Array ( [content] => [params] => Array ( [0] => /forum/index.php?threads/u-s-ramps-up-chip-production-efforts-with-taxpayers-footing-part-of-the-bill-is-it-worth-it.20126/ ) [addOns] => Array ( [DL6/MLTP] => 13 [Hampel/TimeZoneDebug] => 1000070 [SV/ChangePostDate] => 2010200 [SemiWiki/Newsletter] => 1000010 [SemiWiki/WPMenu] => 1000010 [SemiWiki/XPressExtend] => 1000010 [ThemeHouse/XLink] => 1000970 [ThemeHouse/XPress] => 1010570 [XF] => 2021370 [XFI] => 1050270 ) [wordpress] => /var/www/html )
It remains to be seen if TSMC's 5-4nm factory, which will only operate in a couple of years, will have much clientele. Especially when the chip prices will be much higher than for chips made in Taiwan. Everything is more expensive in the US including salaries.
Previous experience with TSMC's WaferTech subsidiary and Samsung's Austin fab has not been exactly positive for either company.
Intel's EUV expansion was highly necessary and should have happened sooner. Because Intel is an IDM they can better absorb the additional foundry costs than a pure play foundry like TSMC.
Conclusion: making chip in the US is very expensive. more money is needed because National security.
Intel's Pat is trying hard to sell foundry services to Elon and Altman![]()
This is not the impression I got from speeches by the TSMC CEO last year. I hope you are right.Chip prices will not be higher. Manufacturing costs might me higher but wafer prices will be the same as Taiwan fabs.
China, Taiwan, Japan, South Korea, and Singapore also subsidize factories. And they pay lower salaries to workers while making them work longer hours.You also have to factor in the CHIPs Act monies for AZ.
They definitively won't be making any revenue until they start operating them and delivering product.TSMC is NOT losing money on the AZ fabs.
I am just assuming it will take them time to ramp up production.TSMC AZ fab begins production this year. Your view of TSMC seems jaded, why is that?
I am just assuming it will take them time to ramp up production.
Is it worth it? YES
It is not just for national security. Semiconductor Fabs are amazing economy boosts, high paying job boosts and the impact lasts a long time. It is the best thing you can do for the people in the US. Having Global, Intel, TSMC, Samsung, Micron all getting money will have a measurable impact on the whole nation. It is the best use of money by the government in 50 years. (now the rest of the 200B slush fund might get wasted we wont know where it went). But the Fabs subsidies will be game changing. It will pay for itself.
AND ALL other countries give subsidies to these companies because is SO smart to do so. The difference is where the fabs are built.
They should hire me to do PR LOL
It remains to be seen if TSMC's 5-4nm factory, which will only operate in a couple of years, will have much clientele. Especially when the chip prices will be much higher than for chips made in Taiwan. Everything is more expensive in the US including salaries.
Previous experience with TSMC's WaferTech subsidiary and Samsung's Austin fab has not been exactly positive for either company.
Intel's EUV expansion was highly necessary and should have happened sooner. Because Intel is an IDM they can better absorb the additional foundry costs than a pure play foundry like TSMC.
This is not the impression I got from speeches by the TSMC CEO last year. I hope you are right.
China, Taiwan, Japan, South Korea, and Singapore also subsidize factories. And they pay lower salaries to workers while making them work longer hours.
In China, for example, you get the land for free. You pay 0% corporate tax. And if you do a joint venture with the Chinese government they might pay the cost of half or more of your factory. You will get money from both the local and central government.
In other countries in the Far East the subsidies vary but are also quite extensive.
For example the Japanese government subsidized 40% of the cost of the latest TSMC fab there.
They definitively won't be making any revenue until they start operating them and delivering product.
I am just assuming it will take them time to ramp up production.
Luckily Semi fabs are much better for air quality today than 90% of industries. I have heard stories about Intel fab 1,3,4 that are wild .... those days are over..If you look at the Federal budget/deficit, in comparison I think the CHIPs Act is money well spent. Hopefully there will be more to come.
By the way, I was working in Silicon Valley when we had fabs here. They were dirty and the EPA tossed them overseas so it wasn't all about cost. I was also in Hsinchu in the 1990s. You could taste the air back then.
OK, but what's the limit on what the subsidy should be then ?Is it worth it? YES
It is not just for national security. Semiconductor Fabs are amazing economy boosts, high paying job boosts and the impact lasts a long time. It is the best thing you can do for the people in the US. Having Global, Intel, TSMC, Samsung, Micron all getting money will have a measurable impact on the whole nation. It is the best use of money by the government in 50 years. (now the rest of the 200B slush fund might get wasted we wont know where it went). But the Fabs subsidies will be game changing. It will pay for itself.
