Rising Costs in Semiconductor Manufacturing: The Shift to 2nm Chips
As the semiconductor industry advances towards smaller and more efficient chips, the cost savings of transitioning to new manufacturing processes are diminishing. Taiwanese chip giant, TSMC, is set to mass-produce 2nm chips next year, with a staggering cost of $30,000 per wafer, marking a significant increase from previous generations. This shift highlights the increasing financial demands of cutting-edge technology development.
Increasing Cost Per Wafer
TSMC’s production costs illustrate the steep climb in expenses with each new generation of chips. A 300mm wafer of 3nm chips currently costs around $20,000, in comparison to $16,000 for the older 5nm technology. With the 2nm wafers, costs are projected to soar to $30,000, and further enhancements to the lithography could see these prices rise to $45,000.
Industry Adapts to Rising Costs
Despite the rising costs, manufacturers are not shying away from 2nm technology. According to Taiwanese publication Commercial Times, developing a 2nm chip could set clients back approximately $725 million. Leading companies like AMD announced that their Venice family of EPYC server processors will adopt this technology. MediaTek plans to unveil its 2nm Dimensity 9600 processors in December.
Tech Giants Embrace 2nm Innovation
Next year, Apple and Qualcomm are anticipated to introduce flagship chips using the 2nm process, specifically the A20/M6 and Snapdragon 8 Elite Gen 3, respectively. Cloud service giants are also gearing up to launch their 2nm processors by 2027. Google is slated to debut the Trillium v8, AWS (Amazon) the Trainium 4, and Microsoft aims to release the Maia 300 in late 2026.
These moves are strategic, as cloud providers strive for independence from dominant players like AMD and Nvidia, while optimizing costs. Notably, in 2023, AWS is leading with the highest growth in deploying its proprietary processors.
