You are currently viewing SemiWiki as a guest which gives you limited access to the site. To view blog comments and experience other SemiWiki features you must be a registered member. Registration is fast, simple, and absolutely free so please, join our community today!
CNA file photo
Taipei, March 31 (CNA) Taiwan Semiconductor Manufacturing Co. (TSMC) received government approval on Tuesday to deploy its advanced 3-nanometer (3nm) process at its second fab currently under construction in Japan, according to a news release.
The Ministry of Economic Affairs (MOEA) green-lit the plan for the facility in Kumamoto, which is scheduled to start installing equipment and come online in 2028 with a monthly production capacity of 15,000 12-inch wafers, the ministry said in the release.
The MOEA's Department of Investment Review in June 2024 authorized a US$5.26 billion investment for the facility, slated to manufacture 6- to 12nm chips, significantly less advanced than 3nm process.
At a meeting with Japanese Prime Minister Sanae Takaichi in Tokyo this February, TSMC CEO C.C. Wei (魏哲家) had stated that the shift in plan for the second fab was a response to soaring global demand for artificial intelligence.
TSMC's first Kumamoto fab, operated by Japan Advanced Semiconductor Manufacturing, Inc. (JASM), entered mass production in late 2024, primarily focusing on 12nm-28nm logic chips for the automotive and industrial sectors.
Taiwan Semiconductor Manufacturing Co. (TSMC) received government approval on Tuesday to deploy its advanced 3-nanometer (3nm) process at its second fab currently under construction in Japan, according to a news release.
Great news for N3 customers! Japan fabs go up quickly. Japan neighbors calls them Night Castles because work is done around the clock.
The real reason is that 6-12nm wafers are commodities selling for cheap from Samsung and GF. 3nm has a much bigger margin so this is a strong financial move.
Yes. Samsung is buying customers at 7-4nm which is why TSMC is moving capacity down to 3nm because Samsung S3F 100% failed. This is a good thing for the industry. It is unfortunate for Intel Foundry that they are not in the position to buy customers like Samsung is.
Samsung is still at a serious ecosystem disadvantage but if you can't get into TSMC Samsung is a viable alternative for sure. Or if lower wafer prices are worth the risk of doing business with Samsung.
TSMC N2 is still dominating design wins and I do not see that changing especially with the massive TSMC CAPEX for 2026. TSMC builds capacity to orders, right?
focustaiwan.tw