I think I've covered the technical and financial aspects quite clearly. The political side does not have much room here but I will add a link to where politics meets law and some scenarios that could result.
There is no point discussing more on the subject as you simply have a pre-formed opinion and ignore anything that doesn't fit into it.
From a technical perspective your understanding of current events is severely lacking.
Military hardware does not, and never has, employed the latest nodes for semiconductor usage. It is not mature enough and costs are prohibitive.
Military hardware is a combination of bespoke silicon and off-the-shelf silicon.
Procurement has always been a huge problem for military usage as hardware has a very long service life.
When using commercial off-the-shelf options, that long service life often means there are procurement issues down the line as the nodes in question are often not economically viable for outputting semiconductors for military usage.
That is a major consideration. It applies to bespoke silicon for military usage too. Ironically, commercial bespoke silicon has seen high demand over the last few years but there is a huge difference.
Military usage does not generate the unit numbers required to make things financially viable - even on old nodes. The military simply does not order enough product. It is therefore an expensive business.
Where does this tie in with your China claims?
China already has 14nm and 7nm capacity.
In military terms, yields and cost are completely irrelevant in the face of sanctions. They don't need to pump out millions of processors and have all the finances needed to pay for it!
Sanctions have zero impact on China's military ambitions. No one (well, except you for sure) doubts that from a military perspective.
Away from military hardware but still on military usage, AI can benefit from the latest advances but China for sure has plenty of AI capacity already. What AI needs is to be 'fed' and it gorges on data. And who is better placed to get data into those systems? It certainly isn't the US.
So, from a 'national security' perspective, the sanctions mean literally nothing.
This is about technology and the rise of China as a competitor to the US. Nothing more.
OK, as promised, the legal side:
Take note of the contents of this paper (.Pdf ) and try to imagine what the possible outcomes may be.
I think I've covered the technical and financial aspects quite clearly. The political side does not have much room here but I will add a link to where politics meets law and some scenarios that could result.
There is no point discussing more on the subject as you simply have a pre-formed opinion and ignore anything that doesn't fit into it.
From a technical perspective your understanding of current events is severely lacking.
Military hardware does not, and never has, employed the latest nodes for semiconductor usage. It is not mature enough and costs are prohibitive.
Military hardware is a combination of bespoke silicon and off-the-shelf silicon.
Procurement has always been a huge problem for military usage as hardware has a very long service life.
When using commercial off-the-shelf options, that long service life often means there are procurement issues down the line as the nodes in question are often not economically viable for outputting semiconductors for military usage.
That is a major consideration. It applies to bespoke silicon for military usage too. Ironically, commercial bespoke silicon using various nodes has seen high demand over the last few years but there is a huge difference.
Military usage does not generate the unit numbers required to make things financially viable - even on old nodes. The military simply does not order enough product. It is therefore an expensive business.
Google, Meta et tal can go that route because they can order thousands upon thousands of units.
Where does this tie in with your China claims?
China already has 14nm and 7nm capacity. The US does not have 7nm capacity. That was one of the key takeaways from the Techinsights report. It was a wake up call because no one thought it could happen. Yourself included. Much less last year (which is when they began shipping the hardware). I'm sure there was a certain amount of panic running through the corridors of US power on that news.
In military terms, yields and cost are completely irrelevant in the face of sanctions. They don't need to pump out millions of processors and have all the finances needed to pay for it!
Sanctions have zero impact on China's military ambitions. No one (well, except you, for sure) doubts that from a military perspective. No one questions that
Away from military hardware but still on military usage, AI can benefit from the latest advances but China for sure has plenty of AI capacity already. What AI needs is to be 'fed' and it gorges on data. And who is better placed to get data into those systems? It certainly isn't the US.
So, from a 'national security' perspective, the sanctions mean literally nothing.
This is about technology and the rise of China as a competitor to the US. Nothing more.
OK, as promised, the legal side:
Take note of the contents of this paper (.Pdf ) and try to imagine what the possible outcomes may be.
That's it from me. I've spelt things out to you in clear and simple terms while trying to focus on the technological side of things.
The rest is simply a question of time.
You have a weak point in your arguments. You don't provide facts to support your statements. For example, Military hardware does not, and never has, employed the latest nodes for semiconductor usage.
How do you know? You need to provide facts to make people believe you.
I think I've covered the technical and financial aspects quite clearly. The political side does not have much room here but I will add a link to where politics meets law and some scenarios that could result.
There is no point discussing more on the subject as you simply have a pre-formed opinion and ignore anything that doesn't fit into it.
From a technical perspective your understanding of current events is severely lacking.
Military hardware does not, and never has, employed the latest nodes for semiconductor usage. It is not mature enough and costs are prohibitive.
Military hardware is a combination of bespoke silicon and off-the-shelf silicon.
Procurement has always been a huge problem for military usage as hardware has a very long service life.
When using commercial off-the-shelf options, that long service life often means there are procurement issues down the line as the nodes in question are often not economically viable for outputting semiconductors for military usage.
That is a major consideration. It applies to bespoke silicon for military usage too. Ironically, commercial bespoke silicon using various nodes has seen high demand over the last few years but there is a huge difference.
Military usage does not generate the unit numbers required to make things financially viable - even on old nodes. The military simply does not order enough product. It is therefore an expensive business.
Google, Meta et tal can go that route because they can order thousands upon thousands of units.
Where does this tie in with your China claims?
China already has 14nm and 7nm capacity. The US does not have 7nm capacity. That was one of the key takeaways from the Techinsights report. It was a wake up call because no one thought it could happen. Yourself included. Much less last year (which is when they began shipping the hardware). I'm sure there was a certain amount of panic running through the corridors of US power on that news.
In military terms, yields and cost are completely irrelevant in the face of sanctions. They don't need to pump out millions of processors and have all the finances needed to pay for it!
Sanctions have zero impact on China's military ambitions. No one (well, except you, for sure) doubts that from a military perspective. No one questions that
Away from military hardware but still on military usage, AI can benefit from the latest advances but China for sure has plenty of AI capacity already. What AI needs is to be 'fed' and it gorges on data. And who is better placed to get data into those systems? It certainly isn't the US.
So, from a 'national security' perspective, the sanctions mean literally nothing.
This is about technology and the rise of China as a competitor to the US. Nothing more.
OK, as promised, the legal side:
Take note of the contents of this paper (.Pdf ) and try to imagine what the possible outcomes may be.
That's it from me. I've spelt things out to you in clear and simple terms while trying to focus on the technological side of things.
The rest is simply a question of time.
You have a weak point in your arguments. You don't provide facts to support your statements. For example, Military hardware does not, and never has, employed the latest nodes for semiconductor usage.
How do you know? You need to provide facts to make people believe you.
You are aware that Military hardware designs initiated today will be based on the best available technology, not older technology. So the U.S. has the TSMC fab coming on line in 2024 at 5nm, and that is going to be the production node for many of those components, given the security of having production in the U.S. Intel will also open its fab in Ohio to custom orders. In the meantime, companies can still source a number of different nodes from TSMC in Taiwan, including leading edge nodes.
How do I know? I manufacture RF chassis components for various electronic manufacturers that are located in my area, all of which are in the supply chain for the military. They are designing to the latest available technology knowing that all of the finished product will likely have a long life cycle. So even if they are not using the latest SOC's, or any SOC's at all, there are others in the supply chain that do require the latest generations. Most of this has to do with performance per watt, especially in misses and aircraft, where weight and bulk need to be minimized.
So China, being restricted on Western technology, is faced with limited production at 14nm, and even less production at 7nm. While some of the military hardware will use that in the future, there is no reliable timeline when anything better becomes available. China will trail behind the West and that is a massive disadvantage, of three to four generations of SOC's, ie, bigger, bulkier. hotter electronics for their missiles and aircraft, again, a disadvantage.
I think I've covered the technical and financial aspects quite clearly. The political side does not have much room here but I will add a link to where politics meets law and some scenarios that could result.
There is no point discussing more on the subject as you simply have a pre-formed opinion and ignore anything that doesn't fit into it.
From a technical perspective your understanding of current events is severely lacking.
Military hardware does not, and never has, employed the latest nodes for semiconductor usage. It is not mature enough and costs are prohibitive.
Military hardware is a combination of bespoke silicon and off-the-shelf silicon.
Procurement has always been a huge problem for military usage as hardware has a very long service life.
When using commercial off-the-shelf options, that long service life often means there are procurement issues down the line as the nodes in question are often not economically viable for outputting semiconductors for military usage.
That is a major consideration. It applies to bespoke silicon for military usage too. Ironically, commercial bespoke silicon using various nodes has seen high demand over the last few years but there is a huge difference.
Military usage does not generate the unit numbers required to make things financially viable - even on old nodes. The military simply does not order enough product. It is therefore an expensive business.
Google, Meta et tal can go that route because they can order thousands upon thousands of units.
Where does this tie in with your China claims?
China already has 14nm and 7nm capacity. The US does not have 7nm capacity. That was one of the key takeaways from the Techinsights report. It was a wake up call because no one thought it could happen. Yourself included. Much less last year (which is when they began shipping the hardware). I'm sure there was a certain amount of panic running through the corridors of US power on that news.
In military terms, yields and cost are completely irrelevant in the face of sanctions. They don't need to pump out millions of processors and have all the finances needed to pay for it!
Sanctions have zero impact on China's military ambitions. No one (well, except you, for sure) doubts that from a military perspective. No one questions that
Away from military hardware but still on military usage, AI can benefit from the latest advances but China for sure has plenty of AI capacity already. What AI needs is to be 'fed' and it gorges on data. And who is better placed to get data into those systems? It certainly isn't the US.
So, from a 'national security' perspective, the sanctions mean literally nothing.
This is about technology and the rise of China as a competitor to the US. Nothing more.
OK, as promised, the legal side:
Take note of the contents of this paper (.Pdf ) and try to imagine what the possible outcomes may be.
That's it from me. I've spelt things out to you in clear and simple terms while trying to focus on the technological side of things.
The rest is simply a question of time.
You have a weak point in your arguments. You don't provide facts to support your statements. For example, Military hardware does not, and never has, employed the latest nodes for semiconductor usage.
How do you know? You need to provide facts to make people believe you.
It should be common knowledge for anyone who has even a passing interest in processor development and to be honest I've given Tmay lots of links over the years.
Here is the first article that Google returned (published yesterday):
"The fiction of “barring military use” is only to provide the fig leaf of a cover under the World Trade Organization (WTO) exceptions on having to provide market access to all WTO members. Most military applications use older-generation chips and not the latest versions."
New process nodes bring new bugs. Software tools need to be updated etc.
You just can't rush the latest process node into your products. Often it isn't even necessary. Most of the cutting edge nodes end up in commercial products because those products can benefit from the power and power efficiencies they bring and be viable economically speaking. Cost is a major concern as some military semiconductor hardware has to be 'hardened' too. For example radiation hardened or heat hardened. Even in a commercial setting, manufacturers may decide to go with 'last year's' design in order to avoid problems. Huawei did that on one of its flagship phones when it chose to run with an older ARM design on its Soc one year. The reason given was immaturity.
Military designs are long lifespan. They get upgraded along the way but only with mature, tried and tested product components although the ultimate test is always in battle scenarios.
Typically, newer nodes bring more transistor density and power and size efficiencies. If your design can function well with current mature technology there is just no good reason to move up. That is precisely why 98% of the world's semiconductor output is nowhere near the cutting edge nodes. There is literally no need for it in IoT, industrial applications, the automotive industry etc.
As a stop gap solution, it is rumored that Huawei will bring to market a new chip stacked 14nm Kirin next year. There is nothing really new about chip stacking except that in this case the actual process is said to be much cheaper and very efficient. It should remain competitive even in 2023. And that would be 14nm.
With regards to off the shelf semiconductors, there is actually a military term for that, the name of which escapes me but it has 'off-the-shelf' in it.
I think I've covered the technical and financial aspects quite clearly. The political side does not have much room here but I will add a link to where politics meets law and some scenarios that could result.
There is no point discussing more on the subject as you simply have a pre-formed opinion and ignore anything that doesn't fit into it.
From a technical perspective your understanding of current events is severely lacking.
Military hardware does not, and never has, employed the latest nodes for semiconductor usage. It is not mature enough and costs are prohibitive.
Military hardware is a combination of bespoke silicon and off-the-shelf silicon.
