Book Event: Chris Miller's "Chip War: The Fight for the World's Most Critical Technology"

welcome to csis I'm sujay Shiva Kumar senior fellow and director of the renewing American innovation project here at csis securing the the future of the semiconductor industry is vital to renewing American innovation this is why I'm particularly delighted to welcome Professor Chris Miller author of the Chip War I'm equally pleased to welcome our interviewer professor Charles Wesner of Georgetown University and a senior advisor to csis he is also an expert on this ON Semiconductor policy I will now turn to Dr wessner to more fully introduce and welcome our guest good morning I'm very pleased to introduce Chris Miller who's an associate professor of international history where his research focuses on technology geopolitics economics International Affairs and Russia in short everything we think about on a on an international basis most importantly he is author of the Chip War the fight for the world's most critical technology a geopolitical history of the computer chip and the current state of the industry and the international competition for it he is also the gene Kirkpatrick visiting fellow at the American Enterprise Institute and the Eurasia director at the foreign policy Research Institute Chris welcome to csis thank you for having me I think you should be congratulated on this outstanding work uh your book is a pleasure to read it's easy to grasp and at the same time you address a wide variety of very complex policy issues it's a it's really a very refreshing read and as we were saying it couldn't be more timely uh my first question is just a collegial one being a Fletcher graduate myself but what motivated you as a professor of international history at the Fletcher School and an expert on Russian and Soviet economic history to write a book about computer chips well I didn't actually start out intending to write a book about ships I realized over the course of my research just how important semiconductors are my initial question was about the evolution of military technology during the Cold War on the puzzle that interests me is why was it that in the early Cold War both the U.S and the Soviet Union could produce the key Technologies Atomic weapons long-range Delivery Systems but by the end of the Cold War The Cutting Edge of military technology was something the U.S could do but the Soviets couldn't and you see this really visibly in the 1991 Persian Gulf War where the U.S began to deploy Precision strike capabilities that were far beyond what the Soviet Union could do itself and and I want to understand well why was this because the Soviet Union had a lot of capabilities that you might think would be core developing technology brilliant physicists a lot of capital investment but it became clear to me that the key transformative force in military power over the period of the Cold War was the application of computing to the military systems and to understand where computing power comes from because to start with semiconductors and it's the production of semiconductors their design and their proliferation throughout society that is explained not only the shifts of military power but I also came to realize the broader economic and technological importance of ships and I was someone who had taken advantage of chips in my smartphones and PCs but never really thought about how important they were and over the course of my research I realized that the really interesting question wasn't the shift in military technology per se but rather how chips had transformed economy and Society in ways that we were only really dimly aware of even though we relied on them for all aspects of everyday life well that's great I wish more people understood what you what you came to understand um you know you've referred and I think rightly to the some of the battles in the Reagan America administration of computer chips potato chips you know what's the difference you treat that very very effectively in fact why don't you make a quick reference why why what's the difference between potato chips and computer chips well one of the things that most people don't realize is that the the process of Designing and Manufacturing ships is the most complex technological process humans have ever undertaken Bar None the supply chain needed to produce an advanced ship stretches across multiple continents it involves some of the most purified materials some of the most precise Machine Tools ever made so there's really nothing harder to produce than an advanced chip which is why their production is both very difficult but also very important when it comes to which companies and countries control that process and that's why it's not simply a question of business or economics or technology but also a question of really top level political relevances to which countries have these capabilities and which don't but well said in fact an add-on to you mentioned the first Gulf War and the shock that many many of our opponents had to see the deployment of these precision but it's interesting um you can understand why the Soviet Union could not do that um why would you think that the Putin regime if we can call it that uh still can't well if you look at the world today most countries can't make advanced chips in fact the capability to make Cutting Edge chips is really in the hands of just a couple of countries and even many advanced economies don't have any capacity domestically to make an advanced semiconductor so it's not just Russia or um or China that can't make advanced ships just a couple of companies can and the reason is is twofold first off there's just a ton of precision knowledge involved in producing semiconductors and the tools that are used to make them in the materials that can't be simply replicated it's built up in companies over years if not decades and there's just no way to copy and paste that into a new organization and secondly it's brutally expensive right an advanced shipping facility is one of the most expensive factories in human history and the amount of capital investment across the supply chain to make an advanced ship is extraordinary which is why no country can go it alone and most countries