AND ALL other countries give subsidies to these companies because is SO smart to do so. The difference is where the fabs are built.
They should hire me to do PR LOL
Interesting question. What is TSMC worth to Taiwan, given the probable reality of the "silicon shield"? Trillions of dollars? What are advanced chips worth to China, in light of the US technology embargoes? Pretty much any amount I'd guess.OK, but what's the limit on what the subsidy should be then ?
Let's keep things simple and assume govt X is subsidising a fixed percentage of company Y's leading edge (let's call it around 3nm) logic wafer fab and no ongoing handouts. What's the rough % estimate where the subsidies outweigh the benefits ? 20% ? 50% ? 80% ?
In the case of the US, the US government can and has held equity stakes in companies. Politically it's not popular, and seldom done. It was discussed when the CHIPS Act was being written, but obviously Congress wrote in a very clumsy profit-sharing plan instead (which IMO is a profoundly dumb idea). An equity stake implies the government is going to invest for profits, which is frowned on, or a social agenda, which is not in keeping with the US free enterprise agenda. Also, in the US there is resistance to government control over private corporations, so board seats or special reporting structures are usually out of the question.I'd like to understand a bit better where the crossover might be.
I'm also curious why governments don't want an equity stake when they're putting up this sort of cash.
The US has its share of fiascos too. The planned Foxconn electric car plant in Lordstown, Ohio comes to mind as an example in the fiasco category.I'm also a little sceptical when I recall that sometime during the early/mid 1990s the UK government was so persuaded by this argument that it ended up with Wales in a bidding war with the North East of England (UK bidding against the UK) for a memory wafer fab that either never got built or shut after a year or two. Memory being memory and the fab being planned/built at the peak of the market just before the next bust. Still, I'm sure LG were grateful.
I have worked on these agreement details in the past. they have standard requirements.OK, but what's the limit on what the subsidy should be then ?
Let's keep things simple and assume govt X is subsidising a fixed percentage of company Y's leading edge (let's call it around 3nm) logic wafer fab and no ongoing handouts. What's the rough % estimate where the subsidies outweigh the benefits ? 20% ? 50% ? 80% ?
I'd like to understand a bit better where the crossover might be.
I'm also curious why governments don't want an equity stake when they're putting up this sort of cash.
I'm also a little sceptical when I recall that sometime during the early/mid 1990s the UK government was so persuaded by this argument that it ended up with Wales in a bidding war with the North East of England (UK bidding against the UK) for a memory wafer fab that either never got built or shut after a year or two. Memory being memory and the fab being planned/built at the peak of the market just before the next bust. Still, I'm sure LG were grateful.
A simplified contract says: we pay 30% of all capex. show us receipts, get a check . no prepayments. you must have payroll of 3000 people, 80% must be local hires. 10M fine per month if you fail this. We pay for highways and traffic. you pay for all driveways and connecting road to highway. Failure to spend 10B and hire 3000 people in the next 5 years is default, Fine of 125% of subsidy. In Other countries, government takes property on default.... everything is renegotiated based on these penalties.
agreed. But actually a lot of this issue had nothing to do with incentive subsidies. Building infrastructure and timing is always a hard part of any development. and every public work project is a study in inefficiency and cronyism.Here’s how not to do it - stupid politics just to claim a photo-op. By the time the clowns who put this deal together were out of office, the state and local governments had spent $1B to create a few jobs inside essentially empty buildings doing nothing. The successors were left to pick up the pieces based on a loophole-filled contract.
![]()
Wisconsin Foxconn Deal Cost Taxpayers Millions—And It Will Continue To Cost More Millions
In exchange for a $3 billion dollar subsidy, the company Foxconn promised to build a $10 billion factory that would create 13,000 jobs in Wisconsin. That promise still hasn’t come to pass, and there’s no evidence that it ever will.www.strongtowns.org
I had made my way to South Wales by then and had an interview with LG!OK, but what's the limit on what the subsidy should be then ?
Let's keep things simple and assume govt X is subsidising a fixed percentage of company Y's leading edge (let's call it around 3nm) logic wafer fab and no ongoing handouts. What's the rough % estimate where the subsidies outweigh the benefits ? 20% ? 50% ? 80% ?
I'd like to understand a bit better where the crossover might be.
I'm also curious why governments don't want an equity stake when they're putting up this sort of cash.
I'm also a little sceptical when I recall that sometime during the early/mid 1990s the UK government was so persuaded by this argument that it ended up with Wales in a bidding war with the North East of England (UK bidding against the UK) for a memory wafer fab that either never got built or shut after a year or two. Memory being memory and the fab being planned/built at the peak of the market just before the next bust. Still, I'm sure LG were grateful.