Procurement has always been a huge problem for military usage as hardware has a very long service life.
When using commercial off-the-shelf options, that long service life often means there are procurement issues down the line as the nodes in question are often not economically viable for outputting semiconductors for military usage.
That is a major consideration. It applies to bespoke silicon for military usage too. Ironically, commercial bespoke silicon using various nodes has seen high demand over the last few years but there is a huge difference.
Military usage does not generate the unit numbers required to make things financially viable - even on old nodes. The military simply does not order enough product. It is therefore an expensive business.
Google, Meta et tal can go that route because they can order thousands upon thousands of units.
Where does this tie in with your China claims?
China already has 14nm and 7nm capacity. The US does not have 7nm capacity. That was one of the key takeaways from the Techinsights report. It was a wake up call because no one thought it could happen. Yourself included. Much less last year (which is when they began shipping the hardware). I'm sure there was a certain amount of panic running through the corridors of US power on that news.
In military terms, yields and cost are completely irrelevant in the face of sanctions. They don't need to pump out millions of processors and have all the finances needed to pay for it!
Sanctions have zero impact on China's military ambitions. No one (well, except you, for sure) doubts that from a military perspective. No one questions that
Away from military hardware but still on military usage, AI can benefit from the latest advances but China for sure has plenty of AI capacity already. What AI needs is to be 'fed' and it gorges on data. And who is better placed to get data into those systems? It certainly isn't the US.
So, from a 'national security' perspective, the sanctions mean literally nothing.
This is about technology and the rise of China as a competitor to the US. Nothing more.
OK, as promised, the legal side:
Take note of the contents of this paper (.Pdf ) and try to imagine what the possible outcomes may be.
That's it from me. I've spelt things out to you in clear and simple terms while trying to focus on the technological side of things.
The rest is simply a question of time.
You have a weak point in your arguments. You don't provide facts to support your statements. For example, Military hardware does not, and never has, employed the latest nodes for semiconductor usage.
How do you know? You need to provide facts to make people believe you.
It should be common knowledge for anyone who has even a passing interest in processor development and to be honest I've given Tmay lots of links over the years.
Here is the first article that Google returned (published yesterday):
"The fiction of “barring military use” is only to provide the fig leaf of a cover under the World Trade Organization (WTO) exceptions on having to provide market access to all WTO members. Most military applications use older-generation chips and not the latest versions."
New process nodes bring new bugs. Software tools need to be updated etc.
You just can't rush the latest process node into your products. Often it isn't even necessary. Most of the cutting edge nodes end up in commercial products because those products can benefit from the power and power efficiencies they bring and be viable economically speaking. Cost is a major concern as some military semiconductor hardware has to be 'hardened' too. For example radiation hardened or heat hardened. Even in a commercial setting, manufacturers may decide to go with 'last year's' design in order to avoid problems. Huawei did that on one of its flagship phones when it chose to run with an older ARM design on its Soc one year. The reason given was immaturity.
Military designs are long lifespan. They get upgraded along the way but only with mature, tried and tested product components although the ultimate test is always in battle scenarios.
Typically, newer nodes bring more transistor density and power and size efficiencies. If your design can function well with current mature technology there is just no good reason to move up. That is precisely why 98% of the world's semiconductor output is nowhere near the cutting edge nodes. There is literally no need for it in IoT, industrial applications, the automotive industry etc.
As a stop gap solution, it is rumored that Huawei will bring to market a new chip stacked 14nm Kirin next year. There is nothing really new about chip stacking except that in this case the actual process is said to be much cheaper and very efficient. It should remain competitive even in 2023. And that would be 14nm.
With regards to off the shelf semiconductors, there is actually a military term for that, the name of which escapes me but it has 'off-the-shelf' in it.
Great. You posted a link to asiatimes, and we now have the Indian point of view, and you are still attempting to post as if you have a clue about the military.
For the record, India has been attempting to wean itself off of cheap Russian Military equipment, and now finds itself stranded as little of that equipment is available to export. More to the point, India stands to gain from resourcing supply chains removed from China. I'm sure the French will sell some aircraft to them, given that their home-brew program is again a fail.
Should I mention that the B-21 will be unveiled to the public on December 2? Yet another program that will benefit from 5nm.
You probably can't imagine the number of new and ongoing programs recently spurred on by the Russian invasion of Ukraine, and the ascent of Xi for yet another term, but I'll bet it's numerous.
Oh, and for the record, it appears that China doesn't get to have technology that they will ultimately use against us.
I think I've covered the technical and financial aspects quite clearly. The political side does not have much room here but I will add a link to where politics meets law and some scenarios that could result.
There is no point discussing more on the subject as you simply have a pre-formed opinion and ignore anything that doesn't fit into it.
From a technical perspective your understanding of current events is severely lacking.
Military hardware does not, and never has, employed the latest nodes for semiconductor usage. It is not mature enough and costs are prohibitive.
Military hardware is a combination of bespoke silicon and off-the-shelf silicon.
Procurement has always been a huge problem for military usage as hardware has a very long service life.
When using commercial off-the-shelf options, that long service life often means there are procurement issues down the line as the nodes in question are often not economically viable for outputting semiconductors for military usage.
That is a major consideration. It applies to bespoke silicon for military usage too. Ironically, commercial bespoke silicon using various nodes has seen high demand over the last few years but there is a huge difference.
Military usage does not generate the unit numbers required to make things financially viable - even on old nodes. The military simply does not order enough product. It is therefore an expensive business.
Google, Meta et tal can go that route because they can order thousands upon thousands of units.
Where does this tie in with your China claims?
China already has 14nm and 7nm capacity. The US does not have 7nm capacity. That was one of the key takeaways from the Techinsights report. It was a wake up call because no one thought it could happen. Yourself included. Much less last year (which is when they began shipping the hardware). I'm sure there was a certain amount of panic running through the corridors of US power on that news.
In military terms, yields and cost are completely irrelevant in the face of sanctions. They don't need to pump out millions of processors and have all the finances needed to pay for it!
Sanctions have zero impact on China's military ambitions. No one (well, except you, for sure) doubts that from a military perspective. No one questions that
Away from military hardware but still on military usage, AI can benefit from the latest advances but China for sure has plenty of AI capacity already. What AI needs is to be 'fed' and it gorges on data. And who is better placed to get data into those systems? It certainly isn't the US.
So, from a 'national security' perspective, the sanctions mean literally nothing.
This is about technology and the rise of China as a competitor to the US. Nothing more.
OK, as promised, the legal side:
Take note of the contents of this paper (.Pdf ) and try to imagine what the possible outcomes may be.
That's it from me. I've spelt things out to you in clear and simple terms while trying to focus on the technological side of things.
The rest is simply a question of time.
You have a weak point in your arguments. You don't provide facts to support your statements. For example, Military hardware does not, and never has, employed the latest nodes for semiconductor usage.
How do you know? You need to provide facts to make people believe you.
You are aware that Military hardware designs initiated today will be based on the best available technology, not older technology. So the U.S. has the TSMC fab coming on line in 2024 at 5nm, and that is going to be the production node for many of those components, given the security of having production in the U.S. Intel will also open its fab in Ohio to custom orders. In the meantime, companies can still source a number of different nodes from TSMC in Taiwan, including leading edge nodes.
How do I know? I manufacture RF chassis components for various electronic manufacturers that are located in my area, all of which are in the supply chain for the military. They are designing to the latest available technology knowing that all of the finished product will likely have a long life cycle. So even if they are not using the latest SOC's, or any SOC's at all, there are others in the supply chain that do require the latest generations. Most of this has to do with performance per watt, especially in misses and aircraft, where weight and bulk need to be minimized.
So China, being restricted on Western technology, is faced with limited production at 14nm, and even less production at 7nm. While some of the military hardware will use that in the future, there is no reliable timeline when anything better becomes available. China will trail behind the West and that is a massive disadvantage, of three to four generations of SOC's, ie, bigger, bulkier. hotter electronics for their missiles and aircraft, again, a disadvantage.
You remind me of one thing. The Military Requirement requires chips should work in very harsh environment. The materials TSMC uses in the fabs do not satisfy it.
I think I've covered the technical and financial aspects quite clearly. The political side does not have much room here but I will add a link to where politics meets law and some scenarios that could result.
There is no point discussing more on the subject as you simply have a pre-formed opinion and ignore anything that doesn't fit into it.
From a technical perspective your understanding of current events is severely lacking.
Military hardware does not, and never has, employed the latest nodes for semiconductor usage. It is not mature enough and costs are prohibitive.
Military hardware is a combination of bespoke silicon and off-the-shelf silicon.
Procurement has always been a huge problem for military usage as hardware has a very long service life.
When using commercial off-the-shelf options, that long service life often means there are procurement issues down the line as the nodes in question are often not economically viable for outputting semiconductors for military usage.
That is a major consideration. It applies to bespoke silicon for military usage too. Ironically, commercial bespoke silicon using various nodes has seen high demand over the last few years but there is a huge difference.
Military usage does not generate the unit numbers required to make things financially viable - even on old nodes. The military simply does not order enough product. It is therefore an expensive business.
Google, Meta et tal can go that route because they can order thousands upon thousands of units.
Where does this tie in with your China claims?
China already has 14nm and 7nm capacity. The US does not have 7nm capacity. That was one of the key takeaways from the Techinsights report. It was a wake up call because no one thought it could happen. Yourself included. Much less last year (which is when they began shipping the hardware). I'm sure there was a certain amount of panic running through the corridors of US power on that news.
In military terms, yields and cost are completely irrelevant in the face of sanctions. They don't need to pump out millions of processors and have all the finances needed to pay for it!
Sanctions have zero impact on China's military ambitions. No one (well, except you, for sure) doubts that from a military perspective. No one questions that
Away from military hardware but still on military usage, AI can benefit from the latest advances but China for sure has plenty of AI capacity already. What AI needs is to be 'fed' and it gorges on data. And who is better placed to get data into those systems? It certainly isn't the US.
So, from a 'national security' perspective, the sanctions mean literally nothing.
This is about technology and the rise of China as a competitor to the US. Nothing more.
OK, as promised, the legal side:
Take note of the contents of this paper (.Pdf ) and try to imagine what the possible outcomes may be.
That's it from me. I've spelt things out to you in clear and simple terms while trying to focus on the technological side of things.
The rest is simply a question of time.
You have a weak point in your arguments. You don't provide facts to support your statements. For example, Military hardware does not, and never has, employed the latest nodes for semiconductor usage.
How do you know? You need to provide facts to make people believe you.
You are aware that Military hardware designs initiated today will be based on the best available technology, not older technology. So the U.S. has the TSMC fab coming on line in 2024 at 5nm, and that is going to be the production node for many of those components, given the security of having production in the U.S. Intel will also open its fab in Ohio to custom orders. In the meantime, companies can still source a number of different nodes from TSMC in Taiwan, including leading edge nodes.
How do I know? I manufacture RF chassis components for various electronic manufacturers that are located in my area, all of which are in the supply chain for the military. They are designing to the latest available technology knowing that all of the finished product will likely have a long life cycle. So even if they are not using the latest SOC's, or any SOC's at all, there are others in the supply chain that do require the latest generations. Most of this has to do with performance per watt, especially in misses and aircraft, where weight and bulk need to be minimized.
So China, being restricted on Western technology, is faced with limited production at 14nm, and even less production at 7nm. While some of the military hardware will use that in the future, there is no reliable timeline when anything better becomes available. China will trail behind the West and that is a massive disadvantage, of three to four generations of SOC's, ie, bigger, bulkier. hotter electronics for their missiles and aircraft, again, a disadvantage.
You remind me of one thing. The Military Requirement requires chips should work in very harsh environment. The materials TSMC uses in the fabs do not satisfy it.
About 2,470,000 results (0.49 seconds)
Search Results
Featured snippet from the web
U.S. dependency on Taiwanese production of chips for defense systems extends beyond AI. TSMC makes semiconductors used in F-35 fighters and a wide range of “military-grade” devices used by the U.S. Department of Defense (DOD)
All major U.S. defense systems and platforms rely on semiconductors for their performance. Consequently, the erosion of U.S. capabilities in microelectronics is a direct threat to the United States’ ability to defend itself and its allies. Moreover, the U.S. civilian economy is deeply dependent on semiconductor-based platforms for its daily operations. Ensuring U.S. leadership in semiconductor technology and securing the integrity of the value chains that design, manufacture, package, and distribute these chips are perhaps the preeminent economic and national security concerns of the modern era.