can't afford to get anywhere close to The Cutting Edge that's that's interesting and and I couldn't agree with you more and yet there have been countries which started from basically nowhere and over time it created knowledge I think something that you point out very interestingly is the not just the equipment but the no-haul know-how that goes with that that seems to be crucial and yet Taiwan under Morris Chang's leadership um and you must tell us at some point how did you get in to see Morris Chang there's a long line but in any case um you talk about the taiwanization of the chip industry and that suggests a much more Dynamic there's countries that have been in the lead and uh Japan comes to mind and who are no longer and America may be headed in that direction although hopefully not but what was the secret to the taiwanization and is that reversible over time well I think the first thing to recognize about taiwan's role is that on the one hand Taiwan is critically important Taiwan produces 90 of the world's most advanced processor chips and over one-third of the world's computing power that's added each year comes from ships made in Taiwan so you you can't imagine a world without taiwan-made ships but tsmc's success has been not by doing things on their own but by embedding themselves very very deeply in the international chip supply chain so buying Machine Tools from the US and other limbs buying materials and chemicals from leading Japanese firms producing ships designed by companies largely in Silicon Valley and so the Taiwan success story is not a success story about Taiwan going alone it's a success story of integration of the best of Silicon Valley and the best of Taiwan and that's that's the first thing to put taiwan's success in context but secondly tsmc deserves an extraordinary amount of credit for what they've done a lot of this is the business vision of Morris Chang the founder of tsmc who realized based in the late 1980s that the industry was shifting to a model rather than integrating design and Manufacturing in one company of splitting design from manufacturing and he was right to realize at the time that if he focused solely on manufacturing he could eat out more efficiencies and grow to a much larger scale and so today tsmc the company he founded as the world's biggest ship maker and in part because of that it's also the world's most advanced because they're able to hone their production processes over more and more chips but tsmc is is not a company that can do it alone it's a company that works deeply with U.S tool makers with U.S chip designers and so it's part of this symbiosis between Silicon Valley and Taiwan that's made both more successful I think that's a very interesting term symbiosis of what some people forget is that outstanding American companies like Qualcomm have had their success enabled by the by the capabilities it's one thing to design it but someone has to figure out how to manufacture it and manufacture it efficiently that's right with enormous Capital requirements when it's been a great enabler too of American design successes to not have to do the capital expenditure because you can't do a chip startup in the manufacturing business if a new Factory costs 20 billion dollars no Venture firm is going to give you 20 billion dollars to start a new company but if your only costs are hiring a number of Chip designers buying some software licenses it's much more plausible to do startups there so actually Morris Chang and tsmc have enabled a startup ecosystem in California and in Texas and elsewhere in the US that previously didn't exist because they're able to provide Manufacturing Services that lower the cost of Chip startups in the design space well we have the massively successful design firms like Qualcomm perhaps to some degree in Nvidia and AMD I wonder how have you seen a really vibrant ecosystem for new design firms that are coming on stream you do and you see it in a couple different ways part of it is that you have new startups that are being created especially around designing ships for artificial intelligence applications there's a lot of activity in that sphere but next to that the sort of small startups you have big firms big tech companies that are getting into the chip design business so for example Google and Amazon are designing their own ships largely produced by companies like tsmc and that's a big change in the industry but it's important because if you're running a vast data center like Google or Amazon do the more you can design your chips specifically around the processes you're running the more efficient they can become and so a lot of the design Innovation is actually happening in big companies that are venturing into the chip design business for the first time isn't it a lot more expensive for tsmc to manufacture with our in essence small batch designs there are certainly economies of scale for producing a lot more chips but tsmc has lots of customers some of which like apple are producing millions and millions of chips for all the iPhones that are created others that are much smaller and so foundries like the SMC can produce for companies with smaller production runs too well you know tsmc is a success is is quite remarkable do you see that as continuing with the with the top suppliers narrowing or do you think we can at least maintain the current distribution with tsmc and Samsung which are arguably national champions and Intel which we are reluctant to admit might be a national champion but how do you see that working out over time is there any clear path well I think it's it's difficult to predict the future and so I would I would be hesitant to make any uh confident predictions and and I would say that because 10 years ago the pecking order in terms of Technology was different Intel had the most advanced process technology and tsmc was behind so 10 years into the future I don't know who's going to be had behead in that race I think what we can say is that we're unlikely to see any or perhaps many new entrants into that market um at the high at the high end at the high end that's right the amount of knowledge is very very specialized is just too high and the capital expenditure is too high for anyone