Taiwan Semiconductor Manufacturing Co. (TSMC) has unveiled its new-generation chip technology, the A14 process, which is scheduled for deployment in 2028. The new process is aimed at fulfilling the growing need for AI-focused computing, providing dramatic enhancements in performance and efficiency compared to existing models.
TSMC says that the A14 process will offer chips that are as much as 15percent faster or 30 percent more power-efficient than its existing industry-leading N2 chips, which will go into production later this year. Furthermore, A14 will offer an over 20 percent boost in logic density, allowing even higher computational capability within smaller chip sizes.
Along with the A14, TSMC announced its latest "System on Wafer-X" technology. The technology can integrate as many as 16 large computing chips with memory and optical links in a single wafer, significantly improving system performance and power efficiency. To satisfy increased demand, most notably from leading AI customers like Nvidia and AMD, TSMC will build two new production facilities in Arizona.
“TSMC’s cutting-edge logic technologies like A14 are part of a comprehensive suite of solutions that connect the physical and digital worlds to unleash our customers’ innovation for advancing the AI future,” said TSMC, CEO, C.C. Wei in a prepared statement.
Based in Hsinchu, Taiwan, TSMC is the largest chipmaking contract company globally, and its semiconductors power some of the fastest chips for such devices as smartphones and data centers.