On Wednesday, Taiwan Semiconductor Manufacturing Co (TSMC) introduced cutting-edge chip technology designed to enhance speed and efficiency, particularly for artificial intelligence applications. The company announced the 2028 debut of a new production process for the A14 chip, which it claims will offer chips 15% faster than the N2 chips slated to begin production later this year, all at equivalent power levels. An alternative would be for the A14 chips to work at the N2s' clock rate at a power level that is 30% less.

As the world’s leading contract chipmaker, TSMC whose clients include NVidia and Advanced Micro Devices (AMD) also unveiled a new packaging innovation called "System on Wafer-X." It is a working platform that will integrate, at the very least, 16 large computing chips with memory chips, high-speed optical interconnections, and advanced power delivery technology capable of supporting thousands of watts. These components will be packaged together to form a package approximately the size of a dinner plate.

This platform will integrate at least 16 of its large computing chips along with memory chips, high-speed optical interconnections, and advanced power delivery technology that is capable of supporting thousands of watts. These will be packaged to sizes that are about the size of a dinner plate. To support this initiative, TSMC plans to construct two new facilities near its existing Arizona-based chip plants.

Meanwhile, Intel is preparing to unveil its own advancements in manufacturing technology next week. Last year, Intel asserted that it would surpass TSMC in the race to produce the world’s fastest chips.

As demand surges for large-scale AI chips assembled into unified packages, the competition between TSMC and Intel is passing from the onset of chip making into integration, which is inherently technical and requires working deep with the customers.