else to get close or indeed is some European countries have learned for any country to do that's it's just it's just prohibitive um but but it will be interesting you you know you've talked to well there are a number of things that's clearly related to China but you've talked about how or why don't you talk about taiwan's and tsmc's importance to the Chinese economy and where the you you have a very interesting chapter on some of the mutual dependencies there could you elaborate on that a little well DBC produces chips that almost everyone in the world touches on a daily basis inside most smartphones almost all smartphones there's at least one sometimes multiple tsmc produced chips inside many PCs have tsmc produced chips inside basically every data center uses chips that tsmc produces every cell phone tower and that's not to include dishwashers and Automobiles and microwaves so you can't really go a day in the life of a typical American without touching dozens of tsmc products now you never see them and they don't say tsmc on them but we're all deeply reliant on tsmc and it's not just United States it's Europe it's Japan and it's China and China is a country that is more dependent on tsmc than arguably any other because it has no Cutting Edge domestic capabilities but it Imports tons of chips because its economy is dependent on assembling Electronics uh today China spends as much money importing semiconductor resistance importing oil an extraordinary data point that changed the way I see the global economy and the Chinese leadership realizes that this is a vulnerability because their most advanced Computing Technologies are imported from Taiwan from South Korea from Japan and from the United States all geopolitical competitors or Rivals and so China's been trying to find ways to become less dependent on imported ships and less dependent on tsmc and had some success but also remains very Reliant to this day well you know that's that's interesting you say some success you know they've had the well they had majors in 2010 and then they had the 2020 plan and now the 2025 plan and um something is given our Fletcher backgrounds it's a little naive to say but you that China seems to be creating tensions and competing with countries that it's extremely dependent on and I don't see where that leads them on the on the long term perhaps one of the most interesting things in the last six months is the administer Biden administration's decision to uh to block access to Top End chips and people often forget but also the importance of equipment manufacturing equipment from Applied Materials and KLA and lamb here how do you see that playing out there's the Peterson Institute is argued that um they'll be able to um make good in three or four years and circumvent these with domestic production based on what you've written I would be interested to hear your views if you look at the controls that were put in place in October they prevent the transfer of advanced chip making tools from U.S firms to China there's discussions underway but also preventing the transfer of tools made in the Netherlands in Japan which are two other leaders in this year also to China right now you can't make an advanced ship or even a sort of second generation most advanced ship without using U.S tools and in terms of the tools that China produces domestically they're far far behind what U.S Japanese or Dutch firms can produce we interrupted for a second what do you mean by far behind some and sometimes I've seen National Security documents or pcast the president's advisory committee on Science and Technology who say that we have to keep them two generations back which is two years or maybe three and and that seems like a fairly narrow margin to me it was it's a complicated question because not every aspect of the chip industry moves along the same schedule um so when you talk about one or two generation it's a pretty loose term actually um in logic chip making the manufacturer of processor chips uh there's a fair amount of similarity between when firms talk about a given generation in the Next Generation and because of the Dynamics of Moore's Law which has predicted accurately that the computing power provided by a chip will double once every year or two between every Generations there's a big gap in computing power and so two generations actually implies a pretty substantial differential in terms of what a chip can do but as you say it's not a huge differential which is why the US has shifted its approach someone over the past couple of months for a long time the two generation gap was informal US government policy but in a speech in September of last year Jake Sullivan said he wants to grow the Gap so it's no longer going to be a sliding scale where if the U.S moves forward export controls will advance to allow the transfer of older equipment to China now it's going to be a hard stop us is trying to move forward and not let China move forward which will be harder to enforce but also China's fairly weak position in tooling means that it I think is plausible at least over the short medium term short Dominion term meaning at least a number of years five years perhaps I think at least because the trick in the chip industry is that the question isn't at what point will China reach 2023 levels of Technology because we're going to be far ahead of 2023 levels of Technology big question is when can China catch up to where we are in the future that's a really hard thing to do and no country has ever caught up in the chip industry except by deeply integrating into International Supply chains I think it's a very interesting point the although some of our our goals seem to be pushed together to use a very technical term but we seem to want to block them on on artificial intelligence advances and perhaps in quantum and as I understand that the absence of of access will be a major impediment for them on the other hand we 'd be interested in your views we kind of derivatively referred to Legacy chips which is kind of funny because that's what runs everything you know our banking system our our Airlines in fact we could probably use more chips for the airline based on what's been happening but how do you how do you see