News
NVIDIA Rumored as TSMC A16's First Customer
According to reports from Wccftech, NVIDIA is rumored to be the first customer for TSMC's cutting-edge A16 (1.6nm) process. If confirmed, this would mark a major strategic shift for NVIDIA, which has historically relied on TSMC's slightly older nodes rather than its absolute latest process technology.
Traditionally, companies like Apple, MediaTek, and Qualcomm have been early adopters of TSMC's most advanced nodes, while NVIDIA has focused primarily on architectural innovations to boost performance. This move could see NVIDIA returning to TSMC's leading-edge process since its use of the 0.11-micron node, promising significant performance gains and further progress along Moore's Law.
The A16 node introduces key technologies including GAAFET architecture and SPR back-side power delivery, with the latter regarded as the most impactful breakthrough. Market expectations suggest NVIDIA could adopt A16 around late 2027 or early 2028, aligning with TSMC's plan to begin mass production by the end of 2026.
According to TSMC's roadmap, the A16 process may first power Rubin Ultra or Feynman GPUs, with Feynman being the more likely candidate due to its integration of multiple technological advances.
For TSMC, this development is positive news. Early adoption of its most advanced nodes by major players helps optimize process yields, accelerate production ramp-up, and drive higher revenue and operational efficiency.
