Overall, it looks great. We absolutely need to fund more semiconductor manufacturing. Losing that ability to China/Taiwan was just a terrible strategic mistake.
The US hasn’t lost this battle (over semiconductors) just yet. While current manufacturing is indeed dominated by East Asian countries, the vast majority of research still happens in US universities. This legislation allocates more funding for that research, fantastic.
Theres some political gamesmanship over some silly stock options of congresspeople… my response is, who gives a single shit. The country needs this bill.
main problem is that we are rewarding companies for being disloyal and offshoring in the first place, now they win again with free money after raking in profits for decades. We should have some funding for smaller companies and R&D but also massive tariffs to force companies to bring manufacturing back
There are basically only 3 close-to-cutting edge foundries left - Intel (American), Samsung (South Korean) and TSMC (Taiwanese), and TSMC is in the lead by quite a bit. We're getting extremely close to the end-game of moore's law, so the lifetime of a <7nm fab is likely to be long, and there are likely to only ever be a handful built because they are absurdly expensive (20+ Billion dollars), particularly if you can't keep them fully utilized. There are no 'smaller companies' with cutting edge fabs, and these aren't really American companies who outsourced and are no bringing them back - Intel has always had a large manufacturing presence in the US, and a large part of this is trying to get foreign companies to 'offshore' fabs to the US rather than continuing to concentrate in south korea / taiwan.
TSMC straight out competed the 1st movers in the chip space. They’re not dominant because American companies moved their own factories to Taiwan. It was homegrown and they really nailed it.
> main problem is that we are rewarding companies for being disloyal and offshoring in the first place, now they win again with free money after raking in profits for decades. We should have some funding for smaller companies and R&D but also massive tariffs to force companies to bring manufacturing back
I'm no corporate apologist but the current mess is the result of a government policy designed to encourage globalization. An American company that builds stuff here will always be at a competitive disadvantage versus a foreign company that builds stuff using dramatically cheaper labor.
So I don't think it's really correct to think of these subsidies as "rewards" - it's done strategically to keep American manufacturing alive.
I would suggest that the real alternative is not to fund the smaller companies, but to simply nationalize Intel.
>>While current manufacturing is indeed dominated by East Asian countries, the vast majority of research still happens in US universities. This legislation allocates more funding for that research…>>
I suspect there’s more to it than this, but to the extent there’s not, this doesn’t make much sense. If the monetization of our research ends up in China’s control, how does it help to fund more research. Apparently research is not the problem, manufacturing is.
Can someone else corroborate that is something the US lost because of labor costs? We are constantly told it's R&D and infrastructure costs. How much of that is labor costs?
It almost feels like labor costs is a good story to salve the ego when the US was simply beat on tech and tech investment.
90 plants in the US, of which 5 have been built or under construction within the last 5 years. That's the same number of new fabs as Taiwan, China and Japan.
It is true that Taiwan has the technical lead at the moment, but that's a very recent development, yet everyone's talking as though the US hasn't built a chip since the 90s. Most of the strategically most important chips for military and industrial applications aren't at the cutting edge anyway.
It's an important strategic sector, sure, but none of these subsidies will actually address the most advanced tech segment and most of it will disappear as political pork. I mean it's your money, I'm a Brit, but to me this all just seems like absurd posturing.
>We need more partners in this not just a handout.
This rings hollow when all of America's major partners are pumping massive subsidies to their domestic companies to win. Case and point SK is subsidizing their Fabs to the tune of $450 Bn.[1]
> Losing that ability to China/Taiwan was just a terrible strategic mistake.
More like losing that ability to Taiwan's TSMC and South Korea's Samsung. China's SMIC is very far behind and has insignificant market share. Additionally, it's rather insulting to non-communists when you mishmash together China and Taiwan.
I think you misread the thrust of this comment. Commenter was likely alluding not to R&D, but an amphibious invasion, as to how it would end up being China. Obviously I don't know how much of that infrastructure would persist to be carried off after that invasion, but it would reasonably be considered 'lost' to the US.
This is always framed in terms of USA vs. China/Taiwan but what about Europe?
