According to reports on October 18， TSMC’s 2nm process node， N2， is nearing mass production， signaling a future move into the angstrom era. The A16 process is slated for release next year， followed by the A14 node.
The A14， previously referred to as the 1.4nm process， is expected to represent a more significant technological leap than the transitional A16. It will incorporate second-generation GAAFET nanosheet transistors and a novel NanoFlex Pro standard cell architecture， promising substantial advancements.
When compared with the N2 process， the A14 is projected to deliver up to a 15% speed improvement at the same power consumption. Alternatively， it could achieve a power reduction of up to 30% at equivalent speeds， along with an increase in logic density exceeding 20%.
It is worth noting that TSMC’s comparative approach， benchmarking A14 against the N2 process rather than the immediately preceding A16， serves to highlight the magnitude of its advancements. This strategic comparison emphasizes the generational leap offered by the A14.
Regarding its production timeline， TSMC has consistently indicated a 2028 launch for the A14. More concrete evidence has recently emerged as TSMC has reportedly applied for permits to commence construction of its A14 factory in Central Taiwan Science Park. Construction is anticipated to begin in Taichung by the end of this year， and the company is actively recruiting personnel.
For the 1.4nm node， TSMC is highly likely to utilize High NA EUV lithography machines. These next-generation machines， priced at $400 million each， are significantly more expensive than the current $200 million EUV machines. Consequently， the investment required for building 1.4nm chip fabrication plants will be substantial， with initial estimates reaching $49 billion， approximately 350 billion RMB.
However， TSMC appears confident in its ability to recoup these investments. Rumors suggest that the pricing for the 1.4nm node will see another increase， reaching $45，000 per wafer. This translates to a manufacturing cost of roughly 320，000 RMB per wafer， a considerable jump from the $30，000 price point for the 2nm process.
If this trend of escalating prices continues， it is plausible that only companies with exceptionally high-profit margins， such as NVIDIA for its AI chips， will be able to afford the 1.4nm process in the future. The cost may prove prohibitive for mobile processor manufacturers， and it’s conceivable that PC gamers could see fewer CPUs and GPUs built on this cutting-edge technology.
