Davos 2024: The Battle for Chips

This transcript has been generated using speech recognition software and may contain errors. Please check its accuracy against the audio.

Nicholas Thompson: All right. I admit it. I was trying to become the minister of telecommunications in India and I just got demoted. Alright, let's get cracking.

My name is Nicholas Thompson. I'm CEO of the Atlantic I'll be the moderator and facilitator today. We should have an amazing, extremely important discussion about the hottest topic in the world, which is chips. Very exciting geopolitical implications.

We're gonna be talking about manufacturing but also packaging, supply chain. We have incredible people on this panel. We have three incredible ones. We might have a fourth incredible one who'll be showing up depending on the traffic situation outside but let's get cracking.

So, Arati Prabhakar, she is the director of OSTP [Office of Science and Technology Policy] for the Biden administration. In the green room, she was asked what company she worked for, she said the US government.

Arati Prabhakar: That's a big one.

Nicholas Thompson: We have Ashwini Vaishnaw who is – this is amazing guys – minister of railways, minister of communications, minister of electronics, information, tech of India. He does one on ones with all of his 1.2 million direct reports every week.

We have Chris Miller, who is an associate professor at the Fletcher School of Law and Diplomacy. And who wrote the best-timed book that anybody has written since my friend wrote a book about Osama bin Laden published September 10 2001. He wrote a book about chip wars right before US industrial policy was announced. So let's get going.

So, Arati Prabhakar. United States. The last year has been rolling out a series of policies to build up chip industry in the United States. We need to slow down the chip industry in some other countries. Let's start this out by explaining where we are, very quickly, what's been rolled out and very quickly, what's to come.

Arati Prabhakar: OK and I will start with a personal lens. Forty years ago, this year, I finished my PhD thesis in applied physics. It was about semiconductor materials and I've spent those four decades watching the US semiconductor industry grow, get globalized, which was totally fine and then get dangerously concentrated in one part of the world.

Dangerous geopolitically, dangerous in terms of any kinds of natural shocks. And for decades, we knew this was happening, it was a concern, it was a national security concern, it was an economic concern.

And then to my surprise and delight, in the summer of 2022, on a bipartisan basis, Congress passed the Chips and Science Act, which took the first really significant step that the government has ever taken to focus on semiconductor manufacturing in the United States.

And that was a package of $52 billion, about 39 billion of that is manufacturing incentives to bring bleeding edge but also critical other nodes of semiconductors and make sure that they were in the United States and part of still a global ecosystem. And there's an additional, the rest of that was 39 billion, the rest is R&D because we all understand that this is a fast-moving technology-intensive business.

That work is getting executed largely by the Department of Commerce Undersecretary Gina Raimondo's leadership and they have just moved with urgency and focus and I think with just great professionalism to take these financial tools to achieve what is a national purpose. So that's what's going on.

I think it's important for the semiconductor industry in the US but it's also critically important for the world because these semiconductors are the underpinnings for the entire information revolution. So, every time you hear about artificial intelligence here at Davos, you're hearing about something that's been enabled by the semiconductor revolution.

And so it's important, I think, for the whole world and we are very clear in the work that we are doing, we're doing it for US economic and national security advantage and purposes but we also want to do it with our partners around the world.

Nicholas Thompson: So let's stay on that to the point about it being good for the world. So certainly, there's at least one country with more than a billion people that probably doesn't love all this. There must be other countries that say, wait a second, US is dumping, you know, billions and billions of dollars in their chip industry, we're not so psyched about that now the US is going to like be subsidised just as we're always complaining about other subsidies elsewhere.

How is this something that's good for the world when it's so US-centric?

Arati Prabhakar: Well, first of all, we were really the only country that wanted semiconductor manufacturing that wasn't subsidizing manufacturing. And it just became table stakes because of what other nations were doing. If you want semiconductor manufacturing, especially at the bleeding edge, it has to be supported by the government of any country. And so this and in fact, if you look at the scale of the subsidies in other parts of the world, other parts of the world, this is not that huge.

But again, I want to be really clear that this isn't about saying, you know, Fortress America, we're going to put up the gates and just do everything here. We know that this is a deeply globalized industry. And what we want is for it to work and for supply chains to be robust and to have those manufacturing jobs in the US as well as other parts of the world.

Nicholas Thompson: Minister Vaishnaw, I can just ask you this question: US passes this industrial policy, do you like it in India?

In fact, I had someone asked him one night at dinner and I said, do you think the minister in India likes US industrial chip policy. And they said yeah, probably because anything that wrenches it out of East Asia means there's more room for India to play, there are more things to get involved in. Is that accurate?

