There's a lot of commotion within Samsung's foundry division. The Korean giant had high hopes for this business, imagining it would challenge the dominance of TSMC, which manufactures chips for Qualcomm, MediaTek, and, of course, Apple. But according to TrendForce, Samsung's foundry represented only 7.7% of the market in the first quarter, compared to 67.6% for its Taiwanese rival. Profitability is reportedly in the doldrums, with losses piling up.
Google is switching sides for its in-house chip
And to make matters worse, Google has reportedly decided to move to TSMC for the future Tensor chip for the Pixel 10. According to a report from The Bell, Samsung's yield for 3nm chips is reportedly insufficient, and there is also a lack of expertise in designing IP blocks (pre-designed and reusable circuit components, such as a processor or memory controller) suited to the needs of Google, which has ambitions for foldable smartphones... in direct competition with Samsung.
The move to TSMC is not just technical, it is also strategic. According to industry sources, TSMC would have exclusivity on Google chips up to the Pixel 14! Samsung is therefore losing an important customer, and a highly symbolic one since the Tensor chips are (or were?) based on the company's Exynos chips.
The loss of the Google customer would therefore explain the current introspection within Samsung, which is seeking to understand what is blocking its foundry. An audit has been launched to identify the weak points and learn lessons from the situation to get back on track. Several measures are already being considered, such as the splitting of the foundry division, a partial reorganization, or diversification into other industries such as automotive or robotics.
Above all, Samsung must demonstrate that its own Exynos chips can compete with the best, in other words, Qualcomm's Snapdragons. The Exynos 2600, intended for the Galaxy S26 scheduled for next January, would be engraved in 2nm.
Source: The Bell
0 Comments