TSMC could reach its 2027 2nm capacity targets a year early, according to a report from Taiwan that points to strong orders from Nvidia, AMD and other major technology companies, including Apple and Qualcomm.
2nm is the most advanced manufacturing process TSMC currently has in production. For AI chips and data center processors, process maturity matters most: a node that has been in production long enough lets customers eliminate performance and power problems under heavy load. The company's roadmap calls for the N2P family in 2026 and the N2X family in 2027, the latter aimed at high-performance devices and data center computing.
According to the report cited by Taiwanese paper UDN, 2nm capacity could reach 120,000 wafers per month by the end of 2026, about 20% more than the previous estimate of 100,000. UDN had previously indicated this target, and a later report from the Commercial Times described a further acceleration. The end-2027 target is 140,000 wafers per month; if the estimate holds, TSMC would hit that level a year early.
Bank of America's estimates
A Bank of America analysis supports the trend. Ahead of TSMC's earnings, the bank expects 2nm's contribution to revenue to rise from 3% to 11% in the third quarter. The same estimate puts utilization of the 3nm and 2nm nodes above 90%, with a tendency to grow. By the third quarter of 2028, Bank of America expects wafer supply to exceed 500,000 per month, with utilization reaching 105%.
The bank also suggests TSMC's leading-edge lines could be booked until 2028, a shift in a historically cyclical industry. This remains an estimate not confirmed by other independent sources.
No comments yet. Be the first!