Of course, U.S. leaders have recognized the nexus between semiconductors and national security for decades. In the mid-1980s, President Reagan articulated the need to retain U.S. global leadership in chips to counter the Soviet bloc’s numerically superior military forces through “smart” systems driven by semiconductors—satellites, stealth aircraft, cruise missiles, and the like. Today, the United States faces an even greater challenge—it is confronted by a geopolitical rival that is stronger economically and technologically than the Soviet Union ever was. At the same time, the United States is struggling to halt and reverse the loss of its leading capabilities in manufacturing microelectronics. The recent shortage of automotive chips has driven home the damage that the steady erosion of capabilities in chip manufacturing can inflict on the U.S. civilian economy. It is equally important to appreciate the national defense implications of recent adverse trends.
The current chip legislation, which at this writing is pending reconciliation in Congress, represents the most comprehensive program of federal support ever undertaken to ensure the viability of a domestic semiconductor industry—that is, an industry whose critical functions are located within the geography of the United States. The legislation is both essential and overdue. Despite the dwindling number of critics who oppose the legislation as “industrial policy,” federal measures to ensure the existence of a robust domestic commercial semiconductor industry—as President Reagan recognized long ago—are vital to U.S. national defense and to the strength and resilience of the U.S. national economy.
Declining Onshore Capabilities
In terms of onshore capability, the U.S. semiconductor industry now has significant gaps in its production chain. The United States remains the unchallenged world leader in semiconductor design, controlling about 85 percent of the world market for electronic design automation (EDA) tools, which are necessary for the design of the most advanced chips. However, there are factors to consider, such as:
,,,
AI Challenge from China
These production chain gaps are problematic from a U.S. national security perspective. China has emerged as a major strategic challenger to the United States and is investing heavily in developing its military power and defense industrial base, placing a priority on overtaking the United States and its allies in semiconductor technology. As an independent commission established by Congress recently concluded: “If a potential adversary bests the United States in semiconductors over the long term or suddenly cuts off U.S. access to cutting-edge chips entirely, it could gain the upper hand in every domain of warfare.” U.S. vulnerability is particularly acute with respect to the most advanced chips currently in production, which are essential to the creation and application of artificial intelligence (AI)—intelligence generated by machines—which is expected to revolutionize warfare.
China’s leaders have set a goal to build a “fully modern” military by 2027 based on “informatization,” “intelligentization,” and “mechanization,” investing heavily in technical areas which support such an approach, such as AI, quantum computing, hypersonics, and microelectronics. AI enables computer systems to solve problems and address tasks that normally require human intelligence, ultimately at speeds and performance levels that vastly exceed those of humans. In the words of the Japanese National Institute for Defense Studies: “As AI does not get fatigued, does not forget, and has no emotional fluctuation, AI is expected to be able to help commanders make decisions by processing large quantities of data quickly and accurately.” In a future war, according to U.S. Senator Mike Rounds, “defending against AI-capable adversaries operating at machine speeds without employing AI is an invitation to disaster.” Human operators cannot outmatch multiple machines making thousands of decisions per second coordinated across various systems, nor will they be able to counter an adversary’s AI-enabled missile attack, strike against communications satellites and infrastructure, or coordinate firepower strikes, drone swarms, cyberattacks, and other twenty-first century threats. For this reason, the U.S. armed forces’ current technical edge over all potential adversaries “could be lost within the next decade if they do not accelerate the adoption of AI across their missions.” A massive AI-driven Chinese attack could overwhelm U.S. defenses.
AI systems operate on a foundation of interconnected computer hardware driven by cutting-edge semiconductor devices. “Cutting-edge” is a critical term in this context—it is almost impossible to overstate the performance disparity between advanced AI chips and conventional semiconductor central processing units (CPU). The most advanced chips are tens or even thousands of times faster than CPUs in the development of AI algorithms. “An AI chip a thousand times as efficient as a CPU provides an improvement equivalent to 26 years of Moore’s Law-driven CPU improvements,” according to Georgetown’s Center for Security and Emerging Technology. “State-of-the-art AI chips are necessary for the cost-effective, fast development, and deployment of advanced security-relevant AI systems.”
The most advanced AI systems require semiconductor chips based on 7 nm to 5 nm design rules, which are not currently manufacturable in the United States. Intel makes field-programmable gate arrays (FPGA), which are incorporated in AI systems, based on 10 nm design rules—a generation behind 7 nm. The company will begin U.S.-based production of 7 nm chips in 2022. Taiwan Semiconductor Manufacturing Co. (TSMC) is building a fab in Arizona which will operate at the 5 nm node beginning around 2024. But by this time, the state of the art is likely to have moved to 3 nm chips, all of which will be made in Taiwan.
So, China doesn't get cutting edge semiconductors. Problem solved.
I think I've covered the technical and financial aspects quite clearly. The political side does not have much room here but I will add a link to where politics meets law and some scenarios that could result.
There is no point discussing more on the subject as you simply have a pre-formed opinion and ignore anything that doesn't fit into it.
From a technical perspective your understanding of current events is severely lacking.
Military hardware does not, and never has, employed the latest nodes for semiconductor usage. It is not mature enough and costs are prohibitive.
Military hardware is a combination of bespoke silicon and off-the-shelf silicon.
Procurement has always been a huge problem for military usage as hardware has a very long service life.
When using commercial off-the-shelf options, that long service life often means there are procurement issues down the line as the nodes in question are often not economically viable for outputting semiconductors for military usage.
That is a major consideration. It applies to bespoke silicon for military usage too. Ironically, commercial bespoke silicon using various nodes has seen high demand over the last few years but there is a huge difference.
Military usage does not generate the unit numbers required to make things financially viable - even on old nodes. The military simply does not order enough product. It is therefore an expensive business.
Google, Meta et tal can go that route because they can order thousands upon thousands of units.
Where does this tie in with your China claims?
China already has 14nm and 7nm capacity. The US does not have 7nm capacity. That was one of the key takeaways from the Techinsights report. It was a wake up call because no one thought it could happen. Yourself included. Much less last year (which is when they began shipping the hardware). I'm sure there was a certain amount of panic running through the corridors of US power on that news.
In military terms, yields and cost are completely irrelevant in the face of sanctions. They don't need to pump out millions of processors and have all the finances needed to pay for it!
Sanctions have zero impact on China's military ambitions. No one (well, except you, for sure) doubts that from a military perspective. No one questions that
Away from military hardware but still on military usage, AI can benefit from the latest advances but China for sure has plenty of AI capacity already. What AI needs is to be 'fed' and it gorges on data. And who is better placed to get data into those systems? It certainly isn't the US.
So, from a 'national security' perspective, the sanctions mean literally nothing.
This is about technology and the rise of China as a competitor to the US. Nothing more.
OK, as promised, the legal side:
Take note of the contents of this paper (.Pdf ) and try to imagine what the possible outcomes may be.
That's it from me. I've spelt things out to you in clear and simple terms while trying to focus on the technological side of things.
The rest is simply a question of time.
You have a weak point in your arguments. You don't provide facts to support your statements. For example, Military hardware does not, and never has, employed the latest nodes for semiconductor usage.
How do you know? You need to provide facts to make people believe you.
You are aware that Military hardware designs initiated today will be based on the best available technology, not older technology. So the U.S. has the TSMC fab coming on line in 2024 at 5nm, and that is going to be the production node for many of those components, given the security of having production in the U.S. Intel will also open its fab in Ohio to custom orders. In the meantime, companies can still source a number of different nodes from TSMC in Taiwan, including leading edge nodes.
How do I know? I manufacture RF chassis components for various electronic manufacturers that are located in my area, all of which are in the supply chain for the military. They are designing to the latest available technology knowing that all of the finished product will likely have a long life cycle. So even if they are not using the latest SOC's, or any SOC's at all, there are others in the supply chain that do require the latest generations. Most of this has to do with performance per watt, especially in misses and aircraft, where weight and bulk need to be minimized.
So China, being restricted on Western technology, is faced with limited production at 14nm, and even less production at 7nm. While some of the military hardware will use that in the future, there is no reliable timeline when anything better becomes available. China will trail behind the West and that is a massive disadvantage, of three to four generations of SOC's, ie, bigger, bulkier. hotter electronics for their missiles and aircraft, again, a disadvantage.
You remind me of one thing. The Military Requirement requires chips should work in very harsh environment. The materials TSMC uses in the fabs do not satisfy it.
About 2,470,000 results (0.49 seconds)
Search Results
Featured snippet from the web
U.S. dependency on Taiwanese production of chips for defense systems extends beyond AI. TSMC makes semiconductors used in F-35 fighters and a wide range of “military-grade” devices used by the U.S. Department of Defense (DOD)
All major U.S. defense systems and platforms rely on semiconductors for their performance. Consequently, the erosion of U.S. capabilities in microelectronics is a direct threat to the United States’ ability to defend itself and its allies. Moreover, the U.S. civilian economy is deeply dependent on semiconductor-based platforms for its daily operations. Ensuring U.S. leadership in semiconductor technology and securing the integrity of the value chains that design, manufacture, package, and distribute these chips are perhaps the preeminent economic and national security concerns of the modern era.
Of course, U.S. leaders have recognized the nexus between semiconductors and national security for decades. In the mid-1980s, President Reagan articulated the need to retain U.S. global leadership in chips to counter the Soviet bloc’s numerically superior military forces through “smart” systems driven by semiconductors—satellites, stealth aircraft, cruise missiles, and the like. Today, the United States faces an even greater challenge—it is confronted by a geopolitical rival that is stronger economically and technologically than the Soviet Union ever was. At the same time, the United States is struggling to halt and reverse the loss of its leading capabilities in manufacturing microelectronics. The recent shortage of automotive chips has driven home the damage that the steady erosion of capabilities in chip manufacturing can inflict on the U.S. civilian economy. It is equally important to appreciate the national defense implications of recent adverse trends.
The current chip legislation, which at this writing is pending reconciliation in Congress, represents the most comprehensive program of federal support ever undertaken to ensure the viability of a domestic semiconductor industry—that is, an industry whose critical functions are located within the geography of the United States. The legislation is both essential and overdue. Despite the dwindling number of critics who oppose the legislation as “industrial policy,” federal measures to ensure the existence of a robust domestic commercial semiconductor industry—as President Reagan recognized long ago—are vital to U.S. national defense and to the strength and resilience of the U.S. national economy.
Declining Onshore Capabilities
In terms of onshore capability, the U.S. semiconductor industry now has significant gaps in its production chain. The United States remains the unchallenged world leader in semiconductor design, controlling about 85 percent of the world market for electronic design automation (EDA) tools, which are necessary for the design of the most advanced chips. However, there are factors to consider, such as:
,,,
AI Challenge from China
These production chain gaps are problematic from a U.S. national security perspective. China has emerged as a major strategic challenger to the United States and is investing heavily in developing its military power and defense industrial base, placing a priority on overtaking the United States and its allies in semiconductor technology. As an independent commission established by Congress recently concluded: “If a potential adversary bests the United States in semiconductors over the long term or suddenly cuts off U.S. access to cutting-edge chips entirely, it could gain the upper hand in every domain of warfare.” U.S. vulnerability is particularly acute with respect to the most advanced chips currently in production, which are essential to the creation and application of artificial intelligence (AI)—intelligence generated by machines—which is expected to revolutionize warfare.
China’s leaders have set a goal to build a “fully modern” military by 2027 based on “informatization,” “intelligentization,” and “mechanization,” investing heavily in technical areas which support such an approach, such as AI, quantum computing, hypersonics, and microelectronics. AI enables computer systems to solve problems and address tasks that normally require human intelligence, ultimately at speeds and performance levels that vastly exceed those of humans. In the words of the Japanese National Institute for Defense Studies: “As AI does not get fatigued, does not forget, and has no emotional fluctuation, AI is expected to be able to help commanders make decisions by processing large quantities of data quickly and accurately.” In a future war, according to U.S. Senator Mike Rounds, “defending against AI-capable adversaries operating at machine speeds without employing AI is an invitation to disaster.” Human operators cannot outmatch multiple machines making thousands of decisions per second coordinated across various systems, nor will they be able to counter an adversary’s AI-enabled missile attack, strike against communications satellites and infrastructure, or coordinate firepower strikes, drone swarms, cyberattacks, and other twenty-first century threats. For this reason, the U.S. armed forces’ current technical edge over all potential adversaries “could be lost within the next decade if they do not accelerate the adoption of AI across their missions.” A massive AI-driven Chinese attack could overwhelm U.S. defenses.
