TSMC accelerates move towards 1.6nm chips to power next-generation AI and robotics

Taiwan Semiconductor Manufacturing Company revealed plans to transition from its 2nm-class to 1.6nm-class production, aiming to boost efficiency for AI, robotics, and autonomous vehicles amid shifting timelines and technological advances.

Taiwan Semiconductor Manufacturing Company has set out its next steps in advanced chipmaking, using a technology summit to outline how it plans to move from its 2-nanometre-class generation to 1.6-nanometre-class production. The company presented the roadmap as part of a broader pitch that the latest wave of artificial intelligence will only reach its full potential if it is paired with far more efficient silicon, particularly for robotics, autonomous vehicles and advanced health applications.

According to Tom’s Hardware, TSMC first disclosed its 1.6nm-class A16 process in April 2024. The design combines backside power delivery, which moves power routing to the rear of the chip, with gate-all-around nanosheet transistors. Those changes are intended to raise clock speed while lowering power use, and to improve transistor density compared with the company’s N2P node.

TSMC’s 2nm-class N2 process is itself a key milestone, with volume production beginning in the fourth quarter of 2025, according to Tom’s Hardware. The same report said N2 is the company’s first node to use gate-all-around transistors, while enhanced versions such as N2P and A16 are meant to extend the efficiency gains for more demanding processors.

There is, however, some uncertainty over timing. Tom’s Hardware reported in one account that A16 remains on track for mass production in late 2026, while another report said the company’s 2026 roadmap has shifted A16 to 2027 and groups future nodes around different customer needs. Either way, the direction is clear: TSMC is preparing a staged transition from 2nm-class chips towards more power-efficient platforms for AI and high-performance computing.