Ashwini Vaishnaw: We don't look at it that way. We look at this entire semiconductor mission – we call it India semiconductor mission – in a very comprehensive way. We know that this is a very important industry which India is the right place for developing it because we have a very large pool of talent already there. Close to 300,000 design engineers, designing practically every complex chip which is manufactured in the world. So it's very natural to move on to the value chain and move towards manufacturing.

So we are looking at developing the complete ecosystem of semiconductors, which means design capabilities, the software part of it, manufacturing the hardware part of it and talent. We have in our programme, we have promised, we have announced and committed developing 85,000 talent, which includes engineers, BTech, MTech, PhD and technicians for managing the clean rooms, working out the machines, the entire thing, entire spectrum over the next 10 years.

And within the short time span, we announced this policy in January of 2022. It's just a few like what 14,16 months, 18 months and sorry, 24 months now and within this short timeframe, we have tied up with 104 universities, redesigned the entire course curriculum, tied up with international institutions including University of Purdue.

We have had the first plant, which was announced during Prime Minister Modi's US state visit June 23. And within 90 days of that announcement, the construction of the first plant started in September 23. So it's a very rapid progress, very focused. Yes, unlike the US which has $54 billion, our programme is at this point of time, 10 billion US dollars but it's a very comprehensive programme.

Nicholas Thompson: And is the ambition that in 10 years, you will be manufacturing the most cutting edge chips or is the idea that in 10 years you will just have a much bigger, larger role throughout the chip process?

Ashwini Vaishnaw: So, all the experts that we have been discussing with, they say that the entire semiconductor market has evolved in a very interesting way. There is a huge market share and huge demand for chips which are not on that sub seven nanometer, two nanometer kind of cutting edge technologies.

There is a significant demand for 28 nanometer and above, for automotive, for telecom, for power, for power electronics, for train sets, for everything that goes in the day-to-day household unit to go. So at this point of time, we are focused on getting into that market segment and then we move on, once the ecosystem starts developing, then we move on to more bleeding edges.

Nicholas Thompson: OK, so you want to move down for 28 nanometer but you're not quite sure exactly yet. Chris, you're just there. Is this a good policy or is it doomed for failure? Because it's so competitive and they've set their sights too high for India?

Chris Miller: Yeah. Well, I think for India, you've got to look at the historical context. Taiwan, South Korea started a half century ago, developing the chip industries and today they're among the world leaders. And so I think there's no doubt that India is beginning to follow that path. There's a lot of interest among investors at really every stage of the supply chain in India. And that's really positive.

You see Apple shifting iPhone assembly to India, you see design firms investing in India, you see announcements of assembly manufacturing facilities but I also think it's gonna take a long time because these processes always take a long time. And so you shouldn't expect the supply chain to reorient itself overnight because there were tens of billions of dollars every year for the past several decades invested in the supply chain way structured today. And so if you want to shift it, it's going to take a long time.

Nicholas Thompson: Well, let me ask it this way. What is, give me one mistake you think they're making. Overall, you think it's a good idea but tell me a mistake – you've studied very carefully what a country needs to do to build its chip ecosystem? Tell me something that India is doing wrong.

Chris Miller: Well, I'll tell you a mistake that many countries make, which is that most political leaders want a picture of themselves in front of a big manufacturing facility. Most of the money in the chip industry accrues to people who design chips and don't manufacture them. And so in general, we all collectively overstate the importance of manufacturing, which is very important but understate the importance of design.

Nicholas Thompson: That's interesting, Arati, have you ever been photographed in front of a big chip manufacturing factory?

Arati Prabhakar: It's been a long time since I've been in a bunny suit in a semiconductor fab but I'm sure there's a picture somewhere.

Nicholas Thompson: But tell me, I mean this actually. Welcome. Hello? Yeah. You know, be quiet. Of course you're gonna be noticed. You've watched, right? I mean.

Micky Adriaansens: I was in a traffic jam for quite a while. So sorry. Sorry for that.

Nicholas Thompson: We have Micky Adriaansens, the minister of economic affairs and climate policy of the Netherlands. Welcome to our chip panel. We're discussing chips. We're discussing India's industrial policy, United States industrial policy right here.

Tell me how you think about the different parts of the chip sector and where the United States puts its emphasis. What is the most important thing that we need to fix right now in the US chip manufacturing process?

Arati Prabhakar: To Chris's point, design is an incredibly important part of it when the fabs business model emerged, we had the rise of fabs semiconductor companies and we are blessed with incredible fabs semiconductor companies in the United States, Nvidia, Qualcomm, so many major players it's also a place where startups have been possible for a very long time.