AI systems operate on a foundation of interconnected computer hardware driven by cutting-edge semiconductor devices. “Cutting-edge” is a critical term in this context—it is almost impossible to overstate the performance disparity between advanced AI chips and conventional semiconductor central processing units (CPU). The most advanced chips are tens or even thousands of times faster than CPUs in the development of AI algorithms. “An AI chip a thousand times as efficient as a CPU provides an improvement equivalent to 26 years of Moore’s Law-driven CPU improvements,” according to Georgetown’s Center for Security and Emerging Technology. “State-of-the-art AI chips are necessary for the cost-effective, fast development, and deployment of advanced security-relevant AI systems.”
The most advanced AI systems require semiconductor chips based on 7 nm to 5 nm design rules, which are not currently manufacturable in the United States. Intel makes field-programmable gate arrays (FPGA), which are incorporated in AI systems, based on 10 nm design rules—a generation behind 7 nm. The company will begin U.S.-based production of 7 nm chips in 2022. Taiwan Semiconductor Manufacturing Co. (TSMC) is building a fab in Arizona which will operate at the 5 nm node beginning around 2024. But by this time, the state of the art is likely to have moved to 3 nm chips, all of which will be made in Taiwan.
So, China doesn't get cutting edge semiconductors. Problem solved.
Now, what else to you have...
You believe the article? What devices TSMC make? LOL
I think I've covered the technical and financial aspects quite clearly. The political side does not have much room here but I will add a link to where politics meets law and some scenarios that could result.
There is no point discussing more on the subject as you simply have a pre-formed opinion and ignore anything that doesn't fit into it.
From a technical perspective your understanding of current events is severely lacking.
Military hardware does not, and never has, employed the latest nodes for semiconductor usage. It is not mature enough and costs are prohibitive.
Military hardware is a combination of bespoke silicon and off-the-shelf silicon.
Procurement has always been a huge problem for military usage as hardware has a very long service life.
When using commercial off-the-shelf options, that long service life often means there are procurement issues down the line as the nodes in question are often not economically viable for outputting semiconductors for military usage.
That is a major consideration. It applies to bespoke silicon for military usage too. Ironically, commercial bespoke silicon using various nodes has seen high demand over the last few years but there is a huge difference.
Military usage does not generate the unit numbers required to make things financially viable - even on old nodes. The military simply does not order enough product. It is therefore an expensive business.
Google, Meta et tal can go that route because they can order thousands upon thousands of units.
Where does this tie in with your China claims?
China already has 14nm and 7nm capacity. The US does not have 7nm capacity. That was one of the key takeaways from the Techinsights report. It was a wake up call because no one thought it could happen. Yourself included. Much less last year (which is when they began shipping the hardware). I'm sure there was a certain amount of panic running through the corridors of US power on that news.
In military terms, yields and cost are completely irrelevant in the face of sanctions. They don't need to pump out millions of processors and have all the finances needed to pay for it!
Sanctions have zero impact on China's military ambitions. No one (well, except you, for sure) doubts that from a military perspective. No one questions that
Away from military hardware but still on military usage, AI can benefit from the latest advances but China for sure has plenty of AI capacity already. What AI needs is to be 'fed' and it gorges on data. And who is better placed to get data into those systems? It certainly isn't the US.
So, from a 'national security' perspective, the sanctions mean literally nothing.
This is about technology and the rise of China as a competitor to the US. Nothing more.
OK, as promised, the legal side:
Take note of the contents of this paper (.Pdf ) and try to imagine what the possible outcomes may be.
That's it from me. I've spelt things out to you in clear and simple terms while trying to focus on the technological side of things.
The rest is simply a question of time.
You have a weak point in your arguments. You don't provide facts to support your statements. For example, Military hardware does not, and never has, employed the latest nodes for semiconductor usage.
How do you know? You need to provide facts to make people believe you.
You are aware that Military hardware designs initiated today will be based on the best available technology, not older technology. So the U.S. has the TSMC fab coming on line in 2024 at 5nm, and that is going to be the production node for many of those components, given the security of having production in the U.S. Intel will also open its fab in Ohio to custom orders. In the meantime, companies can still source a number of different nodes from TSMC in Taiwan, including leading edge nodes.
How do I know? I manufacture RF chassis components for various electronic manufacturers that are located in my area, all of which are in the supply chain for the military. They are designing to the latest available technology knowing that all of the finished product will likely have a long life cycle. So even if they are not using the latest SOC's, or any SOC's at all, there are others in the supply chain that do require the latest generations. Most of this has to do with performance per watt, especially in misses and aircraft, where weight and bulk need to be minimized.
So China, being restricted on Western technology, is faced with limited production at 14nm, and even less production at 7nm. While some of the military hardware will use that in the future, there is no reliable timeline when anything better becomes available. China will trail behind the West and that is a massive disadvantage, of three to four generations of SOC's, ie, bigger, bulkier. hotter electronics for their missiles and aircraft, again, a disadvantage.
You remind me of one thing. The Military Requirement requires chips should work in very harsh environment. The materials TSMC uses in the fabs do not satisfy it.
About 2,470,000 results (0.49 seconds)
Search Results
Featured snippet from the web
U.S. dependency on Taiwanese production of chips for defense systems extends beyond AI. TSMC makes semiconductors used in F-35 fighters and a wide range of “military-grade” devices used by the U.S. Department of Defense (DOD)
All major U.S. defense systems and platforms rely on semiconductors for their performance. Consequently, the erosion of U.S. capabilities in microelectronics is a direct threat to the United States’ ability to defend itself and its allies. Moreover, the U.S. civilian economy is deeply dependent on semiconductor-based platforms for its daily operations. Ensuring U.S. leadership in semiconductor technology and securing the integrity of the value chains that design, manufacture, package, and distribute these chips are perhaps the preeminent economic and national security concerns of the modern era.
Of course, U.S. leaders have recognized the nexus between semiconductors and national security for decades. In the mid-1980s, President Reagan articulated the need to retain U.S. global leadership in chips to counter the Soviet bloc’s numerically superior military forces through “smart” systems driven by semiconductors—satellites, stealth aircraft, cruise missiles, and the like. Today, the United States faces an even greater challenge—it is confronted by a geopolitical rival that is stronger economically and technologically than the Soviet Union ever was. At the same time, the United States is struggling to halt and reverse the loss of its leading capabilities in manufacturing microelectronics. The recent shortage of automotive chips has driven home the damage that the steady erosion of capabilities in chip manufacturing can inflict on the U.S. civilian economy. It is equally important to appreciate the national defense implications of recent adverse trends.
The current chip legislation, which at this writing is pending reconciliation in Congress, represents the most comprehensive program of federal support ever undertaken to ensure the viability of a domestic semiconductor industry—that is, an industry whose critical functions are located within the geography of the United States. The legislation is both essential and overdue. Despite the dwindling number of critics who oppose the legislation as “industrial policy,” federal measures to ensure the existence of a robust domestic commercial semiconductor industry—as President Reagan recognized long ago—are vital to U.S. national defense and to the strength and resilience of the U.S. national economy.
Declining Onshore Capabilities
In terms of onshore capability, the U.S. semiconductor industry now has significant gaps in its production chain. The United States remains the unchallenged world leader in semiconductor design, controlling about 85 percent of the world market for electronic design automation (EDA) tools, which are necessary for the design of the most advanced chips. However, there are factors to consider, such as:
,,,
AI Challenge from China
These production chain gaps are problematic from a U.S. national security perspective. China has emerged as a major strategic challenger to the United States and is investing heavily in developing its military power and defense industrial base, placing a priority on overtaking the United States and its allies in semiconductor technology. As an independent commission established by Congress recently concluded: “If a potential adversary bests the United States in semiconductors over the long term or suddenly cuts off U.S. access to cutting-edge chips entirely, it could gain the upper hand in every domain of warfare.” U.S. vulnerability is particularly acute with respect to the most advanced chips currently in production, which are essential to the creation and application of artificial intelligence (AI)—intelligence generated by machines—which is expected to revolutionize warfare.
China’s leaders have set a goal to build a “fully modern” military by 2027 based on “informatization,” “intelligentization,” and “mechanization,” investing heavily in technical areas which support such an approach, such as AI, quantum computing, hypersonics, and microelectronics. AI enables computer systems to solve problems and address tasks that normally require human intelligence, ultimately at speeds and performance levels that vastly exceed those of humans. In the words of the Japanese National Institute for Defense Studies: “As AI does not get fatigued, does not forget, and has no emotional fluctuation, AI is expected to be able to help commanders make decisions by processing large quantities of data quickly and accurately.” In a future war, according to U.S. Senator Mike Rounds, “defending against AI-capable adversaries operating at machine speeds without employing AI is an invitation to disaster.” Human operators cannot outmatch multiple machines making thousands of decisions per second coordinated across various systems, nor will they be able to counter an adversary’s AI-enabled missile attack, strike against communications satellites and infrastructure, or coordinate firepower strikes, drone swarms, cyberattacks, and other twenty-first century threats. For this reason, the U.S. armed forces’ current technical edge over all potential adversaries “could be lost within the next decade if they do not accelerate the adoption of AI across their missions.” A massive AI-driven Chinese attack could overwhelm U.S. defenses.
AI systems operate on a foundation of interconnected computer hardware driven by cutting-edge semiconductor devices. “Cutting-edge” is a critical term in this context—it is almost impossible to overstate the performance disparity between advanced AI chips and conventional semiconductor central processing units (CPU). The most advanced chips are tens or even thousands of times faster than CPUs in the development of AI algorithms. “An AI chip a thousand times as efficient as a CPU provides an improvement equivalent to 26 years of Moore’s Law-driven CPU improvements,” according to Georgetown’s Center for Security and Emerging Technology. “State-of-the-art AI chips are necessary for the cost-effective, fast development, and deployment of advanced security-relevant AI systems.”
The most advanced AI systems require semiconductor chips based on 7 nm to 5 nm design rules, which are not currently manufacturable in the United States. Intel makes field-programmable gate arrays (FPGA), which are incorporated in AI systems, based on 10 nm design rules—a generation behind 7 nm. The company will begin U.S.-based production of 7 nm chips in 2022. Taiwan Semiconductor Manufacturing Co. (TSMC) is building a fab in Arizona which will operate at the 5 nm node beginning around 2024. But by this time, the state of the art is likely to have moved to 3 nm chips, all of which will be made in Taiwan.
So, China doesn't get cutting edge semiconductors. Problem solved.
Now, what else to you have...
You believe the article? What devices TSMC make? LOL
It doesn't matter what i believe. It does matter what the U.S. and the West have decided to do about it.
More to the point, Xi gets his third term as Chairman of a Party that is having to deal with a slowing economy, a collapse in real estate, and worsening relations with the rest of the world, all while still having COVID lockdowns. Sanctions of Western technology to China are the least of Xi's problems. Meanwhile, avonb7 is still attempting to defend China's growing threat to the rules of order in the West.
I think I've covered the technical and financial aspects quite clearly. The political side does not have much room here but I will add a link to where politics meets law and some scenarios that could result.
There is no point discussing more on the subject as you simply have a pre-formed opinion and ignore anything that doesn't fit into it.
From a technical perspective your understanding of current events is severely lacking.
Military hardware does not, and never has, employed the latest nodes for semiconductor usage. It is not mature enough and costs are prohibitive.
Military hardware is a combination of bespoke silicon and off-the-shelf silicon.
Procurement has always been a huge problem for military usage as hardware has a very long service life.
When using commercial off-the-shelf options, that long service life often means there are procurement issues down the line as the nodes in question are often not economically viable for outputting semiconductors for military usage.
That is a major consideration. It applies to bespoke silicon for military usage too. Ironically, commercial bespoke silicon using various nodes has seen high demand over the last few years but there is a huge difference.
Military usage does not generate the unit numbers required to make things financially viable - even on old nodes. The military simply does not order enough product. It is therefore an expensive business.
Google, Meta et tal can go that route because they can order thousands upon thousands of units.
Where does this tie in with your China claims?
China already has 14nm and 7nm capacity. The US does not have 7nm capacity. That was one of the key takeaways from the Techinsights report. It was a wake up call because no one thought it could happen. Yourself included. Much less last year (which is when they began shipping the hardware). I'm sure there was a certain amount of panic running through the corridors of US power on that news.
