TSMC’s plans for the next few years remain mostly unchanged – at the end of 2025 the company is preparing to start mass production of chips using 2nm technology, and at the end of 2026 A16 technology (1.6nm class) will be ready. The manufacturer will announce the Open Innovation Platform held in Amsterdam in 2024 He said the following at the (OIP) conference.

At its event in the Netherlands, TSMC presented a roadmap that clearly shows its plans for the implementation of new solutions. By the end of 2025, N2 technology is planned to be deployed – this will be the first in the 2nm class; Improved N2P and N2X variants will follow within a year, and then it will be the A16’s turn. The last three are expected by the end of 2026, the company said. Other details were not available because the contract manufacturer was unable to announce products in advance for its largest customers.

Technically, N2, N2P, N2X and A16 have a lot in common; they are all based on Gate-All-Around (GAA) transistors. The N2 series uses high-performance metal-insulator-metal (Super High Performance Metal-Insulator-Metal, SHPMIM) capacitors, which help reduce the resistance of the transistor and therefore increase performance; and in the A16 — backside power supply (Backside Power Distribution Network, BSPDN) providing even higher performance. However, the last solution also has a disadvantage; additional heat generation to be taken into account.

Therefore, chips based on A16 technology with BSPDN in its current application are the most suitable for artificial intelligence accelerators to be used in data centers. N2P offers performance gains over base N2 without the added complexity, making it the technology best suited for mobile processors and entry-level computer chips. Finally, N2X means increased performance and higher voltage; It is an ideal solution for high-performance CPUs.