And really, I think if you think about the Chips Act, which is focused on manufacturing, it is about looking at the entire ecosystem, recognizing that we have, we had and continue to have a tremendous strength in design but when all of the bleeding edge chips that are critical to our infrastructure to AI, to our national security ambitions, to automotive manufacturing, even are being built in one part, a fragile part of the world, that that's really not...

Nicholas Thompson: That fragile part of the world?

Arati Prabhakar: Taiwan. Absolutely. And I think this is not news. This has been true for a number of years. What happened, of course, is the pandemic just rang the bell. Everyone saw how critically vulnerable that this whole situation was. And when auto manufacturers weren't able to manufacture cars because chips weren't available, it just showed us what some of the issues were and I think that was an important wake-up call. I think that's why Congress took this incredibly bold and important step.

And success for us looks like continuing to have a flourishing design industry but seeing manufacturing happen in many parts of the world, including in the US and many you design, you manufacture but all the backend things that have to happen before you actually have a system happen globally. So it's a much wider network. And I always think of these chips as more globally well-travelled than any of us before they actually end up somewhere, and I think that will continue to be the case.

Nicholas Thompson: So, Minister Adriaansens.

Micky Adriaansens: Correct.

Nicholas Thompson: OK, great. We've been talking about the US industrial chip policy, we've been talking about the Chips Act. I am very curious for your view because it has made it, I suppose somewhat better for Dutch companies that US is building some better stuff can work you know, with advanced lithography, maybe there are ways with ASML that ties in with US manufacturing but it's also made it harder for you because you can't sell as much to East Asia.

So what do you think of our policy? Do you think it's good or do you think it's harmful?

Micky Adriaansens: I think that we're entering a new world. I think that what we had all these years all these decades, all these centuries that we had this open trade possibilities and now it's changing. Like the economy is more a battle. It's a battlefield. So we have to be aware of the new situation.

So I think that we have to change into more, the policy needs to change and it's about protecting what is not good to use by bad parties, that countries that we want not this knowledge to be used in the wrong way. So we have to protect ourselves.

And I think we have seen a lot of that in the world today that we see how important it is that we do that, that we protect ourselves. And we do that also in the Netherlands. So it's new policy that we have to.

But on the other hand, I think that the real solution is in the promote side that we need to have technological leadership. We cannot only protect and close our borders. We really have to think very carefully, what we have to protect in what way and that we keep investing in new technologies and have this technological leadership. And that's a very delicate matter.

It's not like good or it's not good or bad or something. It's not that simple. And I think that working together with like-minded countries and seeing what parts of the supply chain we need to invest more, like the design phase. It's very interesting to have more cooperation over there, we really need to do that.

So then we can be very strong and that also good for regarding TSMC [Taiwan Semiconductor Manufacturing Company], they are building now a new factory in Germany, which is I think very interesting because all European parties are investing also there. So that's the new world.

Nicholas Thompson: All right. So Minister Vaishnaw, we have this battle. We have the West is taking a side, it's taken a side against China. You're in the middle, geographically, how are you going to navigate this battle?

Ashwini Vaishnaw: If you look at how India has conducted its foreign policy, its economic policy, the way it has conducted itself. The entire world has developed a very big amount of trust on India. That trust is what is our biggest capital that trust holds kind of gels the entire world into a very harmonious way.

And that is what is bringing us. For example, we have signed memorandum of cooperation with US, with Europe, with Japan. On semiconductors, we are working with the South Korean government and South Korean companies. It's like practically every possible company wants to collaborate with India in one way or the other.

So we don't think that it's a battle. We think that there is enough for everybody. It's a question of how important, how much importance we give and how much talent we put in it and how much focused we are.

Nicholas Thompson: But that works for now. Right? Once you start building high-end chips, you know United States gonna come to you and say, yep, by the way, you know, that deal you have with, you know, the Chinese large language model company, sorry, that feels off. And you're gonna say, wait a second, right? That's not gonna be a great situation to be in, right?

Ashwini Vaishnaw: Listen, this technology will, this entire sector will evolve in multiple ways. I don't think anybody can see what's going to happen in 10 years from now. It's something which is too complex and too dynamic and too uncertain for anybody to make a bet that okay, this is how the sector will be in 10 years from now.

Just take five years from now, five years back, nobody thought that 28 nanometers or 40 nanometers would be viable anymore, right. But the way electric vehicle industry has evolved, the way the telecom has become all-pervasive, those legacy nodes, as they were called, have become very important market share today. So things will change will keep evolving with it.

