Samsung races to increase 3nm AI chip yields for Nvidia

Geopolitical tensions in Taiwan may lead Nvidia to switch AI chip production from TSMC to Samsung. The company is currently facing yield issues with its cutting-edge 3nm semiconductor fabrication node GAA.

Nvidia CEO recently stated that “agility and their capability to respond to our needs is just incredible,” but “if necessary, of course, we can always bring up others.” Hence, Samsung has an urgent need to enhance its 3nm yield for AI processors.

The top executive mentioned that Nvidia could switch suppliers if necessary, due to concerns like rising geopolitical tensions in Taiwan. However, switching semiconductor production from TSMC might “result in lower-quality products.”

The South Korean tech giant is scrambling to improve its manufacturing yield, which refers to the ratio of good and error-free units to the total number of units produced. But the number is not improving as fast as Samsung hopes.

Samsung failed to secure Nvidia orders for advanced AI processors such as H Series or Blackwell. Last year, the company secured Nvidia’s orders for 8-nanometer-based chips like the Tegra for automobiles and the RTX-3000 GPU in 2020.

The company’s 3nm yield rate remained single digit until the first quarter of the year. This impacted the mass production of the Exynos 2500 chipset, which was rumored to be utilized in the Galaxy S25 series next year.