AI, Supply Chains, and the Future of Economic Power

Our models need to be superior than all the other models on Earth. We need to have the world's best AI innovation on a qualitative basis, so we need to have the most market share. If we have exquisite models that no one uses, we're kind of irrelevant. Americans are positive some thinkers. We don't see the world as a finite pie that needs to be subdivided into smaller slices. We see a world where the pie expands, which is the lifeblood of starting a business. What does it take to win the AI race? At the A16Z American Dunemism Summit, I sat down with Jacob Helberg, under Secretary of State for Economic Affairs, to discuss AI, manufacturing, supply chains, and the new geopolitics of technology. Drawing on themes from his book, The Wires of War, we discussed why hardware, industrial capacity, and secure supply chains are now essential to both economic strength and national security. Jacob also impacts what it means to lead in AI, from model innovation and global adoption to energy, compute, tariffs, and the reindustrialization of the US. Jacob, welcome to the podcast. Thanks for having me. It's good to be here. We're going to get into a lot of exciting conversation about future economy, AI, manufacturing, supply chain, but I want to start at some of your intellectual origin. You wrote a book The Wires of War, came out in 2021. When you briefly explained the high level thesis of the book, and then I'm curious if you wrote an addendum today about how the last five years have played out what you might write. Yeah, so first of all, it's great to be here and it's great to see you. I'm excited for this conversation and then have a day opportunity to lay out some of the things that the administration's been working on. In 2021, I came out with a book called The Wires of War that was basically proposing a thesis that at the time was somewhat controversial but has become common wisdom since then, which is that the technology world is really in the grips of a two front geopolitical technology war. There's the software end and the hardware end on the software side. At the time, we were seeing the emergence of things like foreign information operations. And on the hardware side, which the whole premise of the book was really to actually say that the hardware end was the decisive battlefield. There was obviously the battle against Huawei and really the infrastructure of the internet. And the basic reason why the book argued that the hardware end was the most important one is because if a country controls the internet at the hardware level, they can compromise and control everything that runs on top of it. In our system, we have private companies that live very independently from the US government and that operate in a private capacity, as we know. And that system is not the case in other places, including in China, where the companies are obviously much more subordinate to the centralized government. And so ultimately, when you have a foreign government that can whip up its companies as instruments of political power that's very concerning for what it means for national sovereignty. And it's funny because you started your career on the software side at Google, right? That's right. And you realize that hardware now you're spending more time there in your role. When you flush out exactly what were the sort of key risks you identified on the hardware side and how would they played out, how would they change or say the same? Yeah, so really what we're talking about is if you have a foreign government that controls that can instrumentalize its tech companies that operate the hardware of the internet, it fundamentally challenges our traditional conception of sovereignty. Because it raises a question of what does it mean to be sovereign when you have a foreign power that knows all the deepest darkest secrets of all the judges, journalists, and jurists in your country and politicians in your country. And the definition of sovereignty starts to look very, very different in a world like that. And so the first Trump administration carried out and prosecuted a very comprehensive campaign that basically argued that Huawei and ZT are national security threats. That campaign has been sustained across administrations since then, including this one. And so today, obviously to your earlier question, we live in a world that's the same in some ways, but that has evolved and mutated in really interesting ways with the AI revolution. It actually was arguably prompted by the release of Chattachee, BTA, year after my book. And the AI revolution has really heightened the stakes of this whole technology race in pretty fundamental ways because it's really put on steroids the capabilities that governments can have to actually carry out these types of operations. Yeah. And so let's not get to your work in the administration. Why don't you talk about your main priorities, goals, and achievements today? Yeah. So our cognizant and recognizing the fact that the AI revolution is going to totally transform economic power and military power. And I'd love to talk a little bit more about that. The administration has laid out a policy to win the AI race. The president proclaimed a very, very historic speech last July at a summit that David Saxon, I co-hosted called Winning the AI Race where he rolled out the AI Action Plan. And in that speech, he basically laid out that he acknowledged that America was in a race and he proclaimed that America was going to win that race. And it was very aspirational and David likes to reference that it was almost reminiscent of JFK speech to the moon, which I actually happened to totally agree with him on that. I think it very much echoed that. And so then that raises a question for agencies like ours. How do we actually help him achieve that goal? From our vantage point, we see three conditions that need to be met to win the AI race. We need to have the world's best technology and AI innovation on a qualitative basis. So our models need to be superior than all the other models on Earth. That's number one. Number two, we need to have the most market