Nicholas Thompson: All right. So Chris, let me ask you a meta question. So there's some people in this room who run AI companies desperate for chips, right and they will be able to make better products, serve the world better, improve economic efficiency, the more amazing chips are available at low prices.

Does the policy of my government, does that actually help or does it hurt? Certainly, it helps interest of US chip makers but does it make it better for the developers, the AI companies around the world? Or does it make it worse?

Chris Miller: Well, I think if you look at AI developers in most countries, it probably has a marginally positive impact. If you're subsidising chip manufacturing, the cost of chip manufacturing falls a bit although it's not going to be decisive. I think the real challenge with this question is the countries that can't get access to the advanced chips because of the rules the US rolled out in 2022.

And there's a number of countries who most important is China, where the US is actively trying to restrict their access to the most advanced chips on the theory that access to advanced compute is critical for advances in AI and that's a theory that has been borne out historically. And the US is betting that that will remain true for the foreseeable future.

And that's why the US is focusing on GPUs, the very chips that are critical for AI as the centre of their tech restrictions.

Nicholas Thompson: OK, so you think that it's generally, generally a net positive because most of the companies are there but it's terrible for Chinese AI companies done?

Chris Miller: If not terrible, it's certainly a headwind.

Nicholas Thompson: So, let me run a theory by you or let me run a theory by both of you. So, I've been at actually two events now where different people have made the same point and they've said actually, the US restrictions have just been good for Huawei, right because there are big AI companies.

We've all been in panels with Kai-Fu Lee, he's just released a large language model, it's now second on the Alpaca leaderboards, it's eighth on the Hugging Face leaderboard. It is clearly working and they are using chips most of which were stockpiled before the US ban went into effect but the argument that he and other people have made is, look, they're gonna make the chips. So now they're just going to Huawei, right?

They're not buying from them from Nvidia. They're not going to ASML and Huawei is just going to build up its capacity much faster. So we're gonna have the world completely split and we're not really gonna set China back.

Chris Miller: Well, I think there's a lot of uncertainty but I would say a couple things. One, always production capacity is probably not gonna be that high this year. We'll see where it is next year. Second, the quality of those chips are probably worse than Nvidia's. Third, every AI model in China has been trained in an Nvidia ecosystem. And so there's gonna be switching costs, which is why in the US and the rest of the world people aren't switching.

And so when you put all that together? I think it's pretty clear that there are costs involved. Now, is it insurmountable? No, it's not insurmountable in it's going to totally stop AI in China. Obviously not. But is it going to throw sand in the gears? That's the theory of US policymakers and it seems pretty plausible to me.

Arati Prabhakar: Yeah, I think that's exactly right. And none of these regulations or export control policies are absolute or complete but it's about slowing and creating time for us to be able to advance and understand and be able to shape the direction that AI is going, especially for national security purposes. And I want to come back and put this conversation about export controls in context.

The semiconductor market is vast, it's over half a trillion dollars a year in revenues. And what we're talking about here is some of the most advanced technology components that GPUs power, the most advanced AI and that is a very deliberate plant strategy on the part of the United States that's focused on a specific set of components, some of the underlying manufacturing technologies that are key for those advanced process nodes and it is specifically designed at aiming at the national security capabilities, the military capabilities that China's building on top of AI.

A decade ago, I was running DARPA, so I know how important artificial intelligence is going to be to the next generation of military systems. Those components are how you get there. And, you know, just for the United States, that makes no sense that our most advanced AI components should be helping accelerate China's military ambitions. I think that's the very specific target of these export controls.

Nicholas Thompson: Can you stay on this for a second? Because it's a topic that's really quite interesting and it's the degree to which – and either of you or anybody here can answer this question, which is the degree to which one thinks the US policy is based on slowing down China as an economic competitor, slowing down China as a potential military competitor or third, developing capacities around the world, in the event that China eventually controls TSMC.

Arati Prabhakar: I will tell you that the purpose of our export control policies is national security. And then I will also tell you that in the world in which we operate today, information technologies and now this wave of AI are so integral to both economic progress and military national security capabilities that working on the national security front, of course, has economic implications.

So we seek to be economic, of course, we're going to economic competitors with China. Everyone around the world is in an economic competition. But that's about the marketplace. Military and national security matters are definitely in a different category.

Nicholas Thompson: So, there are two categories of military concerns, number one being worried about Chinese military advancements and number two, being worried about our own capacity to develop advanced chips in the event that China were to control TSMC – which of those two is more important?