that playing out if you look at The Cutting Edge um and especially The Cutting Edge of the types of chips that are used in artificial intelligence systems um what you find is that all of the advances in AI over the past decade have come thanks to advances in chips the key differential between ai's capabilities today and 10 years ago is not primarily that we have smarter computer scientists it's that they have access to several orders of magnitude more transistors than they used to and they can process a lot more data train smarter algorithms because you want to train an algorithm to recognize a cat versus a dog you need to show it a lot of pictures of cats a lot of pictures of dogs and that's a lot of transistors that need to do that work for you Advanced chips give you those additional transistors give you that processing power and let you teach systems to become smarter and there's only a couple of companies that can design the types of chips that are used to train AI systems and only a very small number of companies tsmc being the most important that can actually manufacture the chips involved and China has no Cutting Edge manufacturing capabilities and only a small number of firms that are anywhere close on the design side and so if they can't get access to the couple of types of high-end ships that are used to train AI systems they won't be able to keep up in the AI race well there's they're kind of a series of you know interlocking questions from that um one is that it's it's not clear to to me and I'd be interested to know what you think that you do need the advanced chips for hypersonics for Hypersonic missiles and others as I understand it but a lot of the weapon systems can run on higher node chips on what we call Legacy chips and there may be incentives for the Chinese to produce a great deal more of the Legacy chips although the question is will we make the equipment available at that node do you have any idea where we'll go on that because if there is representatives of asml and others have argued that if we block them at the advanced chips they may Mass produce like Japan in the 80s and swamp Global markets with uh something that you made clear in in the book is the if you don't have to make if a company doesn't have to make a profit and goes only for market share they're very formidable competitors so I I think you know first off on the relationship between The Cutting Edge and semiconductors and Legacy nodes or lower Tech nodes if you think The Cutting Edge is really important you also have to think Legacy chips are more important because the more advanced their computational capabilities become the more benefit there is to have computation embedded across the edge of networks as well and and Edge systems will sometimes have Leading Edge ships but we'll also have a lot of lagging Azure Legacy chips in them to undertake the activities they need to undertake and so it's not that Leading Edge is becoming more important it's that advances in Leading Edge and advances artificial intelligence that the Leading Edge enables makes it more useful to network more things together and so we need both and it's absolutely right that China's got the capability domestically to produce lacking Edge ships and we can talk about the specific definition of lagging Edge there's a lot of complexities in there and it's also absolutely right that China's been investing primarily in building manufacturing capacity in the lacking Edge and I think the administration and the Us and other countries have been right to focus on the Leading Edge first because it really is a critical question when you're looking at chips used for AI but it's also important to ask where are we sourcing our lagging Edge chips from and our China's policies calling into question the profitability of companies in the US and Europe and Japan and actually interestingly in Taiwan which is in some ways the most exposed to Chinese competition at the Lakme Edge simply because Taiwan produces a lot of lagging as chips that's an interest that's an interesting point one One requirement that we've heard and believe is probably accurate is for these export controls to be effective over time we need the active support of our allies I'd be interested how you see whether we'll be able to get the Japanese and particularly the Dutch support uh because they're direct economic costs for doing that at the same time it there's a question and it's a real question not a point of view is um the U.S has the authority simply to block the exports because it's our equipment that and our components even for a company like asml that are critical how do you you know we really need allies but under underneath that it seems to be that perhaps we don't well it's an interesting question um because it's true that as Supply chains have gotten more and more complex and as the number of components in advanced manufacturing equipment has gotten more complexions to give you one data point inside of a asml euv lithography system one of the key component systems is a laser produced by a German company called Trump incredible company and they have in their laser 457 000 component parts so you know the likelihood that one of those components comes from the United States is very high and in any system that has a hundred thousand components you're going to be sourcing some components from a wide number of countries and so that gives the Us and other countries more ability than ever before to uh insert political considerations into Supply chains but there are risks in doing that too much because if a country develops a reputation of meddling and Supply chains and of imposing export controls on use components it also creates an incentive for companies in other countries not to put us components into their system so there's a real balance to be struck I think you write that today the US has really extraordinary power over Tech Supply chains because every Tech system uses U.S components but if we were to use all that power today we'd have a lot less of it in the future because we'd create a really strong incentive for companies in Japan Europe and elsewhere to not use our components in their next Generation system just design them out exactly well we've