In military terms, yields and cost are completely irrelevant in the face of sanctions. They don't need to pump out millions of processors and have all the finances needed to pay for it!
Sanctions have zero impact on China's military ambitions. No one (well, except you, for sure) doubts that from a military perspective. No one questions that
Away from military hardware but still on military usage, AI can benefit from the latest advances but China for sure has plenty of AI capacity already. What AI needs is to be 'fed' and it gorges on data. And who is better placed to get data into those systems? It certainly isn't the US.
So, from a 'national security' perspective, the sanctions mean literally nothing.
This is about technology and the rise of China as a competitor to the US. Nothing more.
OK, as promised, the legal side:
Take note of the contents of this paper (.Pdf ) and try to imagine what the possible outcomes may be.
That's it from me. I've spelt things out to you in clear and simple terms while trying to focus on the technological side of things.
The rest is simply a question of time.
You have a weak point in your arguments. You don't provide facts to support your statements. For example, Military hardware does not, and never has, employed the latest nodes for semiconductor usage.
How do you know? You need to provide facts to make people believe you.
You are aware that Military hardware designs initiated today will be based on the best available technology, not older technology. So the U.S. has the TSMC fab coming on line in 2024 at 5nm, and that is going to be the production node for many of those components, given the security of having production in the U.S. Intel will also open its fab in Ohio to custom orders. In the meantime, companies can still source a number of different nodes from TSMC in Taiwan, including leading edge nodes.
How do I know? I manufacture RF chassis components for various electronic manufacturers that are located in my area, all of which are in the supply chain for the military. They are designing to the latest available technology knowing that all of the finished product will likely have a long life cycle. So even if they are not using the latest SOC's, or any SOC's at all, there are others in the supply chain that do require the latest generations. Most of this has to do with performance per watt, especially in misses and aircraft, where weight and bulk need to be minimized.
So China, being restricted on Western technology, is faced with limited production at 14nm, and even less production at 7nm. While some of the military hardware will use that in the future, there is no reliable timeline when anything better becomes available. China will trail behind the West and that is a massive disadvantage, of three to four generations of SOC's, ie, bigger, bulkier. hotter electronics for their missiles and aircraft, again, a disadvantage.
You remind me of one thing. The Military Requirement requires chips should work in very harsh environment. The materials TSMC uses in the fabs do not satisfy it.
About 2,470,000 results (0.49 seconds)
Search Results
Featured snippet from the web
U.S. dependency on Taiwanese production of chips for defense systems extends beyond AI. TSMC makes semiconductors used in F-35 fighters and a wide range of “military-grade” devices used by the U.S. Department of Defense (DOD)
All major U.S. defense systems and platforms rely on semiconductors for their performance. Consequently, the erosion of U.S. capabilities in microelectronics is a direct threat to the United States’ ability to defend itself and its allies. Moreover, the U.S. civilian economy is deeply dependent on semiconductor-based platforms for its daily operations. Ensuring U.S. leadership in semiconductor technology and securing the integrity of the value chains that design, manufacture, package, and distribute these chips are perhaps the preeminent economic and national security concerns of the modern era.
Of course, U.S. leaders have recognized the nexus between semiconductors and national security for decades. In the mid-1980s, President Reagan articulated the need to retain U.S. global leadership in chips to counter the Soviet bloc’s numerically superior military forces through “smart” systems driven by semiconductors—satellites, stealth aircraft, cruise missiles, and the like. Today, the United States faces an even greater challenge—it is confronted by a geopolitical rival that is stronger economically and technologically than the Soviet Union ever was. At the same time, the United States is struggling to halt and reverse the loss of its leading capabilities in manufacturing microelectronics. The recent shortage of automotive chips has driven home the damage that the steady erosion of capabilities in chip manufacturing can inflict on the U.S. civilian economy. It is equally important to appreciate the national defense implications of recent adverse trends.
The current chip legislation, which at this writing is pending reconciliation in Congress, represents the most comprehensive program of federal support ever undertaken to ensure the viability of a domestic semiconductor industry—that is, an industry whose critical functions are located within the geography of the United States. The legislation is both essential and overdue. Despite the dwindling number of critics who oppose the legislation as “industrial policy,” federal measures to ensure the existence of a robust domestic commercial semiconductor industry—as President Reagan recognized long ago—are vital to U.S. national defense and to the strength and resilience of the U.S. national economy.
Declining Onshore Capabilities
In terms of onshore capability, the U.S. semiconductor industry now has significant gaps in its production chain. The United States remains the unchallenged world leader in semiconductor design, controlling about 85 percent of the world market for electronic design automation (EDA) tools, which are necessary for the design of the most advanced chips. However, there are factors to consider, such as:
,,,
AI Challenge from China
These production chain gaps are problematic from a U.S. national security perspective. China has emerged as a major strategic challenger to the United States and is investing heavily in developing its military power and defense industrial base, placing a priority on overtaking the United States and its allies in semiconductor technology. As an independent commission established by Congress recently concluded: “If a potential adversary bests the United States in semiconductors over the long term or suddenly cuts off U.S. access to cutting-edge chips entirely, it could gain the upper hand in every domain of warfare.” U.S. vulnerability is particularly acute with respect to the most advanced chips currently in production, which are essential to the creation and application of artificial intelligence (AI)—intelligence generated by machines—which is expected to revolutionize warfare.
China’s leaders have set a goal to build a “fully modern” military by 2027 based on “informatization,” “intelligentization,” and “mechanization,” investing heavily in technical areas which support such an approach, such as AI, quantum computing, hypersonics, and microelectronics. AI enables computer systems to solve problems and address tasks that normally require human intelligence, ultimately at speeds and performance levels that vastly exceed those of humans. In the words of the Japanese National Institute for Defense Studies: “As AI does not get fatigued, does not forget, and has no emotional fluctuation, AI is expected to be able to help commanders make decisions by processing large quantities of data quickly and accurately.” In a future war, according to U.S. Senator Mike Rounds, “defending against AI-capable adversaries operating at machine speeds without employing AI is an invitation to disaster.” Human operators cannot outmatch multiple machines making thousands of decisions per second coordinated across various systems, nor will they be able to counter an adversary’s AI-enabled missile attack, strike against communications satellites and infrastructure, or coordinate firepower strikes, drone swarms, cyberattacks, and other twenty-first century threats. For this reason, the U.S. armed forces’ current technical edge over all potential adversaries “could be lost within the next decade if they do not accelerate the adoption of AI across their missions.” A massive AI-driven Chinese attack could overwhelm U.S. defenses.
AI systems operate on a foundation of interconnected computer hardware driven by cutting-edge semiconductor devices. “Cutting-edge” is a critical term in this context—it is almost impossible to overstate the performance disparity between advanced AI chips and conventional semiconductor central processing units (CPU). The most advanced chips are tens or even thousands of times faster than CPUs in the development of AI algorithms. “An AI chip a thousand times as efficient as a CPU provides an improvement equivalent to 26 years of Moore’s Law-driven CPU improvements,” according to Georgetown’s Center for Security and Emerging Technology. “State-of-the-art AI chips are necessary for the cost-effective, fast development, and deployment of advanced security-relevant AI systems.”
The most advanced AI systems require semiconductor chips based on 7 nm to 5 nm design rules, which are not currently manufacturable in the United States. Intel makes field-programmable gate arrays (FPGA), which are incorporated in AI systems, based on 10 nm design rules—a generation behind 7 nm. The company will begin U.S.-based production of 7 nm chips in 2022. Taiwan Semiconductor Manufacturing Co. (TSMC) is building a fab in Arizona which will operate at the 5 nm node beginning around 2024. But by this time, the state of the art is likely to have moved to 3 nm chips, all of which will be made in Taiwan.
So, China doesn't get cutting edge semiconductors. Problem solved.
Now, what else to you have...
I know that article very well. You left out some points that are relevant but...
Now, national security. Those snippets revolve around national security from a technology 'failure' perspective. A US technology failure. Strategic technology errors. Lack of technology funding. Moving slowly and late etc.
It saw semiconductor manufacturing move largely offshore. It failed to even see 5G and understand the implications on the world economy. It failed to give local workers the skills required to make onshore manufacturing plausible.
Every nation on earth (or every terrorist group) is a potential national security threat. Even Trump has been labeled a national security threat! That is one thing. These sanctions are another.
The sanctions have nothing to do with national security. It is about economic value intrinsically tied to high technology. Not military usage. It is all about being dominant technologically because in the digital age technology means everything and everything means money.
The US banned the export of some xeon chips to China a few years because they were planned for use in supercomputers. What happened? The following year China launched a supercomputer that went straight to the top of the world ranking using homegrown technology.
But the point is, if the processors themselves were a problem why is the US allowing China access to hundreds of millions of Qualcomm processors manufacturered using 4nm process nodes? I mean now, in 2022 with all the tension and 'national security' flag waving, there is no issue. Why? Money.
A complete shutdown of US technology in China would kill the entire US semiconductor industry in a very short time.
For military usage China already has the chops and is moving forward. As it should. As everyone with that capability does. Sanctions won't stop that progress. Can you point to anyone that doesn't think the same?
The article you quoted definitely sees China as already being a military threat. Almost all do but don't confuse that with technology, 99% of which has zero military usage.
But, I will repeat, these sanctions are not about national security or the military. They are about trying to slow Chinese influence through technology on the world stage because technology is going to be everything.
That is all!
As this article puts it, the US shouldn't being trying to trip its opponents. It should be running faster:
I think I've covered the technical and financial aspects quite clearly. The political side does not have much room here but I will add a link to where politics meets law and some scenarios that could result.
There is no point discussing more on the subject as you simply have a pre-formed opinion and ignore anything that doesn't fit into it.
From a technical perspective your understanding of current events is severely lacking.
Military hardware does not, and never has, employed the latest nodes for semiconductor usage. It is not mature enough and costs are prohibitive.
Military hardware is a combination of bespoke silicon and off-the-shelf silicon.
Procurement has always been a huge problem for military usage as hardware has a very long service life.
When using commercial off-the-shelf options, that long service life often means there are procurement issues down the line as the nodes in question are often not economically viable for outputting semiconductors for military usage.
That is a major consideration. It applies to bespoke silicon for military usage too. Ironically, commercial bespoke silicon using various nodes has seen high demand over the last few years but there is a huge difference.
Military usage does not generate the unit numbers required to make things financially viable - even on old nodes. The military simply does not order enough product. It is therefore an expensive business.
Google, Meta et tal can go that route because they can order thousands upon thousands of units.
Where does this tie in with your China claims?
China already has 14nm and 7nm capacity. The US does not have 7nm capacity. That was one of the key takeaways from the Techinsights report. It was a wake up call because no one thought it could happen. Yourself included. Much less last year (which is when they began shipping the hardware). I'm sure there was a certain amount of panic running through the corridors of US power on that news.
In military terms, yields and cost are completely irrelevant in the face of sanctions. They don't need to pump out millions of processors and have all the finances needed to pay for it!
Sanctions have zero impact on China's military ambitions. No one (well, except you, for sure) doubts that from a military perspective. No one questions that
Away from military hardware but still on military usage, AI can benefit from the latest advances but China for sure has plenty of AI capacity already. What AI needs is to be 'fed' and it gorges on data. And who is better placed to get data into those systems? It certainly isn't the US.
So, from a 'national security' perspective, the sanctions mean literally nothing.
This is about technology and the rise of China as a competitor to the US. Nothing more.
OK, as promised, the legal side:
Take note of the contents of this paper (.Pdf ) and try to imagine what the possible outcomes may be.
That's it from me. I've spelt things out to you in clear and simple terms while trying to focus on the technological side of things.
The rest is simply a question of time.
You have a weak point in your arguments. You don't provide facts to support your statements. For example, Military hardware does not, and never has, employed the latest nodes for semiconductor usage.
How do you know? You need to provide facts to make people believe you.
It should be common knowledge for anyone who has even a passing interest in processor development and to be honest I've given Tmay lots of links over the years.
Here is the first article that Google returned (published yesterday):
"The fiction of “barring military use” is only to provide the fig leaf of a cover under the World Trade Organization (WTO) exceptions on having to provide market access to all WTO members. Most military applications use older-generation chips and not the latest versions."
New process nodes bring new bugs. Software tools need to be updated etc.