Arati Prabhakar: Well, I think the Taiwanese location for. I mean, first of all Taiwan has just done an amazing thing over these multiple decades that they've built the capacity and watching TSMC become such a leader and driving the bleeding edge of the semiconductor technology. You just have to admire that. I think that's really quite amazing.

But the fact remains that when so much of the capacity at the bleeding edge and so many of these important chips, especially for AI and advanced military systems, go through... it's just a bottleneck and I think that's the fundamental issue. So that is part of the motivation, obviously, behind the Chips Act.

Chris Miller: We've put a number on that – 90% of the most advanced processor chips are made in Taiwan. Your smartphone processor, your PC processor, almost every single AI processor. So you're talking about concentration, this is vastly more concentrated than the oil market. And unlike oil, chips aren't fungible, you need the exact specific type of chip, so if there were a disruption in Taiwan – a blockade or a war.

Bloomberg, last week, estimated the cost at $10 trillion, even if they're 10 times the right number, it's still vast. And so even if you think it's only 10% likely, your willingness to buy insurance should be pretty expensive. And I think that's essentially what the US Chips Act and the EU Chips Act and Japanese policies, their insurance policies against the worst-case scenario.

Nicholas Thompson: Minister Adriaansens, does it help you sleep at night to know that if anybody were to threaten ASML, the United States military would come to your defence?

Micky Adriaansens: Let me rephrase it. What do you want to ask?

Nicholas Thompson: Let me rephrase the question. Explain how you think about cooperation across Europe. Now that you're part of the European Union, because clearly you do the advanced lithography in your country.

But in the ASML machines, you're taking parts from Germany, you're taking parts from ZEISS; explain how you think about the geopolitics of Europe as it comes to your chip manufacturing.

Micky Adriaansens: And I think that the Chips Act is is very good strategy for the European Union for different reasons. It's the first time that we really choose that we need to have a supply chain, which is of such strategic importance. So, it's a learning procedure as well.

So we say that we have these goals and we want to achieve it together, which I think is amazing. And I think that Europe should do that more with regard to these critical technologies we have and we need for the future.

So, that's I think, the main point and the second is that the chips are the most important products to act upon because of the dependencies because of what we see of the risk for national security. And with regard to ASML, it's a Dutch company and we're very proud of that. But it's really a European company, as you mentioned, because I think they have 800 or 900 suppliers and there are very small and they're very big, in size, etc, etc. So they're very much connected.

So that's the reason that a European issue. So it's not just the Netherlands acting here. So I think we have to join forces more. And we are building now the semiconductor board and we try to make strategy and you're more strong, more together. And that's the way to go.

Nicholas Thompson: And how confident are you that ASML will be able to maintain its lead over the next five or 10 years as India rises, as Huawei builds its own capacities, as the US builds its hiring capacity?

Micky Adriaansens: I'm not I'm not a technician but they tell me that they have some years progress in advance and maybe five or 10 and some say even more but with regard to the EUV machine technology. But either way, we have to speed up because India's really a different story from China.

So I would say join forces on the right with like minded countries. So it doesn't matter where we build this capacity as long as it's not there where it's misused or abused or for the wrong reasons. So I think that's our main goal.

Nicholas Thompson: Alright, Minister Vaishnaw, as you've been thinking about building up your industry, you must have studied TSMC and how it was able to grow so incredibly, it's a really an astonishing story. I'm sure you've read Chris's book.

When you thought about the different components where India can be most successful. How did you wait and where did you decide because you've had a $10 billion investment not a $50 billion investment. How are you allocating and tell me where you think you have advantages and market edges?

Ashwini Vaishnaw: So the biggest priority is on design. That's where we have a very big strength. We have as I said about 300,000 design engineers. In design, we are advancing from being a back office to actually producing complete chip, rather actually designing the complete chip to going out to the customers, taking the concept and then developing the entire chip. So that's a very important journey. So this is a very big focus for us.

The second big focus for us is to make sure that talent for all parts of the ecosystem gets developed. And in that, it's very important to continuously upgrade the cost curriculums. It is very important to make sure that we are connected with the companies which are working on it. For example, Applied Materials. We've worked with them and they have approved a $300 million investment in India, where they have set up a big base where they're designing the equipments and that's an important point for ASML.

Also, they are designing the equipment, manufacturing the components there, so that their own supply chain also becomes resilient. So, there is lot of collaboration. There's collaboration with another company where they have set up their most advanced training system called Semiverse, which will be used to train about 6,000 technicians in India, in semiconductor manufacturing. So those kinds of collaborations are what we are focusing on.

