s 3nm offering, and could potentially drive even South Korean clients toward TSMC’s more established 3nm technology.
Samsung’s 3nm GAA technology might be turning into a bit of a headache. This could explain why the company seems to be shifting resources and talent towards its upcoming 2nm node.
Rumors suggest that an unnamed Exynos chipset, codenamed “Ulysses,” is being developed using the “SF2P” technology. This chip is expected to debut in a future Galaxy S model, possibly the S27 in 2027.
It remains to be seen, however, if Samsung can overcome these hurdles and establish its 3nm technology as a viable alternative. If not, the future of the company’s next flagship Exynos 2500 SoC could also be in doubt.
(Source)
Comments