AFAIK (I might know wrong) the consensus among macroeconomists
is that "social market economies" like European ones are better suited (compared to "liberal market economies" like the USA) to investing in large scale manufacturing (see e.g. Varieties of Capitalism [0]—though it is a bit controversial).
Though I do wonder if chip manufacturing might just be too large for the EU to handle.
Many people in Europe keeps saying that e.g. Ericsson could start a fab but I'd be curious to see if a bill like this would succeed in the EU.
The EU passed their own Chips Act earlier this year. It was a point of contention since Intel's CEO has kept threatening to expand in Europe instead of the US if the US CHIPS act isn't passed.
>> We absolutely need to fund more semiconductor manufacturing. Losing that ability to China/Taiwan was just a terrible strategic mistake.
The correct solution to that is tariffs. That would hurt us short term and should have been done long ago before it would have been a problem. This will add to the debt pile, so it will be with us for a long time.
The debt pile won't notice an extra $50 billion. To make a meaningful dent in it, spending on entitlements needs to get cut or taxes on the middle class need to dramatically increase. Neither of those things are going to happen.
Fairly cynical comments on this thread, many trying to uphold American values like "free market", "protectionism is bad", "what about the consumer", "not from my tax money".
I find those remarks pretty naive. Every major power block heavily subsidizes strategic industries, now and forever. It's in no way new or remarkable, nor does it violate some principle. Most of the above concepts are pure fiction.
Semiconductors are not a "free market" in any ordinary sense. It costs tens of billions to enter the market and you'll buy an ASML machine, speaking of dependencies.
It's pretty obvious that this is in response to geopolitical instability. Being geo locked and this not really being a free market at all justify the "unusual" decision.
As is stands, food, energy, and yes...also semiconductors are foundational to a modern digital society. And they're interconnected, without semiconductors you won't have food either.
If you want to know what happens when semiconductors dry up, watch modern Russia. Arguably it's hard to do because of all the propaganda, but pretty much every domestic industry is falling apart, and therefore society falls apart.
Bottom line, don't be naive. Sure it sucks to give rich companies even more money but if that's what it takes, so be it. Also, rich is just one side of the coin, the other is that they spear-headed the information revolution.
Here's my cynical take: This bill will subsidize a lot of jobs when we already have zero slack in our labor market. I think it's an excellent target for spending and would have a positive multiplier in normal times but in 2022 this seems likely to just goose inflation another inch.
Unless we actually start seeing cyclical job losses finally hit right when this money starts being spent. Which is entirely possible.
Fab guy here. This is excellent news and much needed, but vigilence is needed to make sure it doesn't end up being devoured by crony corporate agenda and it actually results in favorouble pro-US climate for semiconductor manufacturing.
I'd like to share personal experience with how we bankrupted American leadership in cutting edge nodes. Although, it is not lithography related, I was part of a few billion $ ROI program where we'd hot test the chip for binning, best I don't disclose too many details. Let's just say, it was critical so much so that I sat in unmarked buildings. I saw that get transfered under my personal watch to China. We had Chinese employees visit US for 6 months at a time and during this rotation, we'd teach them everything. Had to take a Chinese culture course. Process charts, metrics, drawings and schematics, whitepapers, how everything works, be part of troubleshooting process and then test them for their acuity. This was around 2012-2014. Usually, US semiconductor manufacturers do not transfer fab capabilities to China, only assembly/packaging. But, here, the was a clear violation of backend fab activities that were transferred to China and built out. I visited China for 3 months to get things up and running. This was a brand new process that no one in the world has. All custom equipment from a major Japanese equipment manufacturer. This process was so insane that it took 10 years of development internally to come to this point. Even today, in 2022, no one has replicated it.
This should not have happened IMO from a national security standpoint. But, these things continue to happen and US gov does not have enough insight into America's semiconductor industry when it comes to protecting IP. Far too many things do not require ITAR and are exported without oversight.
I am pretty much against over-regulation, but here there needs to be strict regulation for exporting any semiconductor technologies whether it is fab or assembly or what have you. The entire industry needs to be hamstrung with export control.
This is completely the wrong way to go about it (tax individuals and give the money to big corps). If you want a protectionist measure then implement a protectionist measure, like I don't know, tariffs.