Nicholas Thompson: Makes a lot of sense. Dr Prabhakar, what would it take for the US to decide to flip its policy to say you know what, just flip the export control part of the policy, not the invest in US chips but to say, you know what, we're not gonna we're not gonna ban them from selling their stuff and using their stuff to China. In fact, we're gonna let Kai-Fu buy all the chips he wants from Nvidia.

What does it take to get there?

Arati Prabhakar: I don't have the answer to that. The policy was designed to recognize that the technology is moving that what constitutes the bleeding edge is moving and we want to be able to have people around the world be able to access GPUs to build AI systems. It's just really about making sure we get access to the bleeding edge, which is where so much of the interesting stuff is happening.

So, we've already adapted those regulations. Once. You mentioned that they originally came out in October of 2022 and a year later, we updated them. And I think that will continue because this is the fast, the chip technology is moving fast. But of course, AI is morphing even as we speak and so I think that will be an evolving landscape and I think it's hard to speculate about the answer to your question.

Nicholas Thompson: Well, fortunately, we have a professor here who is paid to speculate. So Chris, do you foresee any plausible scenarios where maybe let's not make it as pie in the sky where the US lifts all textbook controls but where the direction of the tension over chips reverses and instead of kind of being pulled more apart, we start to pull more together?

Chris Miller: You know, I think that the technology is downstream of geopolitics. And so there's lots of people in the chip industry say, can't we sort this out? And if it were just a chip question, I'm sure you could sort it out. But it's not it's a chip question, it's a political question. It's about the military balance in East Asia. It's about China's ambitions. It's about the US position and until those issues are sorted out, which might take years or decades.

I think the entire technology competition will be structured by that fact. And so there's no easy fixes here because the core issue is the hardest issue in international politics.

Nicholas Thompson: And do you think that the US will be able to maintain or the West will be able to maintain its lead in bleeding edge technology as we get ever smaller?

Chris Miller: Well, nobody knows because we're predicting the future. But if you asked me would I rather bet on an ecosystem selling to 80% of global GDP or an ecosystem selling to domestic China GDP, 20% of GDP, I think I pretty obviously bet on the former.

Nicholas Thompson: Alright, let's go back to the Netherlands because you have as a second part of your portfolio which is super interesting. You're the minister of climate policy. And one of the big concerns about chips and about AI is the amount of energy they use.

Right now, AI data centres are maybe 2% of all energy emissions but heading towards 20, right? I can't remember the stat that I heard earlier at the World Economic Forum but it's going from small to large. What are you doing to make sure that as we go through this technological revolution, the climate costs are not massive?

Micky Adriaansens: I think that's a question we have to ask ourselves but it's inevitable that we have to deal with it because we need these chips for all the solutions for the challenges we are facing, also for the climate, for energy transition we are trying to achieve etc, etc.

So we need the chips but we need on the other hand also to have in the design phase, taken into account more what they need for energy. And in the Netherlands, we are supporting some initiatives to do that better and also to use chips more in using less energy in production processes.

So, there are different ways of solving this problem but it will not be solved in the end. I mean, it will cost a lot of energy. So, I think that's where we have to deal with but we can diminish the demand.

Nicholas Thompson: Now, is this a major priority for India as you begin this industrialization, you build out this whole new sector of your economy?

Ashwini Vaishnaw: Yes, it is a very important priority sector for us. And another point I would like to add is when we are developing the ecosystem, we have also looked at what is going to be the energy which goes into semiconductor manufacturing.

So we have selected a cluster where close to 30,000 megawatts which is going to be the biggest, largest renewable energy plant at one single location in the world. Currently, the largest is 7,000 megawatts, this will be 30,000 megawatts. So all this green energy will go into manufacturing, which includes semiconductors, which includes data centres also, of course, there will be a series of data centres, which will be fed by renewable energy. So it's a very, very interesting and a very comprehensive way we are looking at.

To the question that you were asking earlier, the power consumption issue will be solved by advanced packaging. A lot of it will come from design and advanced packaging. It will not be simply like, you add five chips and you get the power consumption, which is five times, no. Lots of lots of research is going into that. And that's actually going to define the new Moore's law, that's actually going to define the climate sustainability.

Nicholas Thompson: Well Chris, let me ask you this: one of the things I love about your book is how vivid your descriptions of the manufacturer of chips are and how well you describe different systems and the evolution, how they're made, it's really an extraordinary part of how you do it. And I used to work a lot in the New Yorker, editing stories about how scientists describe and it is a real art.

Help us understand the evolution of how you can make chips that use power but are actually more powerful and explain a little bit about the packaging, about the manufacturing and about your best hope for where this could go.