share because if we have exquisite models that no one uses, as we know from the world of consumer technology, we're kind of irrelevant. So we need to have the most users. We want to have the app store that has the most apps. We want to be the platform for the rest of the world. And number three, we need secure supply chains because if we have technologies that are firing on all cylinders and technology companies that are firing on all cylinders, but if one tiny cog gets loose and trips up the whole machine, we are in a very brittle precarious position. And so we need to make sure that our companies who rely on many different types of inputs and components that come from very different corners of the globe actually have supply chains that are reliable, that are cost effective, and that are secure. And so we're working very hard to do that. We launched an initiative called Paxileca that aims at securing the supply chain and partnering with some of the world's most technological advanced economies to develop partnerships to secure off-take agreements, joint ventures, and the like. And so we're very excited about having this three front strategy to win the AI race. Yeah, let's go through the three fronts that you just mentioned. And you've mentioned a little bit about what we've achieved. Let's talk more about that, but then also what needs to happen in order for those conditions to be met. Yeah, so having the best AI innovation on a qualitative basis is really an effort. So let me just say that we already have the best AI innovation on a qualitative basis. So the question is how do we maintain our lead? One of the big thorny challenges that's new and very apparent to everyone is the challenge of model distillation, which is fundamentally a undermined intellectual property and fundamentally undermines the unit economics of the industry because when companies invest hundreds of billions of dollars in capital investments and then to develop fancy model ways, but then other companies can just take them. That's a major problem. So that is an issue that we obviously are tracking and our administration is the whole is engaged in conversations with companies to try to find solutions to this. But the government, the companies in the AI labs are the ones in the driver's seat when he comes to developing to ensuring our lead on a qualitative basis as they should be. With respect to market chair, we believe in the power of that our products are superior and therefore more people like them. With that being said, we also recognize that they're competing on a playing field that's not always fair and leveled. And so we are actively engaged with governments to basically ensure that we have a preferential preference for technologies that are secure. And this is really an extension of the work that we've done on Huawei. A lot of our focus has been on hardware infrastructure, but I think there's a lot of intellectual work being done for what does that mean at the AI age? What does that mean in a world where you have open model weights that are fundamentally harder to contain? So market chair is very much a part of the thinking. That is the premise for the AI expert program. We want to grow our market share worldwide. David Sachs and Secretary Lottonink have done a lot of really great work and Kratzius has done a lot of really great work to help launch the AI expert program and the State Department is very involved in helping it visualize that and actually onboard more users. We launched the chip concierge service that provides white glove support for the AI expert program for a subset of countries, which are the Paxillica countries. So we're certainly trying to contribute in ways that are constructive in ways that help be an accelerant to get more users. And then for securing our supply chains, that's where I think the government probably can play the most important role because we can catalyze partnerships. We can catalyze joint ventures. Our supply chains are very frail. What we have realized since COVID is obviously China has very vertically integrated supply chains because they have a lot of state-owned companies. Our supply chains are not vertically integrated at all. They are totally geographically dispersed. Thousands of different vendors many of them are mom and pop shops and different parts of the globe, full of information asymmetries, a company in Santa Clara or Santa Jose has no idea where some of these parts come from who makes them because there's so many different layers of vendor partnerships. And so as a result, you end up with a system where companies have very low visibility, there's a lot of inefficiencies by virtue of the sheer number of people involved and the lack of coordination. And so one of the things that we're looking at doing is actually using tools that the government can be a catalyst for to actually bring a little bit more efficiency and coordination to how the supply chains for our large, arguably systemically important companies work. And those tools include off-take agreements. They include joint ventures. They include co-investment opportunities. And the last one is market incentives. So this sort of gets at one of the concepts around creating some sort of preferential trade zone. We have obviously our administration has pursued an incentives-based agenda very heavily with respect to the imposition of tariffs to encourage investment in the US. But the government has tools to actually create, to incentivize behavior and create the environment that will actually lead the private sector to deploy capital in ways that are strategic. - Yeah. And by the way, on that note, what is the right way of making sense of tariffs that the effect of our since liberation day is that they were largely a negotiating tool that was effective in achieving its goals. Is it that it's also an economic policy that had certain goals? What's the right way of making sense of it while also realizing that it's still early? - Yeah, so the fundamental thesis and view is since over the last generation or so, we've lost six thousand manufacturing plans. We've lost over two million manufacturing jobs. Our