heard that from some of the companies and it's an understandable concern at the same time you can't design out asml lithography at least not and have anything remotely that's right resembling and um and if we keep the Japanese with us they're they're as the French say they don't seem to be 36 Solutions and and of course our companies will continue to advance so those parts will become but there is this discussion of are we really volunteeristic uh where we get a whole Coalition with us and um there is a pretty large Hammer under the desk if I I think that's right and I think um you know my sense is that U.S officials and I think especially in Congress um are are aware of the hammer under the desk if you will um the other Dynamic that's really interesting is if if you're one of the countries we mentioned you have a really strong incentive to hide behind the United States because a China might retaliate and we've seen China retire against many of its trading partners it does less so actually against the United States because of our size smaller countries are more at risk of retaliation and B domestically in other countries uh politicians face costs when they take steps that hurt their firms right and so there's a lot of incentive even if political leaders in other countries are supportive or sympathetic of controls like this to try to hide behind the us and not take responsibility for them and that's understandable that's how politically was operating yes no it is it is understandable the and a broader from an economic historian point of views it's interesting that we seem to have moved into a new phase of of uh and the verb matters of recognizing that we do industrial policy you know you often hear debates you know the dreaded words we don't pick winners and losers often wondered what the Department of Agriculture does for the last 150 years of that and the answer would be they picked a winner and that type of long-term care we seem to be coming back to that more you may be familiar with Laura Tyson's a great quote that in the chip industry and instead of the Invisible Hand of the market there's the visible hand of governments and then conversations uh with theraight remember hearing I think it reflects what you were saying that uh some countries don't want to provide the subsidies needed to attract and maintain corporations and they don't have a semiconductor industry it's just Australia and Canada come to mind in that regard do you think we're entering a new phase of more obvious more declarative and perhaps reinforced industrial policy in terms of the level of discussion of the term certainly the answer is is yes in some ways as you alluded to there's always been a fair amount of government involvement in semiconductors for defense uses for intelligence uses for r d funding via the National Science Foundation government's never been absent I think what we're seeing now is that in the political leadership in congress and the White House and elsewhere there's a lot less concern about this term but I think it needs to be approached carefully as well because semiconductor Supply chains are extraordinarily complicated companies inevitably know more about how the supply chain Works than government officials do that's just an information a symmetry that's baked into the way the system works and so government has to be very very careful in the ways that it does try to intervene and so I think if you look at tax policy here's a great example of where other countries have had much more generous tax policy towards semiconductors it's not rocket Sciences to why companies have invested more other countries it's cheaper to do so that's a really easy place where the chips act has actually made a difference by strengthening tax incentives but that's the type of policy where the government can pretty clearly identify what's the problem what's the solution the amount of information you need to devise a smart policy uh in that sphere is a lot less than if you were to try to undertake something more specific more Niche to the industry and I think we do need to be very cognizant of the ways in which government is going to lag behind industry in terms of knowledge and in terms of its predictive capacity as to what the next technology and markets will look like so there's a real tough balance to be strong here I'm very happy to hear you say that because that's basically the rationale for the restoring American innovation program here at csis [Music] I used to work in the I.T field and the director would sometimes remark well as a last resort we could ask the customer what he needs and I think asking industry what they need and what are the consequences of of what we're doing but the you know the chips actually refer to has been a is clearly a massive one designed to address some of these imbalances in other countries subsidies there are two related things one is um one of the things that worry about having been in the government is the the one and done syndrome we we've passed a chip back so we don't have to do any more and my colleague Dr Shiva Kumar has written that actually we're just starting and another one I'd be interested to hear is on is you know we're all environmentalists these days but we have a permitting problem in this country that I'd be interested to hear your your views and how you how do you think that can be addressed I think the permitting question is is absolutely critical there's been great work on the amount of time and the number of Hoops you need to jump through to build a chip making facility in the U.S versus in Korea or Taiwan and what you find is that it's a lot harder in the US for land use for environmental regulation for getting access to the power and the water you need which is substantial for making it's a real Challenge and it's a challenge that camps solely be addressed at the federal level because you have state and local governments adding layers of complication as well so I think sensitizing political leaders to the challenges that chip Farms face is important ultimately it's hard for the us because unlike Taiwan or South Korea where ships are a critical part of the economy and everyone knows that tsmc success or Samsung success is critical for