Chris Miller: Well, I would say that there's always been extraordinary demand from industry for more efficient chips. Energy inefficient computing is where we started and as you want it to miniaturize computing and make it ubiquitous you've had to economize on energy. So this has been a trend not just over the past decade but over the past over half a century. And every use case, whether it's data centres, phones, PCs, everyone is demanding more efficiency because battery life matters because your energy cost matters.

And so there's huge focus in the design stage in the packaging stage and in the manufacturing itself on finding ways to economize and the trend in the industry has been year after year, we get more energy-efficient chips per unit of compute.

Nicholas Thompson: One of the – on this point – one of the critiques of the Chips Act and of the Biden administration's policy and you can answer this critique and I'm sure you have one, is that there hasn't been enough emphasis on quantum chips or chips that are multiple generations out. It's been much more about catching up.

And in fact, you know, let's just say quantum chips, companies can't fund that or most companies can't fund that, only a few places they can fund that. United States government, actually maybe Purdue University which you're working with may be able to fund that but how are we thinking about that kind of far out investment?

Arati Prabhakar: Well, I spent a lot of my career and public sector half of my career at DARPA funding all of those amazing highly advanced new chips, some of them actually turned into businesses and products that surround us in our world today.

They are all fabbed in a fab in Taiwan, I mean this that's actually become a strategic gap in the American semiconductor capability and that's what chips is all about. So of course, we don't want to stop innovating at the bleeding edge and designing these amazing new chips, that's absolutely going on.

But chips is not chips' purpose. Well, I want to tie that actually back to this other conversation. We've talked about the manufacturing incentives, very important. The other critically important part of the Chips Act is semiconductor R&D. And if you think about what the purpose of that is, we always talk about the industry as it is today.

If you read Chris's book, you realize that it wasn't always this way. Right now, it's just companies that design and it's design tool vendors and it's all the supply chain for all the manufacturing equipment and the materials and the chemistry. And then there are fabs and then there's packaging and tests. But you know, that's not written on stone tablets.

That's the consequence of the way that the technology developed and what the Minister is talking about what Chris was talking about, with advanced packaging, actually could be one of the ways that that the industry structure starts changing.

So a strategically important question beyond manufacturing and design is what are the R&D investments that recast the industry and how will advantage shift when that happens? And that's the purpose of the R&D part of the Chips Act and why it is, I think, at least as important as the manufacturing that brings bleeding edge back to the US.

Nicholas Thompson: That makes good sense and that's wonderful background for everyone. I want to spend our last few minutes going through an issue that we've touched on a couple of times but I think is at the core of the war for chips.

So Chris, let me ask you this, if tensions between the United States and China rose, got worse and worse and worse and there's conflict in the South China Sea and terrible things happen, suddenly United States loses access to TSMC, in fact, the whole West loses access to TSMC, what happens to the global economy?

Chris Miller: If the West loses access to TSMC, TSMC shuts down because TSMC has a very complex supply chain that relies on Japan, United States, the Netherlands and others. So there's no scenario in which case TSMC is functioning but not plugged into international supply chains. And what we learned during the pandemic was that it's not just smartphones that require chips. It's everything.

It's cars, it's dishwashers, it's coffee makers. It's refrigerators and I think that the auto industry is the best case study of this. There were car companies that had 999 of the chips they needed for their car. They didn't have the last one and they couldn't produce the cars as a result. And so the car industry alone lost half a trillion dollars worth of sales during the pandemic, despite the fact that during the pandemic the world produced more chips each year, we just had a deficit of certain types of chips.

Now Taiwan produces 90% of most advanced chips, on-third of the world's computing power each year is manufactured in Taiwan. So, when Bloomberg says $10 trillion, I think that sounds ballpark about right.

Nicholas Thompson: So, $10 trillion effect to the US economy, where there's a $10 trillion effect on the world economy

Chris Miller: That's the Bloomberg number.

Nicholas Thompson: All right. So then let's just keep walking through the scenario. So what does the United States do. We quickly try to start building up. We try to go ten times as fast as we can try and correct. How capable how able are we to do that.

Chris Miller: Well, it's hard to do because you need to not only build the facilities, you need to buy the tools and you need to train the workforce. And all that is something that you can spend a lot of money on but ultimately it takes a lot of time to build the equipment to train the workforce.

Nicholas Thompson: Alright, so let's say that India, ups its investment from $10 billion, increase chips investment from $50 billion to $500 billion, European Union throws in another $100 billion dollars. If all that were to happen, how long would it take for the United States or for the West, to be able to survive the loss of TSMC.