industrial base has been completely hollowed out and not by an act of design, destined, manifest destiny, but really by choices, policy choices that have been instituted over the last 25 years that have been terrible. And so the tariffs are actually a recalibration of based on the recognition that we need to have an economic policy that is tethered to our national security imperatives. They're in pure and perfect markets. When you have a trade deficit, typically you have changes in the value of a currency that allows a trade deficit to go back to an equilibrium level of zero, but it is not a natural phenomenon to have consistent trade deficits for a 25 year period on a quarter by quarter basis. That is the result of markets that are not functioning and that are not functioning correctly. And so the tariffs actually helped correct those deficits, those long-term, secular trade deficits. And if I could digress and just really hammer this one point, when the president took office, we had a trade deficit with our top 12 trading partners of over $1 trillion a year. I mean, it was astronomical. And so we took the same view that that level of trade deficit on a year by year basis is completely unsustainable. After Liberation Day and when the president instituted tariffs, we've actually seen a complete change in the trajectory in a lot of these numbers. Our trade deficit was China, which has long been our largest trade deficit country has been plummeting in a way that's really historic. Our trade deficit overall has been receding. And we have actually seen record investment and capital investment in the US, which doesn't even show up in trade data yet because the capital investment hasn't, these things are early indicators. It hasn't yet converted into production capacity because a lot of these plants are being built as we speak. But as someone in the tech industry, you understand the difference between a lagging indicator and an early indicator. And so it's really encouraging and actually quite exciting to see a lot of these investments being made. And so the basic thesis is we used to have a policy worldview that said, potato chips, computer chips, doesn't matter. And you could have a national security untethered from economic policy. We believe that it's important to bring our economic policy in line with our national security. And that includes having an economic policy that does not run trillion dollar trade deficits on a year by year basis. And same more about with the positive benefit of having a much reduced trade deficit. Is it that on the employment side that will now employ more Americans because we're building more here? Is it that will also just take care of some of the industries that are most important for national security? Is it more benefits? Well, first of all, the Treasury will have more resources to actually fund a lot of the things Americans care about. So it has really, it has a very big impact on the resources that are disposal to actually fund the government as well as important programs. Second of all, it has a very big impact on fundamentally, one of the things that we realize during COVID is, it is incredibly, you can't defend a country with Facebook ads. You need to be able to make things. And the national security strategy that we released actually has a line that's very apt, which says the future will belong to builders. And that's our belief. What the tariffs do is they change the economics of building. They provide American builders with a reliveled playing field that makes it economical for them to compete if they build in America. And so, and fundamentally, that's why we've seen historic cat-backs investments because of the tariffs. The ingredients of the Trump industrial revolution are tariffs, deregulation, energy abundance, and tax incentives, which include cat-backs investing through the one big, beautiful bill. Those building blocks totally change the game when it comes to the unit economics of building things in America. And so, and so, it's actually in total contrast to what Europe is doing, sadly, where you have super high energy prices, insane regulation, insanely high taxes. And you see the results, which is a continued sustainment of de-investoralization of Europe, which is very regrettable. And so, we think these things are choices and you can actually reverse them. And we're seeing that play out in real time. There's some people that point to, hey, there's maybe some countries that are increasing their defense spending. Do you think that's for real? Do you think that's good for us? How should we make sense of what's happening here, Brian? So, I think the political will is real. But, and one of the things that I'm sure you and your viewers will appreciate, for folks in the technology industry is the political will can be there. But that, translating that political will into capabilities will require companies. It will require builders building technologies and weapons systems that are at least comparable or superior to what is already on the market. And in order to do that, you need founders who are exceptional, because building companies are very hard, especially hardware companies that are very advanced and capital intensive. And you need an environment that actually makes it somewhat conducive to actually build a company. And it's really hard for Europe to put those pieces together because the culture in Europe has been really not favorable to encouraging risk-taking, which is a necessary, which is the lifeblood of starting a business. You need to have a risk-taking culture. And then they have the Uniteconomics Farm where it's just really expensive to build stuff in Europe because your energy bills are really high. You have to spend a ton of money on legal fees before you even start. And so it's, you know, those parts are very challenging. So it's not that starting things in Europe will be impossible, but they have major headwinds. They will start on day one with significant disadvantages compared to American companies where, fortunately, because we have a federal system, the states are the laboratories of democracy. And so if some