those two economies in the US we don't really think that way maybe we should but we don't and so getting political leader is to understand and treat the chip industry is a bit more important than they have in the past is going to be important in mobilizing political support to help push through improvements in the permitting front and help companies solve problems that they face when they're building new facilities well I think I think you're right there I think part of the problem on the permitting is one of the things we like about this country there's very diffused political power and local entities have have a lot to say you know they can tie up for years a project which given the capital costs essentially make it non non-viable whereas I'm told and it's probably not apocryphal that you learned is going to be building in your neighborhood when the bulldozers arrive in China it's much more direct but I think you raise a very important point that we don't despite the potato chips analogy we don't seem to recognize just how Central this is to not only National Security but Dale Jorgensen at Harvard who's now left us but he documented the massive productivity impact that not only chips have but making chips have and it's a it's it's a question why why did Morris Chang decide to if I can say that uh put a tsmc5 in Arizona and there's a strategic element of that question and there's Manpower issues which we should talk some more about I don't know if you think there's a water issue there it's um you know what's the base there's those elements together I'd be interested here well I think first off why did tsmc decide to open a facility outside of Taiwan in the first place is a new question they've had a couple of small facilities in China lagging ads for some time they already have had a very small facility and a very lagging Edge in Washington state for some time there's an investment made 20 years ago that has been really very small but I think the company realized over the past five years that there is growing concern from other governments and from their customers about the security situation in the Taiwan Straits and as a result they're opening not only a facility in the US but also in Japan and potentially down the road in in Europe or maybe in Singapore as well and they're going to be diversifying their production footprint somewhat and that's something that they didn't want to do but they've been pushed to do by governments and by customers and if you look at their customers apple or Qualcomm or AMD or Nvidia these are all companies that are reading the same news headlines we are and wondering what international trade is going to look like in a couple years and wondering about all the goods that they rely on to be assembled in China what the supply chains will look like and you know every day there's a new headline about a new tech company reassessing their assembly or their Manufacturing in China and tsmc's customers are thinking through what that means for that company's manufacturing footprint and I think that explains why the company has moved away from their history of being Taiwan focused in production towards a much more distributed footprint now why Arizona Arizona is actually now one of the key centers of Chip making in the U.S so in terms of Workforce it's actually got one of the best developed workforces for the chip industry in terms of water perhaps not that's something you'll have to deal with they're building a Water Reclamation plant actually I think precisely to address that issue well Arizona State University seems to be a major asset and not least their flexibility that's a university that can change things and respond the but you know you mentioned the the geopolitical strains you have a particularly interesting section on talking about Chinese interests in Taiwan and tsmc and Taiwan and why things may not be as bad as as some scenarios suggests could you elaborate on that a little bit what is a is there a threat of invasion um and I might just add as a ringer we often talk about the pla we don't talk about the earthquake Gods that's right perhaps you could elaborate on those two points it would be is there a danger of an invasion and what would happen if they did is there a danger of invasion I think there's certainly a risk um what percentage chance should we put in that risk over the next say 10 years you know we can debate the number I think the number is higher today than it was five years back um and every military official or intelligence official in the U.S that testifies to Congress uh makes me think that a number is growing not decreasing and that's important because even if that number is small 10 or 20 percent if you multiply that number by the expected value of the cost of a war it'd be trillions of dollars in disruption to Global manufacturing and so we ought to be willing to spend a lot of resources today as an insurance policy against the risk that something bad happens in the Taiwan Straits because if we were to lose access because of a blockade or a war to Taiwan produced ships we'd face A disruption manufacturing it can I think to 1929 in the Great Depression right auto manufacturing dishwashers household goods consumer electronics we Face immense delays I think the best example of this is in the last couple of years we faced a semiconductor shortage in 2020 2021 and those were two years when chip production globally increased right this demand grew faster that's a great point in your book that it's not a supply chain problem it was a demand yes we wanted too many chips which is a good problem to have for the chip industry and they responded very effectively in some ways uh increasing production by rapid rates but imagine what the world would be like if rather than a double-digit rate of increase but an even bigger demand increase we had a one-third decline in the amount of processing power that was produced Next Year relative expectation it wouldn't just be smartphones or PCS think of cars cars have hundreds of chips inside of them today many of these ships are made in Taiwan and as we found out over the last two years if your car has 99 but not the final chip it needs it can't drive away yeah exactly you know one of the