You just can't rush the latest process node into your products. Often it isn't even necessary. Most of the cutting edge nodes end up in commercial products because those products can benefit from the power and power efficiencies they bring and be viable economically speaking. Cost is a major concern as some military semiconductor hardware has to be 'hardened' too. For example radiation hardened or heat hardened. Even in a commercial setting, manufacturers may decide to go with 'last year's' design in order to avoid problems. Huawei did that on one of its flagship phones when it chose to run with an older ARM design on its Soc one year. The reason given was immaturity.
Military designs are long lifespan. They get upgraded along the way but only with mature, tried and tested product components although the ultimate test is always in battle scenarios.
Typically, newer nodes bring more transistor density and power and size efficiencies. If your design can function well with current mature technology there is just no good reason to move up. That is precisely why 98% of the world's semiconductor output is nowhere near the cutting edge nodes. There is literally no need for it in IoT, industrial applications, the automotive industry etc.
As a stop gap solution, it is rumored that Huawei will bring to market a new chip stacked 14nm Kirin next year. There is nothing really new about chip stacking except that in this case the actual process is said to be much cheaper and very efficient. It should remain competitive even in 2023. And that would be 14nm.
With regards to off the shelf semiconductors, there is actually a military term for that, the name of which escapes me but it has 'off-the-shelf' in it.
Great. You posted a link to asiatimes, and we now have the Indian point of view, and you are still attempting to post as if you have a clue about the military.
For the record, India has been attempting to wean itself off of cheap Russian Military equipment, and now finds itself stranded as little of that equipment is available to export. More to the point, India stands to gain from resourcing supply chains removed from China. I'm sure the French will sell some aircraft to them, given that their home-brew program is again a fail.
Should I mention that the B-21 will be unveiled to the public on December 2? Yet another program that will benefit from 5nm.
You probably can't imagine the number of new and ongoing programs recently spurred on by the Russian invasion of Ukraine, and the ascent of Xi for yet another term, but I'll bet it's numerous.
Oh, and for the record, it appears that China doesn't get to have technology that they will ultimately use against us.
As I said, the link I provided was simply the first one Google provided.
I think I've covered the technical and financial aspects quite clearly. The political side does not have much room here but I will add a link to where politics meets law and some scenarios that could result.
There is no point discussing more on the subject as you simply have a pre-formed opinion and ignore anything that doesn't fit into it.
From a technical perspective your understanding of current events is severely lacking.
Military hardware does not, and never has, employed the latest nodes for semiconductor usage. It is not mature enough and costs are prohibitive.
Military hardware is a combination of bespoke silicon and off-the-shelf silicon.
Procurement has always been a huge problem for military usage as hardware has a very long service life.
When using commercial off-the-shelf options, that long service life often means there are procurement issues down the line as the nodes in question are often not economically viable for outputting semiconductors for military usage.
That is a major consideration. It applies to bespoke silicon for military usage too. Ironically, commercial bespoke silicon using various nodes has seen high demand over the last few years but there is a huge difference.
Military usage does not generate the unit numbers required to make things financially viable - even on old nodes. The military simply does not order enough product. It is therefore an expensive business.
Google, Meta et tal can go that route because they can order thousands upon thousands of units.
Where does this tie in with your China claims?
China already has 14nm and 7nm capacity. The US does not have 7nm capacity. That was one of the key takeaways from the Techinsights report. It was a wake up call because no one thought it could happen. Yourself included. Much less last year (which is when they began shipping the hardware). I'm sure there was a certain amount of panic running through the corridors of US power on that news.
In military terms, yields and cost are completely irrelevant in the face of sanctions. They don't need to pump out millions of processors and have all the finances needed to pay for it!
Sanctions have zero impact on China's military ambitions. No one (well, except you, for sure) doubts that from a military perspective. No one questions that
Away from military hardware but still on military usage, AI can benefit from the latest advances but China for sure has plenty of AI capacity already. What AI needs is to be 'fed' and it gorges on data. And who is better placed to get data into those systems? It certainly isn't the US.
So, from a 'national security' perspective, the sanctions mean literally nothing.
This is about technology and the rise of China as a competitor to the US. Nothing more.
OK, as promised, the legal side:
Take note of the contents of this paper (.Pdf ) and try to imagine what the possible outcomes may be.
That's it from me. I've spelt things out to you in clear and simple terms while trying to focus on the technological side of things.
The rest is simply a question of time.
You have a weak point in your arguments. You don't provide facts to support your statements. For example, Military hardware does not, and never has, employed the latest nodes for semiconductor usage.
How do you know? You need to provide facts to make people believe you.
You are aware that Military hardware designs initiated today will be based on the best available technology, not older technology. So the U.S. has the TSMC fab coming on line in 2024 at 5nm, and that is going to be the production node for many of those components, given the security of having production in the U.S. Intel will also open its fab in Ohio to custom orders. In the meantime, companies can still source a number of different nodes from TSMC in Taiwan, including leading edge nodes.
How do I know? I manufacture RF chassis components for various electronic manufacturers that are located in my area, all of which are in the supply chain for the military. They are designing to the latest available technology knowing that all of the finished product will likely have a long life cycle. So even if they are not using the latest SOC's, or any SOC's at all, there are others in the supply chain that do require the latest generations. Most of this has to do with performance per watt, especially in misses and aircraft, where weight and bulk need to be minimized.
So China, being restricted on Western technology, is faced with limited production at 14nm, and even less production at 7nm. While some of the military hardware will use that in the future, there is no reliable timeline when anything better becomes available. China will trail behind the West and that is a massive disadvantage, of three to four generations of SOC's, ie, bigger, bulkier. hotter electronics for their missiles and aircraft, again, a disadvantage.
You remind me of one thing. The Military Requirement requires chips should work in very harsh environment. The materials TSMC uses in the fabs do not satisfy it.
About 2,470,000 results (0.49 seconds)
Search Results
Featured snippet from the web
U.S. dependency on Taiwanese production of chips for defense systems extends beyond AI. TSMC makes semiconductors used in F-35 fighters and a wide range of “military-grade” devices used by the U.S. Department of Defense (DOD)
All major U.S. defense systems and platforms rely on semiconductors for their performance. Consequently, the erosion of U.S. capabilities in microelectronics is a direct threat to the United States’ ability to defend itself and its allies. Moreover, the U.S. civilian economy is deeply dependent on semiconductor-based platforms for its daily operations. Ensuring U.S. leadership in semiconductor technology and securing the integrity of the value chains that design, manufacture, package, and distribute these chips are perhaps the preeminent economic and national security concerns of the modern era.
Of course, U.S. leaders have recognized the nexus between semiconductors and national security for decades. In the mid-1980s, President Reagan articulated the need to retain U.S. global leadership in chips to counter the Soviet bloc’s numerically superior military forces through “smart” systems driven by semiconductors—satellites, stealth aircraft, cruise missiles, and the like. Today, the United States faces an even greater challenge—it is confronted by a geopolitical rival that is stronger economically and technologically than the Soviet Union ever was. At the same time, the United States is struggling to halt and reverse the loss of its leading capabilities in manufacturing microelectronics. The recent shortage of automotive chips has driven home the damage that the steady erosion of capabilities in chip manufacturing can inflict on the U.S. civilian economy. It is equally important to appreciate the national defense implications of recent adverse trends.
The current chip legislation, which at this writing is pending reconciliation in Congress, represents the most comprehensive program of federal support ever undertaken to ensure the viability of a domestic semiconductor industry—that is, an industry whose critical functions are located within the geography of the United States. The legislation is both essential and overdue. Despite the dwindling number of critics who oppose the legislation as “industrial policy,” federal measures to ensure the existence of a robust domestic commercial semiconductor industry—as President Reagan recognized long ago—are vital to U.S. national defense and to the strength and resilience of the U.S. national economy.
Declining Onshore Capabilities
In terms of onshore capability, the U.S. semiconductor industry now has significant gaps in its production chain. The United States remains the unchallenged world leader in semiconductor design, controlling about 85 percent of the world market for electronic design automation (EDA) tools, which are necessary for the design of the most advanced chips. However, there are factors to consider, such as:
,,,
AI Challenge from China
These production chain gaps are problematic from a U.S. national security perspective. China has emerged as a major strategic challenger to the United States and is investing heavily in developing its military power and defense industrial base, placing a priority on overtaking the United States and its allies in semiconductor technology. As an independent commission established by Congress recently concluded: “If a potential adversary bests the United States in semiconductors over the long term or suddenly cuts off U.S. access to cutting-edge chips entirely, it could gain the upper hand in every domain of warfare.” U.S. vulnerability is particularly acute with respect to the most advanced chips currently in production, which are essential to the creation and application of artificial intelligence (AI)—intelligence generated by machines—which is expected to revolutionize warfare.
China’s leaders have set a goal to build a “fully modern” military by 2027 based on “informatization,” “intelligentization,” and “mechanization,” investing heavily in technical areas which support such an approach, such as AI, quantum computing, hypersonics, and microelectronics. AI enables computer systems to solve problems and address tasks that normally require human intelligence, ultimately at speeds and performance levels that vastly exceed those of humans. In the words of the Japanese National Institute for Defense Studies: “As AI does not get fatigued, does not forget, and has no emotional fluctuation, AI is expected to be able to help commanders make decisions by processing large quantities of data quickly and accurately.” In a future war, according to U.S. Senator Mike Rounds, “defending against AI-capable adversaries operating at machine speeds without employing AI is an invitation to disaster.” Human operators cannot outmatch multiple machines making thousands of decisions per second coordinated across various systems, nor will they be able to counter an adversary’s AI-enabled missile attack, strike against communications satellites and infrastructure, or coordinate firepower strikes, drone swarms, cyberattacks, and other twenty-first century threats. For this reason, the U.S. armed forces’ current technical edge over all potential adversaries “could be lost within the next decade if they do not accelerate the adoption of AI across their missions.” A massive AI-driven Chinese attack could overwhelm U.S. defenses.
AI systems operate on a foundation of interconnected computer hardware driven by cutting-edge semiconductor devices. “Cutting-edge” is a critical term in this context—it is almost impossible to overstate the performance disparity between advanced AI chips and conventional semiconductor central processing units (CPU). The most advanced chips are tens or even thousands of times faster than CPUs in the development of AI algorithms. “An AI chip a thousand times as efficient as a CPU provides an improvement equivalent to 26 years of Moore’s Law-driven CPU improvements,” according to Georgetown’s Center for Security and Emerging Technology. “State-of-the-art AI chips are necessary for the cost-effective, fast development, and deployment of advanced security-relevant AI systems.”
The most advanced AI systems require semiconductor chips based on 7 nm to 5 nm design rules, which are not currently manufacturable in the United States. Intel makes field-programmable gate arrays (FPGA), which are incorporated in AI systems, based on 10 nm design rules—a generation behind 7 nm. The company will begin U.S.-based production of 7 nm chips in 2022. Taiwan Semiconductor Manufacturing Co. (TSMC) is building a fab in Arizona which will operate at the 5 nm node beginning around 2024. But by this time, the state of the art is likely to have moved to 3 nm chips, all of which will be made in Taiwan.
So, China doesn't get cutting edge semiconductors. Problem solved.
Now, what else to you have...
I know that article very well. You left out some points that are relevant but...
Now, national security. Those snippets revolve around national security from a technology 'failure' perspective. A US technology failure. Strategic technology errors. Lack of technology funding. Moving slowly and late etc.
It saw semiconductor manufacturing move largely offshore. It failed to even see 5G and understand the implications on the world economy. It failed to give local workers the skills required to make onshore manufacturing plausible.
Every nation on earth (or every terrorist group) is a potential national security threat. Even Trump has been labeled a national security threat! That is one thing. These sanctions are another.
The sanctions have nothing to do with national security. It is about economic value intrinsically tied to high technology. Not military usage. It is all about being dominant technologically because in the digital age technology means everything and everything means money.
The US banned the export of some xeon chips to China a few years because they were planned for use in supercomputers. What happened? The following year China launched a supercomputer that went straight to the top of the world ranking using homegrown technology.
But the point is, if the processors themselves were a problem why is the US allowing China access to hundreds of millions of Qualcomm processors manufacturered using 4nm process nodes? I mean now, in 2022 with all the tension and 'national security' flag waving, there is no issue. Why? Money.
A complete shutdown of US technology in China would kill the entire US semiconductor industry in a very short time.