companies, many of whom we know, you know, have a hard time building things in one state, they can always move to a different state that has a favorable tax environment and regulatory environment. And so we actually have a system that, you know, it still makes it quite easy to build things. And so we, our policy is we want a strong Europe. And so we have really invested a lot of efforts into having a truthful candidate conversations with our friends in Europe to actually share our concerns. And, you know, the media, some elements of the legacy media has inaccurately portrayed our stance as being antagonistic towards Europe, but it'll be no surprise that that's a total misrepresentation. The truth is is that when you actually care about someone, you can actually be candid with them. Clear is kind. And so, you know, it is because you care if you have a relative who's unwell, is the empathetic view to pretend like they're perfectly fine and tell them everything's gonna be okay when you know that they have cancer or is the empathetic view to actually let them know that there's a problem that they need to take care of. I think the latter is clearly the more empathetic position. We care about Europe, we want them to be strong and so we are trying to draw their attention towards policy issues that we think are serious headwinds for them. - Yeah, similarly, some people critique and say, "Hey, it's better to have these countries dependent on us." Whereas, simply to your example, if you really care about someone, you want them to be sovereign, you want them to be independent, you want them to be strong and that doesn't mean that we're drifting apart and no longer gonna sport each other. It means that perhaps we could support each other better because you bring them to the table. - Exactly. And go for it. - I was just gonna say, and Americans are positive some thinkers. I mean, unlike, and this is fundamentally, you know, something that doesn't always, that people underappreciate overseas is, it's not like there is a central, essentially planned, maniacal plan in Washington to get the rest of the world dependent on America because we are a positive something, or is America, and what that means in practice is, Americans also do a lot of joint ventures, you know, because we don't see the world as a finite pie that needs to be subdivided into smaller slices. We see a world where the pie expands and where there is an opportunity, and a lot of American companies are very comfortable doing co-investments and joint ventures. And so it's just a very different lens. - Yeah. How should we make sense of the sovereign AI investments that are happening in the Middle East? Is that also strategic for us, or say more about that? - Yeah, so fundamentally, the Middle East has a lot of capital. They're going to invest it in artificial intelligence, one way or the other. They are, you know, some of the investors in the region are incredibly sophisticated, and a lot of them have rightly recognized that, they want an economy that is diversified away from hydrocarbons, and AI is fundamentally a path that allows them to shift from a paradigm where the export hydrocarbons to a world where they export intelligence and tokens. They are, as you know, a fundamental building block of training large computing capacity is cheap access to cheap energy. An integral function of the cost of compute is the cost of energy. And in the Middle East, you have a lot of very cheap energy. And so that is the comparative advantage that they have identified accurately and that they want to capitalize on. So our view is, let's work together to have a mutually beneficial approach. We want to enable, you know, because we have a positive, a positive some view of the world, we believe in a world where we help them achieve their aspirations that they've set for themselves in a way that actually grows the pie and offers our company's opportunities to collaborate on something greater. And so we see our approach to the Middle East as an opportunity to allow our companies to expand their lead because it's actually a market that gives our companies access to a lot of cheap energy, the UAE, the Qataris, the Saudis, you know, these relays, all of them are interested in AI. If they want to make available large compute capacity to our companies, we think that's a great thing. And so we've been having a lot of conversations with them, the UAE and Saudi Arabia, we signed an AI deal, we signed two large AI deals with them. I was in Israel where we signed a bilateral joint statement on technology cooperation, which we're very excited about, which included artificial intelligence. And so, you know, we're very supportive. It's a fascinating region because I think for a long time, the stereotype was that the Middle East was an unstable place and East Asia was an engine of growth. And part of what we've seen over the last 20 or so years is really that region emerge economically in a way that has caught a lot of people by surprise. So it's exciting to see a lot of the developments. - Yeah, absolutely. And to the end, you've talked about how if the 20th century ran on oil and steel, the 21st will run on compute and minerals that feed it. Why is the right analogy, you know, compute and oil? - Well, because oil and energy were fundamentally functions of economic growth and economic activity. And when you had the industrial revolution, the integration of industrialized energy, industrial scale energy production really led to a step change and a secular acceleration of economic growth. And part of what we are likely to see with artificial intelligence, and I actually co-authored a paper at the White House Economic Council on the great divergence is the super intelligence revolution I think is likely to lead to a secular acceleration and economic growth similar to what we saw during the first industrial revolution where you start to see the growth band of GDP growth, start to shift to the right, basically. So, you know, when we had societies that were primarily agrarian growth was between zero and 1% year by year. If you're between zero and 1%, that means you don't have