Yet another bill that follows at least the last 40 years of congress doing exactly the wrong thing. (and to be clear this isn't a rant against the Democrats).
I can't help but conclude that this is too little too late. R&D isn't the problem, which is what this bill seems more focused on. The ability to produce affordable chips domestically at affordable prices is the actual problem. North America's first priority isn't access to bleeding edge tech.
> The bill also includes about $100 billion in authorizations over five years for programs such as expanding the National Science Foundation’s work and establishing regional technology hubs to support start-ups in areas of the country that haven’t traditionally drawn big funding for tech.
I can only imagine how bureaucratic the process for accessing those funds will be. More likely, Medium Tech and even Big Tech will be the ones who feed from that trough because they're the only ones willing or able to cut through the red tape.
I've received money from the NSF SBIR grant program and know another half a dozen others who have as well, and while there is certainly a process, it isn't too bad. All the people I know who have received them are at startups and small companies. That said, the NSF does more than just SBIR and I can't speak to any of that.
Readit News
Overall, it looks great. We absolutely need to fund more semiconductor manufacturing. Losing that ability to China/Taiwan was just a terrible strategic mistake.
The US hasn’t lost this battle (over semiconductors) just yet. While current manufacturing is indeed dominated by East Asian countries, the vast majority of research still happens in US universities. This legislation allocates more funding for that research, fantastic.
Theres some political gamesmanship over some silly stock options of congresspeople… my response is, who gives a single shit. The country needs this bill.
I'm no corporate apologist but the current mess is the result of a government policy designed to encourage globalization. An American company that builds stuff here will always be at a competitive disadvantage versus a foreign company that builds stuff using dramatically cheaper labor.
So I don't think it's really correct to think of these subsidies as "rewards" - it's done strategically to keep American manufacturing alive.
I would suggest that the real alternative is not to fund the smaller companies, but to simply nationalize Intel.
I suspect there’s more to it than this, but to the extent there’s not, this doesn’t make much sense. If the monetization of our research ends up in China’s control, how does it help to fund more research. Apparently research is not the problem, manufacturing is.
From the labor cost and natural resource consumption a operation of this size can easily fail. We need more partners in this not just a handout.
This action should have little to no input on any globalized leverage system.
For this to be successful we need more partners
It almost feels like labor costs is a good story to salve the ego when the US was simply beat on tech and tech investment.
https://en.wikipedia.org/wiki/List_of_semiconductor_fabricat...
90 plants in the US, of which 5 have been built or under construction within the last 5 years. That's the same number of new fabs as Taiwan, China and Japan.
It is true that Taiwan has the technical lead at the moment, but that's a very recent development, yet everyone's talking as though the US hasn't built a chip since the 90s. Most of the strategically most important chips for military and industrial applications aren't at the cutting edge anyway.
It's an important strategic sector, sure, but none of these subsidies will actually address the most advanced tech segment and most of it will disappear as political pork. I mean it's your money, I'm a Brit, but to me this all just seems like absurd posturing.
This rings hollow when all of America's major partners are pumping massive subsidies to their domestic companies to win. Case and point SK is subsidizing their Fabs to the tune of $450 Bn.[1]
[1]: https://www.bloomberg.com/news/articles/2021-05-13/korea-unv... [1 Non paywalled]: https://archive.ph/9Gs8q
More like losing that ability to Taiwan's TSMC and South Korea's Samsung. China's SMIC is very far behind and has insignificant market share. Additionally, it's rather insulting to non-communists when you mishmash together China and Taiwan.
I think you misread the thrust of this comment. Commenter was likely alluding not to R&D, but an amphibious invasion, as to how it would end up being China. Obviously I don't know how much of that infrastructure would persist to be carried off after that invasion, but it would reasonably be considered 'lost' to the US.
AFAIK (I might know wrong) the consensus among macroeconomists is that "social market economies" like European ones are better suited (compared to "liberal market economies" like the USA) to investing in large scale manufacturing (see e.g. Varieties of Capitalism [0]—though it is a bit controversial).