For military usage China already has the chops and is moving forward. As it should. As everyone with that capability does. Sanctions won't stop that progress. Can you point to anyone that doesn't think the same?
The article you quoted definitely sees China as already being a military threat. Almost all do but don't confuse that with technology, 99% of which has zero military usage.
But, I will repeat, these sanctions are not about national security or the military. They are about trying to slow Chinese influence through technology on the world stage because technology is going to be everything.
That is all!
As this article puts it, the US shouldn't being trying to trip its opponents. It should be running faster:
Lot's of interesting jousting. Back to the beginning re existential threat US faces w/o AP foundries.
TSMC isn't wrong...it's about the magnitude of engineering human capital required to establish, maintain, and grow the semiconductor industry in the US. The US can throw a ton of money at brick & mortar, but without an almost equal investment in education, training, and establishing true long-lasting engineering competence, these semiconductor fabs will become hollow palaces producing for the handful of US companies still standing. In the short to medium term, we will have to spend real money to entice engineers from Taiwan, Korea, and China to come to the US for as long as it takes to rebuild the semiconductor engineering ecosystem. This is why we should be paying TSMC and others to build many leading edge technology fabs in the US...at whatever the cost. One TSMC fab @5nm is a drop in the ocean...
Lot's of interesting jousting. Back to the beginning re existential threat US faces w/o AP foundries.
TSMC isn't wrong...it's about the magnitude of engineering human capital required to establish, maintain, and grow the semiconductor industry in the US. The US can throw a ton of money at brick & mortar, but without an almost equal investment in education, training, and establishing true long-lasting engineering competence, these semiconductor fabs will become hollow palaces producing for the handful of US companies still standing. In the short to medium term, we will have to spend real money to entice engineers from Taiwan, Korea, and China to come to the US for as long as it takes to rebuild the semiconductor engineering ecosystem. This is why we should be paying TSMC and others to build many leading edge technology fabs in the US...at whatever the cost. One TSMC fab @5nm is a drop in the ocean...
Do you know US population is fifteen times more than Taiwan? US engineers cost more than Taiwan just like the workers.
Lot's of interesting jousting. Back to the beginning re existential threat US faces w/o AP foundries.
TSMC isn't wrong...it's about the magnitude of engineering human capital required to establish, maintain, and grow the semiconductor industry in the US. The US can throw a ton of money at brick & mortar, but without an almost equal investment in education, training, and establishing true long-lasting engineering competence, these semiconductor fabs will become hollow palaces producing for the handful of US companies still standing. In the short to medium term, we will have to spend real money to entice engineers from Taiwan, Korea, and China to come to the US for as long as it takes to rebuild the semiconductor engineering ecosystem. This is why we should be paying TSMC and others to build many leading edge technology fabs in the US...at whatever the cost. One TSMC fab @5nm is a drop in the ocean...
Do you know US population is fifteen times more than Taiwan? US engineers cost more than Taiwan just like the workers.
Don’t forget that population is only relevant if the education and training options are in place.
Also, like the City in London, industries tend to attract talent and peripheral players to their locations. Cork in Ireland is another example.
Taiwan has become a foundry hub that feeds hundreds or thousands of smaller peripheral companies. Its proximity to the Chinese mainland (the world's largest consumer of chips) is also a huge advantage.
As of today industry watchers seem to agree that the US does not have the required skilled labor to satisfy some of the recently announced plans.
Lot's of interesting jousting. Back to the beginning re existential threat US faces w/o AP foundries.
TSMC isn't wrong...it's about the magnitude of engineering human capital required to establish, maintain, and grow the semiconductor industry in the US. The US can throw a ton of money at brick & mortar, but without an almost equal investment in education, training, and establishing true long-lasting engineering competence, these semiconductor fabs will become hollow palaces producing for the handful of US companies still standing. In the short to medium term, we will have to spend real money to entice engineers from Taiwan, Korea, and China to come to the US for as long as it takes to rebuild the semiconductor engineering ecosystem. This is why we should be paying TSMC and others to build many leading edge technology fabs in the US...at whatever the cost. One TSMC fab @5nm is a drop in the ocean...
Do you know US population is fifteen times more than Taiwan? US engineers cost more than Taiwan just like the workers.
Don’t forget that population is only relevant if the education and training options are in place.
Also, like the City in London, industries tend to attract talent and peripheral players to their locations. Cork in Ireland is another example.
Taiwan has become a foundry hub that feeds hundreds or thousands of smaller peripheral companies. Its proximity to the Chinese mainland (the world's largest consumer of chips) is also a huge advantage.
As of today industry watchers seem to agree that the US does not have the required skilled labor to satisfy some of the recently announced plans.
TSMC is building another Fab in Arizona. So it will be more expensive to train skilled labor, but at the same time, that's what the University of Arizona and Stanford have been producing for Intel for decades.
Taiwan's chipmaking giant TSMC is said to be preparing to build another semiconductor fabrication plant in Arizona, alongside the facility it completed this summer, in a move that may be seen as a vindication of the US government’s CHIPS Act funding.
According to reports in the Wall Street Journal, TSMC is planning to announce in the near future that it will build a further factory for making cutting edge chips at a site just north of Phoenix, adjacent to the $12 billion Fab 21 plant the company decided to construct in 2020.
The new facility will be used to manufacture 3nm chips, according to the paper, which cites anonymous sources "familiar with the expansion plans." The scale of this project is expected to be comparable to the existing plant.
You still seem to be carrying water for China, but, you might consider carrying a little bit for the EU, which is attempting to become more relevant in semiconductors.
We have calculated that we would need €500 billion investment in Europe to reach the 20 percent market share goal formulated in the EU Chips Act," the boss of NXP, which makes NFC chips for many of the contactless transport cards in Europe, is quoted as saying.
Reaching 20 percent of the global semiconductor market from today's estimated 10 percent would require a tripling or quadrupling of the chip manufacturing capacity represented among EU countries, a daunting task, especially given the ambitious expansion plans of chipmakers such as TSMC elsewhere in the world.
Andrew Buss, research director for European Enterprise Infrastructure at IDC, said he agreed the current level of funding on offer seemed a little on the low side, but that in an undertaking of this sort it was important to start somewhere.
"Once it does start, you soon see things start to happen," he said, mentioning the American government's US CHIPS Act, which has already proved successful in starting the ball rolling on an ambitious program to overhaul its own semiconductor industry.
However, building up a credible world-class semiconductor industry requires more than just pouring money into the sector, he adds, calling for careful planning to deliver the skills needed to design advanced semiconductors as well as develop Europe's own manufacturing processes instead of relying on companies like Intel and TSMC to build chip fabrication plants here, as is happening with Intel at Magdeburg in Germany.
You keep dissing the U.S. approach to expanding semiconductor production in North America, but it keeps on expanding, and as for skilled labor, good thing that the U.S. allows immigration, and that people actually want to come and live here.
Lot's of interesting jousting. Back to the beginning re existential threat US faces w/o AP foundries.
TSMC isn't wrong...it's about the magnitude of engineering human capital required to establish, maintain, and grow the semiconductor industry in the US. The US can throw a ton of money at brick & mortar, but without an almost equal investment in education, training, and establishing true long-lasting engineering competence, these semiconductor fabs will become hollow palaces producing for the handful of US companies still standing. In the short to medium term, we will have to spend real money to entice engineers from Taiwan, Korea, and China to come to the US for as long as it takes to rebuild the semiconductor engineering ecosystem. This is why we should be paying TSMC and others to build many leading edge technology fabs in the US...at whatever the cost. One TSMC fab @5nm is a drop in the ocean...
Do you know US population is fifteen times more than Taiwan? US engineers cost more than Taiwan just like the workers.
Don’t forget that population is only relevant if the education and training options are in place.
Also, like the City in London, industries tend to attract talent and peripheral players to their locations. Cork in Ireland is another example.
Taiwan has become a foundry hub that feeds hundreds or thousands of smaller peripheral companies. Its proximity to the Chinese mainland (the world's largest consumer of chips) is also a huge advantage.
As of today industry watchers seem to agree that the US does not have the required skilled labor to satisfy some of the recently announced plans.
The European Commission on Thursday urged EU member countries to reduce the risks associated with Chinese telecoms equipment in 5G networks, naming Germany in particular as a country that needs to implement the bloc's joint 5G security guidelines.
"We are urging member states who have not yet imposed restrictions on high-risk suppliers to do that without delay, as a matter of urgency," said Margrethe Vestager, executive vice president of the Commission in charge of digital issues.
"A number of countries have passed legislation but they have not put it into effect ... Making it work is even better," she said.
Lot's of interesting jousting. Back to the beginning re existential threat US faces w/o AP foundries.
TSMC isn't wrong...it's about the magnitude of engineering human capital required to establish, maintain, and grow the semiconductor industry in the US. The US can throw a ton of money at brick & mortar, but without an almost equal investment in education, training, and establishing true long-lasting engineering competence, these semiconductor fabs will become hollow palaces producing for the handful of US companies still standing. In the short to medium term, we will have to spend real money to entice engineers from Taiwan, Korea, and China to come to the US for as long as it takes to rebuild the semiconductor engineering ecosystem. This is why we should be paying TSMC and others to build many leading edge technology fabs in the US...at whatever the cost. One TSMC fab @5nm is a drop in the ocean...
Do you know US population is fifteen times more than Taiwan? US engineers cost more than Taiwan just like the workers.
Don’t forget that population is only relevant if the education and training options are in place.
Also, like the City in London, industries tend to attract talent and peripheral players to their locations. Cork in Ireland is another example.
Taiwan has become a foundry hub that feeds hundreds or thousands of smaller peripheral companies. Its proximity to the Chinese mainland (the world's largest consumer of chips) is also a huge advantage.
As of today industry watchers seem to agree that the US does not have the required skilled labor to satisfy some of the recently announced plans.
TSMC is building another Fab in Arizona. So it will be more expensive to train skilled labor, but at the same time, that's what the University of Arizona and Stanford have been producing for Intel for decades.
Taiwan's chipmaking giant TSMC is said to be preparing to build another semiconductor fabrication plant in Arizona, alongside the facility it completed this summer, in a move that may be seen as a vindication of the US government’s CHIPS Act funding.
According to reports in the Wall Street Journal, TSMC is planning to announce in the near future that it will build a further factory for making cutting edge chips at a site just north of Phoenix, adjacent to the $12 billion Fab 21 plant the company decided to construct in 2020.
The new facility will be used to manufacture 3nm chips, according to the paper, which cites anonymous sources "familiar with the expansion plans." The scale of this project is expected to be comparable to the existing plant.
You still seem to be carrying water for China, but, you might consider carrying a little bit for the EU, which is attempting to become more relevant in semiconductors.
We have calculated that we would need €500 billion investment in Europe to reach the 20 percent market share goal formulated in the EU Chips Act," the boss of NXP, which makes NFC chips for many of the contactless transport cards in Europe, is quoted as saying.
Reaching 20 percent of the global semiconductor market from today's estimated 10 percent would require a tripling or quadrupling of the chip manufacturing capacity represented among EU countries, a daunting task, especially given the ambitious expansion plans of chipmakers such as TSMC elsewhere in the world.
Andrew Buss, research director for European Enterprise Infrastructure at IDC, said he agreed the current level of funding on offer seemed a little on the low side, but that in an undertaking of this sort it was important to start somewhere.
"Once it does start, you soon see things start to happen," he said, mentioning the American government's US CHIPS Act, which has already proved successful in starting the ball rolling on an ambitious program to overhaul its own semiconductor industry.
However, building up a credible world-class semiconductor industry requires more than just pouring money into the sector, he adds, calling for careful planning to deliver the skills needed to design advanced semiconductors as well as develop Europe's own manufacturing processes instead of relying on companies like Intel and TSMC to build chip fabrication plants here, as is happening with Intel at Magdeburg in Germany.
You keep dissing the U.S. approach to expanding semiconductor production in North America, but it keeps on expanding, and as for skilled labor, good thing that the U.S. allows immigration, and that people actually want to come and live here.
Cutting edge, again?
I have asked multiple times now. How much of the industry is made up of those chips?
Virtually nothing, right?
The world is dependent on mature process technology. Very, very dependent. You might have noticed how things wobbled across the globe with the shortage of chips that were far, far removed from the cutting edge.
The CHIPS act is necessary to get a smidgen of manufacturing capacity into the US. Without it the US would be fated to rely on technology from other geographical regions and the global supply chain.
