TSMC commits $100B of capital over three years to add chip-fabrication capacity, making expansion rather than near-term capital restraint its operating priority.
The plan establishes a multiyear spending framework that later supports TSMC's higher 2022 capital-expenditure guidance.
The company’s capital allocation becomes a competitive constraint on other foundries: matching TSMC requires comparable long-duration investment, not simply incremental capacity additions.
Third-order effects
The subsequent U.S. commitments indicate that leading-edge foundry capacity is increasingly shaped by where fabs and supporting packaging operations are located, alongside the size of the capital program.
If this investment pattern persists, semiconductor manufacturing will become more concentrated among operators able to finance repeated, multiyear fab expansions, with margins pressured as overseas fabs ramp.
The trend: Semiconductor manufacturing is moving into a capital-intensive, geographically distributed capacity race led by foundries able to sustain multiyear investment programs.
Intel plans to spend $20B over the next 4 years to dig itself out of a 5yr hole. While TSMC is spending $100B over the next 3 to leapfrog even further. They are just in different leagues now. The kind of lead Intel used to have but squandered. https://asia.nikkei.com/...
Intel in 2007 was concerned about Innovator's Dilemma disruption to its CPU business, and so it started the Atom line to make x86 cheaper. But it somehow overlooked its core competency of the time: chip manufacturing. Now TSMC has fully disrupted them. https://asia.nikkei.com/...
So the U.S. as a country is investing $200B per year on infrastructure as the most ambitious “one in a generation” effort. TSM, a chip manufacturer in Taiwan, is investing $33B/year or 1/6 of the U.S. entire infrastructure plan? https://twitter.com/...
“It's economically unrealistic for all the countries to build additional chip production capacity,” says chairman of Taiwan Semiconductor Manufacturing Co., the world's biggest contract chipmaker. https://asia.nikkei.com/...
Head of TSMC and also Chairman of Taiwan Semiconductor Industry Association defends Taiwan for the criticism that its leading position in global chip production was to blame for the serious chip shortage @NikkeiAsia https://asia.nikkei.com/...
“Liu also confirmed for the first time that there is a serious “double-booking” effect in the industry, in which clients place orders for more chips than they actually intend to use.” https://asia.nikkei.com/...
$TSM guided for 2021 capex of $25B-$28B, and might spend ~$4.5B this year on R&D. This comment suggests their capex/R&D spend will average ~$33.5B-$35B/year in 2022 and 2023, with maybe ~$27.5B-$29B/year spent on capex (assuming R&D grows ~20%/year). https://www.bloombergquint.co…
Yes, been saying this for some time.. The head of world's biggest contract chipmaker has said the rush by major economies to onshore semiconductor production is “unrealistic” and that expanding capacity would not help alleviate the global chip shortage. https://asia.nikkei.com/..…
Wuhan Hongxin Semiconductor Manufacturing is one of six multibillion-dollar chip projects to fail in China in the last two years. (Turns out it ain't that easy.) https://www.npr.org/...