things I thought was the most disturbing in in your analysis was not the invasion as you referred to it not a D-Day invasion but just a blockade or sufficient threats of a blockade to break the confidence that is necessary could do you think that that's a more likely possibility how and how would that work well we know that after Nancy Pelosi visited in Taiwan uh Last Summer the exercises that the pla undertook were to simulate what you'd need to do to begin a blockade and in some ways the placate is an interesting tactic for China to try they wouldn't call it a blockade they call it a quarantine just like we did to Cuba right in the 1960s and It's tricky because the Soviet Union lost the Cuban Missile Crisis because they were far away and the U.S had far more forces to bring to bear against Cuba from from the Southern United States it would be a very challenging dilemma for the US politically economically as well as militarily how would it play out you know it's been war game thousands of times with different results the answer is we don't know but that makes me nervous yes that intermediate State uh makes one makes makes me makes me very nervous as well it's a you know the Chinese curse I mean you live in exciting times so do you think we have Mr clothes here do you think we have the Manpower that we need to I'm interested to hear to say about it Arizona but do you think we have a talent Pipeline and and a related it's a it's a difficult question but some argue that if we don't import high-end Talent from places like India we just can't do this you know we we talk about educating young women and getting them interested at the age of 12 but they won't have a PhD for another 15 years if not longer so well I think that in terms of importing Talent you know it's really been key to Silicon Valley success since the earliest days you look at Fairchild semiconductor one of the two companies that invented the first integrated circuits you know two of the eight Founders were not born in the United States and you can go through the chip industry and find dozens and dozens of uh Americans who were immigrants in the country Morris Chang being the best example unfortunately he immigrated again that was a big problem we faced so immigration is is going to be key to the solution um but I think in terms of thinking about Manpower problems um on the one hand industry is always saying we need better trained Workforce right and that's always true you always want a better trained Workforce I'm skeptical of the thesis that it's going to be a sufficiently large roadblock to really prevent the US from doing what it needs to do in the chip industry if you look at U.S universities they're training not only the next generation of ship making experts in the US but also globally it's still the case that in Taiwan and Korea they're sending many of their best students to U.S universities to study and so I don't see a reason why this isn't a solvable problem well sadly we send them back well that's right that's right and that's certainly a problem um but I think the workforce question is is a question that's solvable um it's going to require some thinking about companies in terms of what other training programs some universities in terms of how do you train people specifically to work in industry and also I think government in connecting universities with companies and more effective ways and just in closing what do you think is the biggest challenge to the chips act being successful and and how would you the classic question how would you measure success well that's the hard question I think you know in some ways I think the easy part of the chips Act is the the manufacturing incentives um because we know how to companies know how to create uh chip making facilities uh and so if we incentivize them to do so we're going to get more to make facilities in the US I think that's a pretty straightforward math I think the hard part of the chips Act is the r d effort and a quarter of the funds are dedicated to R D um and that's trickier harder to measure but also harder to do successfully but it's I think more important I think in 10 years time when we look back on the chips act we're going to assess the success or failure not on whether we slightly increase the share of manufacturing that happens in the US but whether we kept Moore's Law alive and I don't think nearly enough people I wake up in the morning every morning thinking is you know this can be the day when we find out that Moore's Law is finally dead and if so Society is going to change dramatically economic growth is going to slow dramatically and I don't think we're nearly afraid of that uh to the extent that we ought to be uh good point so I wish political leaders would think more about that question and then put the r d infrastructure in place that would solve that problem over the longer run because we've got a line of sight for the next couple of years to keep shrinking transistors but beyond a couple of years it's harder to say and I don't think we're ready for the consequences of a world in which Moore's Law does come to an end well you did raise some very optimistic predictions in in at the end of your book but I remember asking Gordon Moore personally about Moore's Law and and I pointed out that a lot of people were predicting that it was about to come to an end as it was in the early 2000s and he said it may be but they have been predicting that for the last 40 years right so there may be some help there but thank you I really appreciate you taking the time to be with us here at csis and uh look forward to seeing you here again these questions that we've touched on today are as you suggest are not going away and the devil's always in the detail so I look forward to working with you in the future well thank you for having me thank you [Music]

Channel: Center for Strategic & International Studies

Views: 113,828

Keywords: Center for Strategic and International Studies, CSIS, bipartisan, policy, foreign relations, national security, think tank, politics, global food security, food loss, food waste, sustainability, food tech, biotech, #RAI, #Event, #2C

Id: v5yztfMWXqA

Length: 45min 38sec (2738 seconds)

Published: Wed Jan 11 2023