The same applies to the EU.
China obviously has its own plans, too.
Both the EU and China (and India by the way) have plans to reduce dependence on US sourced technology. That started long before the US tried to wreck the global supply chain for semiconductor components.
Those plans, according to various reports, were obviously accelerated in the case of China but now industry watchers are saying that many companies began changing their designs and seeking non-US technology back in 2019. It is even reported that some US companies set up overseas business in order to skirt restrictions. We have yet to see the brunt of those changes but even you should be able to understand that nobody likes their business interests clobbered by extraterritorial, unilateral sanctions. The US has made it semiconductor industry toxic.
If you take a look at the CHIPS act you will also find that it is woefully lacking in funding once everything is broken down. Everyone agrees on that. No one knows how funding will progress in the future either.
Obviously, any funding is better than no funding and the same applies to the EU. However, the EU plan was laid out years ago and is on track. China's too. The US plan was hastily cobbled together when it realised how reliant it was on those offshore fabs and how fast China was progressing.
It is only the US that has more to lose. It has already lost. There is no getting around that. Everyone has something to gain.
It cut off revenues to its own tech industry with the China measures. Those same measures saw market caps of those companies lose over a trillion dollars. Those measures sent that lost business into the hands of competitors.
That all leads to a reduction in funding for R&D going forward.
It will also see less trade with the EU when the EU efforts take off. The EU processor initiative focused on HPC design first but will widen its scope over time.
From dominating the design aspects of the industry and controlling certain key manufacturing areas, it will now have to adjust to a new reality and intense competition from basically, well, everywhere.
"Meanwhile, according to The Economist, Trump’s «success» had a negative side. The Republican Administration overtly ignored the interests of the allies and partners. As a result, western investors began to invest in businesses and supply chains of components that were exempt from the control of the American supervisory bodies. Japanese companies offered a full range of electronic components as produce that was free from technological restrictions imposed by the United States"
Comments
You really do compartmentalize your worldview to make it sync with China. Sad.
https://www.defense.gov/News/News-Stories/Article/Article/3123699/chips-act-advances-dods-emphasis-on-microelectronics/
Military hardware does not, and never has, employed the latest nodes for semiconductor usage.
How do you know? You need to provide facts to make people believe you.
How do I know? I manufacture RF chassis components for various electronic manufacturers that are located in my area, all of which are in the supply chain for the military. They are designing to the latest available technology knowing that all of the finished product will likely have a long life cycle. So even if they are not using the latest SOC's, or any SOC's at all, there are others in the supply chain that do require the latest generations. Most of this has to do with performance per watt, especially in misses and aircraft, where weight and bulk need to be minimized.
So China, being restricted on Western technology, is faced with limited production at 14nm, and even less production at 7nm. While some of the military hardware will use that in the future, there is no reliable timeline when anything better becomes available. China will trail behind the West and that is a massive disadvantage, of three to four generations of SOC's, ie, bigger, bulkier. hotter electronics for their missiles and aircraft, again, a disadvantage.
See the following for why this is an advantage.
"John Boyd, father of the F-16"
https://www.amazon.com/Boyd-Fighter-Pilot-Who-Changed/dp/0316796883
https://en.wikipedia.org/wiki/OODA_loop
https://thedecisionlab.com/reference-guide/computer-science/the-ooda-loop
Here is the first article that Google returned (published yesterday):
"The fiction of “barring military use” is only to provide the fig leaf of a cover under the World Trade Organization (WTO) exceptions on having to provide market access to all WTO members. Most military applications use older-generation chips and not the latest versions."
https://asiatimes.com/2022/10/us-chip-ban-de-facto-declaration-of-war-on-china/
New process nodes bring new bugs. Software tools need to be updated etc.
You just can't rush the latest process node into your products. Often it isn't even necessary. Most of the cutting edge nodes end up in commercial products because those products can benefit from the power and power efficiencies they bring and be viable economically speaking. Cost is a major concern as some military semiconductor hardware has to be 'hardened' too. For example radiation hardened or heat hardened. Even in a commercial setting, manufacturers may decide to go with 'last year's' design in order to avoid problems. Huawei did that on one of its flagship phones when it chose to run with an older ARM design on its Soc one year. The reason given was immaturity.
Military designs are long lifespan. They get upgraded along the way but only with mature, tried and tested product components although the ultimate test is always in battle scenarios.
Typically, newer nodes bring more transistor density and power and size efficiencies. If your design can function well with current mature technology there is just no good reason to move up. That is precisely why 98% of the world's semiconductor output is nowhere near the cutting edge nodes. There is literally no need for it in IoT, industrial applications, the automotive industry etc.
As a stop gap solution, it is rumored that Huawei will bring to market a new chip stacked 14nm Kirin next year. There is nothing really new about chip stacking except that in this case the actual process is said to be much cheaper and very efficient. It should remain competitive even in 2023. And that would be 14nm.
With regards to off the shelf semiconductors, there is actually a military term for that, the name of which escapes me but it has 'off-the-shelf' in it.
For the record, India has been attempting to wean itself off of cheap Russian Military equipment, and now finds itself stranded as little of that equipment is available to export. More to the point, India stands to gain from resourcing supply chains removed from China. I'm sure the French will sell some aircraft to them, given that their home-brew program is again a fail.
Oh, and BTW, the F-35 will likely be using 5nm technology, as it is upgraded to Block 4 through to the end of the decade, followed by the NGAD.
https://breakingdefense.com/2022/04/f-35-modernization-programs-costs-schedule-keeps-growing-gao/
https://www.defensenews.com/air/2022/06/01/the-air-forces-next-gen-fighter-has-moved-into-a-critical-new-phase/
Should I mention that the B-21 will be unveiled to the public on December 2? Yet another program that will benefit from 5nm.
You probably can't imagine the number of new and ongoing programs recently spurred on by the Russian invasion of Ukraine, and the ascent of Xi for yet another term, but I'll bet it's numerous.
Oh, and for the record, it appears that China doesn't get to have technology that they will ultimately use against us.
https://www.csis.org/analysis/semiconductors-and-national-defense-what-are-stakes
So, China doesn't get cutting edge semiconductors. Problem solved.
Now, what else to you have...
More to the point, Xi gets his third term as Chairman of a Party that is having to deal with a slowing economy, a collapse in real estate, and worsening relations with the rest of the world, all while still having COVID lockdowns. Sanctions of Western technology to China are the least of Xi's problems. Meanwhile, avonb7 is still attempting to defend China's growing threat to the rules of order in the West.
Is that what you were aiming for?
Now, national security. Those snippets revolve around national security from a technology 'failure' perspective. A US technology failure. Strategic technology errors. Lack of technology funding. Moving slowly and late etc.
It saw semiconductor manufacturing move largely offshore. It failed to even see 5G and understand the implications on the world economy. It failed to give local workers the skills required to make onshore manufacturing plausible.
Every nation on earth (or every terrorist group) is a potential national security threat. Even Trump has been labeled a national security threat! That is one thing. These sanctions are another.
The sanctions have nothing to do with national security. It is about economic value intrinsically tied to high technology. Not military usage. It is all about being dominant technologically because in the digital age technology means everything and everything means money.
The US banned the export of some xeon chips to China a few years because they were planned for use in supercomputers. What happened? The following year China launched a supercomputer that went straight to the top of the world ranking using homegrown technology.
But the point is, if the processors themselves were a problem why is the US allowing China access to hundreds of millions of Qualcomm processors manufacturered using 4nm process nodes? I mean now, in 2022 with all the tension and 'national security' flag waving, there is no issue. Why? Money.
A complete shutdown of US technology in China would kill the entire US semiconductor industry in a very short time.
For military usage China already has the chops and is moving forward. As it should. As everyone with that capability does. Sanctions won't stop that progress. Can you point to anyone that doesn't think the same?
The article you quoted definitely sees China as already being a military threat. Almost all do but don't confuse that with technology, 99% of which has zero military usage.
But, I will repeat, these sanctions are not about national security or the military. They are about trying to slow Chinese influence through technology on the world stage because technology is going to be everything.
That is all!
As this article puts it, the US shouldn't being trying to trip its opponents. It should be running faster:
https://fortune.com/2022/09/28/chip-export-ban-china-us-asml-nvidia-rakesh-kumar/
Your link just serves to prove my point.
https://nymag.com/intelligencer/2022/10/on-with-kara-swisher-nouriel-roubini-on-the-threat-of-china.html
TSMC isn't wrong...it's about the magnitude of engineering human capital required to establish, maintain, and grow the semiconductor industry in the US. The US can throw a ton of money at brick & mortar, but without an almost equal investment in education, training, and establishing true long-lasting engineering competence, these semiconductor fabs will become hollow palaces producing for the handful of US companies still standing. In the short to medium term, we will have to spend real money to entice engineers from Taiwan, Korea, and China to come to the US for as long as it takes to rebuild the semiconductor engineering ecosystem. This is why we should be paying TSMC and others to build many leading edge technology fabs in the US...at whatever the cost. One TSMC fab @5nm is a drop in the ocean...
Also, like the City in London, industries tend to attract talent and peripheral players to their locations. Cork in Ireland is another example.
Taiwan has become a foundry hub that feeds hundreds or thousands of smaller peripheral companies. Its proximity to the Chinese mainland (the world's largest consumer of chips) is also a huge advantage.
As of today industry watchers seem to agree that the US does not have the required skilled labor to satisfy some of the recently announced plans.
Meanwhile,
https://www.theregister.com/2022/11/09/tsmc_arizona_fab_report/
You still seem to be carrying water for China, but, you might consider carrying a little bit for the EU, which is attempting to become more relevant in semiconductors.
https://www.theregister.com/2022/10/03/eu_needs_more_semiconductor_investment_nxp/
You keep dissing the U.S. approach to expanding semiconductor production in North America, but it keeps on expanding, and as for skilled labor, good thing that the U.S. allows immigration, and that people actually want to come and live here.
https://www.politico.eu/article/eu-nudges-germany-to-cut-down-on-huawei/
I have asked multiple times now. How much of the industry is made up of those chips?
Virtually nothing, right?
The world is dependent on mature process technology. Very, very dependent. You might have noticed how things wobbled across the globe with the shortage of chips that were far, far removed from the cutting edge.
The CHIPS act is necessary to get a smidgen of manufacturing capacity into the US. Without it the US would be fated to rely on technology from other geographical regions and the global supply chain.
The same applies to the EU.
China obviously has its own plans, too.
Both the EU and China (and India by the way) have plans to reduce dependence on US sourced technology. That started long before the US tried to wreck the global supply chain for semiconductor components.
Those plans, according to various reports, were obviously accelerated in the case of China but now industry watchers are saying that many companies began changing their designs and seeking non-US technology back in 2019. It is even reported that some US companies set up overseas business in order to skirt restrictions. We have yet to see the brunt of those changes but even you should be able to understand that nobody likes their business interests clobbered by extraterritorial, unilateral sanctions. The US has made it semiconductor industry toxic.
If you take a look at the CHIPS act you will also find that it is woefully lacking in funding once everything is broken down. Everyone agrees on that. No one knows how funding will progress in the future either.
Obviously, any funding is better than no funding and the same applies to the EU. However, the EU plan was laid out years ago and is on track. China's too. The US plan was hastily cobbled together when it realised how reliant it was on those offshore fabs and how fast China was progressing.
It is only the US that has more to lose. It has already lost. There is no getting around that. Everyone has something to gain.
It cut off revenues to its own tech industry with the China measures. Those same measures saw market caps of those companies lose over a trillion dollars. Those measures sent that lost business into the hands of competitors.
That all leads to a reduction in funding for R&D going forward.
It will also see less trade with the EU when the EU efforts take off. The EU processor initiative focused on HPC design first but will widen its scope over time.
From dominating the design aspects of the industry and controlling certain key manufacturing areas, it will now have to adjust to a new reality and intense competition from basically, well, everywhere.
"Meanwhile, according to The Economist, Trump’s «success» had a negative side. The Republican Administration overtly ignored the interests of the allies and partners. As a result, western investors began to invest in businesses and supply chains of components that were exempt from the control of the American supervisory bodies. Japanese companies offered a full range of electronic components as produce that was free from technological restrictions imposed by the United States"
https://moderndiplomacy.eu/2022/11/06/chip-war-against-china-threatens-to-undermine-america/