TSMC Advances 1nm Chips with 200B Transistor Goal

According to reports, TSMC has begun mass production of its 2nm process, while development and planning for 1nm technology are progressing steadily. The company is reportedly preparing to build a 1nm production hub at its Tainan Shalun Science Park, which spans 531 hectares. The second phase of the environmental review is expected to start in April this year, with final approval targeted for Q3 2027. Once completed, TSMC plans to start initial construction on approximately 200 hectares of the site. The Shalun Park will eventually host six fabs: P1–P3 for the 1.4nm (A14) process, and P4–P6 focused on the 1nm process, with potential expansion toward 0.7nm in the future.

TSMC's roadmap shows that its 2nm process (N2) entered mass production by the end of last year and is expected to see widespread adoption by major tech players like Apple and AMD this year. Meanwhile, the A16 process for NVIDIA's Feynman GPUs (around 1.6nm) is set for trial production by year-end and full-scale production by 2027. Following this, the A14 process (1.4nm) will use second-generation GAA transistors and backside power delivery, targeting a 2028 launch.

Details on the 1nm process (A10) remain scarce, but rumors suggest an upgrade from GAA to CFET transistors and possible integration of 2D materials. The 1nm technology is planned for 2030, with an ambitious goal of integrating 200 billion transistors per chip and achieving a 1 trillion transistor scale using 3D packaging. This aligns with Intel's earlier trillion-transistor vision, though which company will reach it first remains uncertain.