Though I do wonder if chip manufacturing might just be too large for the EU to handle. Many people in Europe keeps saying that e.g. Ericsson could start a fab but I'd be curious to see if a bill like this would succeed in the EU.
[0] https://en.wikipedia.org/wiki/Varieties_of_Capitalism
And there's new legislation encouraging investment in the cutting edge stuff.
Deleted Comment
The correct solution to that is tariffs. That would hurt us short term and should have been done long ago before it would have been a problem. This will add to the debt pile, so it will be with us for a long time.
What are these? Can you quote them?
Pompeo could qualify I suppose? Though he was Secretary of State, which seems to be more than a "personality".
The rest of the "details" are about the opinions of congressmen, not "personalities".
I find those remarks pretty naive. Every major power block heavily subsidizes strategic industries, now and forever. It's in no way new or remarkable, nor does it violate some principle. Most of the above concepts are pure fiction.
Semiconductors are not a "free market" in any ordinary sense. It costs tens of billions to enter the market and you'll buy an ASML machine, speaking of dependencies.
It's pretty obvious that this is in response to geopolitical instability. Being geo locked and this not really being a free market at all justify the "unusual" decision.
As is stands, food, energy, and yes...also semiconductors are foundational to a modern digital society. And they're interconnected, without semiconductors you won't have food either.
If you want to know what happens when semiconductors dry up, watch modern Russia. Arguably it's hard to do because of all the propaganda, but pretty much every domestic industry is falling apart, and therefore society falls apart.
Bottom line, don't be naive. Sure it sucks to give rich companies even more money but if that's what it takes, so be it. Also, rich is just one side of the coin, the other is that they spear-headed the information revolution.
Unless we actually start seeing cyclical job losses finally hit right when this money starts being spent. Which is entirely possible.
Labor is tight, but once production is up the supply curve shifts and prices start going down.
I'd like to share personal experience with how we bankrupted American leadership in cutting edge nodes. Although, it is not lithography related, I was part of a few billion $ ROI program where we'd hot test the chip for binning, best I don't disclose too many details. Let's just say, it was critical so much so that I sat in unmarked buildings. I saw that get transfered under my personal watch to China. We had Chinese employees visit US for 6 months at a time and during this rotation, we'd teach them everything. Had to take a Chinese culture course. Process charts, metrics, drawings and schematics, whitepapers, how everything works, be part of troubleshooting process and then test them for their acuity. This was around 2012-2014. Usually, US semiconductor manufacturers do not transfer fab capabilities to China, only assembly/packaging. But, here, the was a clear violation of backend fab activities that were transferred to China and built out. I visited China for 3 months to get things up and running. This was a brand new process that no one in the world has. All custom equipment from a major Japanese equipment manufacturer. This process was so insane that it took 10 years of development internally to come to this point. Even today, in 2022, no one has replicated it.
This should not have happened IMO from a national security standpoint. But, these things continue to happen and US gov does not have enough insight into America's semiconductor industry when it comes to protecting IP. Far too many things do not require ITAR and are exported without oversight.
I am pretty much against over-regulation, but here there needs to be strict regulation for exporting any semiconductor technologies whether it is fab or assembly or what have you. The entire industry needs to be hamstrung with export control.
The bill says "for section 9902/9906 of Public Law 116-283". This can be found on page 1460 here: https://www.congress.gov/116/plaws/publ283/PLAW-116publ283.p...
This is completely the wrong way to go about it (tax individuals and give the money to big corps). If you want a protectionist measure then implement a protectionist measure, like I don't know, tariffs.
Yet another bill that follows at least the last 40 years of congress doing exactly the wrong thing. (and to be clear this isn't a rant against the Democrats).
> The bill also includes about $100 billion in authorizations over five years for programs such as expanding the National Science Foundation’s work and establishing regional technology hubs to support start-ups in areas of the country that haven’t traditionally drawn big funding for tech.
I can only imagine how bureaucratic the process for accessing those funds will be. More likely, Medium Tech and even Big Tech will be the ones who feed from that trough because they're the only ones willing or able to cut through the red tape.
They should be paying out a large sum ONLY on completion, defined as when the first 10 million semiconductors are delivered to commercial customers.
Deleted Comment