Chris Miller: It would be measured in years and in each one of those years, you would have a trillion-dollar or multiple-trillion dollar hit to world GDP.

Nicholas Thompson: And so every year in which the US catches up, the amount of leverage over TSMC does decline. Correct.

Chris Miller: You mean every year that the US and Europe and Japan produce more chips domestically? Well, that's the theory of the case. You're buying insurance in case you lose access to all the production in Taiwan and in China, which also produces a large number of lower-end.

Nicholas Thompson: Dr Prabhakar, you're nodding at that, is that?

Arati Prabhakar: Let's not have like guilt the catastrophe, let's just make sure this...

Nicholas Thompson: The panel is called the war over chips. It's not called we like chips or the science of chips. It's called the war over chips. So we're sticking on that.

Arati Prabhakar: Economic rather than.

Nicholas Thompson: Okay, fine. But do you agree with this analysis?

Arati Prabhakar: I think Chris is really thoughtful on this. And I think it really points to how important semiconductors are in the economy and then and then this dangerous concentration, that's exactly what the Chips Act is taking action to address.

Nicholas Thompson: And you have a sense of the timeframe in which the West can become wholly resilient and independent?

Arati Prabhakar: Well, wholly resilient, resilient enough, right. I mean, perfection is not on the cards but we can certainly get to a less vulnerable spot than we are and it's not just the West it's really the whole globe that is so critically reliant on this one geography right now. It's the end of this years but you know, it's like Napoleon in the trees when you know, it's gonna take a while start now and that's why it's so important that we have gotten started.

Nicholas Thompson: Wonderful. We have only just a couple minutes left and I'd like to end on a note of optimism. So, I'd love to go around and have each of you talk about what you are the best case scenario that you foresee, like the best reasonable case that can take us away from a pretty tense situation that we've been talking about. So, Minister Adriaansens.

Micky Adriaansens: Yes, well, let's keep it for myself, a small country, the Netherlands. We thrive by trade, for decades, for centuries, our economy is dependent on trade and open trade, and that's changing now. So, it requires new perspective. And I think that we can only do that because with the trade becoming less and open trade is not open anymore. It also concerns welfare in the end. So it will have an effect so we have to act carefully.

And I think that talking about this and having the right leadership not black or white, not the war zone but sensible strategy, working together, finding the like-minded countries and parties and companies to build up this type of capabilities is the right thing to do.

Nicholas Thompson: Minister Vaishnaw?

Ashwini Vaishnaw: I would echo the point that finding right-minded, having the countries which can collaborate with each other and speeding up the collaboration, making sure that the countries which can contribute more, they get that opportunity, they get that investment, they get the technology, that would be the best case scenario for the world economy.

Nicholas Thompson: And that is also a very good scenario for the economy of India.

Ashwini Vaishnaw: Of course.

Nicholas Thompson: Chris?

Chris Miller: I would agree with that. I think the good news is that's already happening. People talk about deglobalization but I see Intel investing more than ever in Ireland and in Germany and Israel TSMC investing in the United States and Japan. Deglobalization is not the story in the chip industry, bifurcation into a China's sphere and a non-China sphere but the rest of the world is getting more globalized than ever.

Nicholas Thompson: So if you were to look at the total amount of trade and the number of chips that are crossing borders and the parts that are crossing borders, that is going way up. It's just the West in China where it's declining.

Chris Miller: Well, its trade flows and there's investment flows and investment flows is where you see the real increase over the past couple of years.

Nicholas Thompson: Wow. And you expect that to continue?

Chris Miller: I don't see any reason why it would reverse.

Nicholas Thompson: OK. Dr Prabhakar.

Arati Prabhakar: I think these are great comments. I'll add two things. In the context of the United States, for a number of years, we basically said globalization and the markets will solve everything; they did amazing things, they definitely did not solve everything.

When as we are changing course and with President Biden's focus on a modern industrial strategy, which is chips and the way we're working with the industry there but also the bipartisan infrastructure law, also the clean energy investments that we're making through the Inflation Reduction Act. What we are not saying is, well now the government's just gonna go do all of this.

What we are crafting is a new kind of partnership with the private sector where we do together, what neither of us can do separately and I think I'm very optimistic about that as being a powerful way to achieve what the country needs but also to tap the power of the market and I that's an area for great optimism for me.

Nicholas Thompson: A final note that talks about public-private sector partnerships and the power of the market is a wonderful way to end the panel on the World Economic Forum. So, thank you very much. Thank you all for a fabulous illuminating conversation.