compounding growth because you're under 1%. After the industrial revolution, we had between, you know, two and 3% of 1% and 3% roughly on a year by year basis, all of a sudden that unlocked compounding growth, that's when you start to see economic growth grow up exponentially like a hockey stick. And part of what we're likely to see with the AI revolution is actually a shift upwards in economic growth where you start to see growth pick up between three and, you know, five or maybe even six percent on a year by year basis, which is really a dramatic acceleration of growth. And the evidence that what evidence do we have to point to that? Well, we've already had growth break 5% in the US. And that's huge because the US is a world's largest economy. So when we're growing at 5%, that's like adding a GDP the size of Saudi Arabia to our economy, I mean, it's massive. And it's a massive movement and creation of wealth. And so that's number one. Number two, the other big indicator is we're seeing productivity break 5% growth in productivity. So if you believe that productivity is, you know, obviously gonna accelerate growth expansion, you're seeing that data. The other early indicators that we're seeing is we're seeing record demand, demand pick up for everything from energy, minerals, components, you know, compute capacity, all of that demand is massively going up. And so we're actually seeing an economy that's almost supply constrained where demand is just so intense, it's outpacing, you know, the rate of supply. And so to me, that shows an economy that's actually really running on all cylinders. And so it's quite exciting to see, you know, this economic acceleration. And it's funny because it really came at the perfect time where for a while the total factor productivity, statistics, you know, weren't super strong, you know, reducing, whereas the debt was also, you know, increasing, you know, local descendants. Let's go, you know, people were moving during COVID and people were saying, oh, you know, is the San Francisco, you know, what's happening there? And then of course, who happens to SF? Who said that? (laughing) Yeah, he's like, some good friends. But also, you know, just for the US in general, it seems like this is our godsend in terms of, you know, because people were saying, you know, during dojo, et cetera, like, you know, we really gotta take care of the debt. And there's just seems to be, you know, you can limit it to some extent, but there just seems to be this spending monsters. And maybe it's just easier to grow our way out of it. Yeah, I mean, it is true how, and it's such a quintessentially American story where, you know, we somehow figure things out at the 11th hour. COVID hits, I mean, it is true that San Francisco did not seem like it was an emergency. I love to, I can't see that. How I mean, we spend some time in Miami for a little bit. Yeah, it's totally good times. (laughing) And so, you know, a lot of people were very, very concerned about the explosion in the deficit and all the rest. And, you know, obviously, Redalio spent a lot of time talking about how we're in a death, you know, death spiral, et cetera. And, you know, what we've seen is the AI revolution catches the economy by storm. San Francisco makes this, you know, shocking come back. And, and, you know, hopefully it lit last. Yeah. And, you know, productivity picks up, growth starts to accelerate. And, and it's, it's actually really amazing because growth solves a lot of problems. I mean, it solves the death problem and solves, you know, a lot of problems. And this is actually something that the president really understands so well. I mean, one of the things that is, one of the things that I really took away from my time in tech is, you know, when you really great founders understand the importance and the power of prioritizing and isolating signal from noise. They understand how they're able to identify what are the one or two things that if I don't get right are decisive. And we are actually really lucky in a way to have a president that really has that kind of thinking and a lot of those attributes because he really understands that if the economy grows really fast, a lot of the other problems go away, just completely on their own. And so it's actually really amazing how how this has actually come together so quickly in a year. Totally. I mean, it's been one year, it's really amazing. Yeah. I want to get towards closing by going back to this idea of economic security that you talk about. I believe China in process like 90% of mineral stuff we were discussing and I'm curious is there a way of actually diversifying or what do you identify as the critical bottlenecks or risks that we have to solve? So I'm actually incredibly bullish that we're going to solve this. So the big picture is obviously as we mentioned earlier is we need to secure our supply chains which requires putting a lot of pieces of the puzzle together. It's very complicated. With that being said, we have the world's most talented founders. Unlike some places, we actually empower those founders. We don't set them hiding to Japan. And we're very creative. And look, we are a country from our earliest days where a country of underdogs. Where our country was started as an army of RACTAG rebels that overtook that rebels against an empire at the time. And that's just part of our DNA. So we tend to perform really well at the 11th hour. And I'm totally confident that we're actually going to solve this. The question is time is how fast can we actually turn, reverse 25 years, essentially? And we'll see. But I will say, President Trump's only been in power a year and just look at how different the world is. I mean, I think it's fair to say that it looks really different. And the running expression-- I mean, it's really one of the fascinating things about doing the role that I'm currently doing in the State Department is I really get a front row seat into having conversations with foreign leaders and getting an up close feel for our country, our administration. And the amount of admiration that they have for President Trump really can't be overstated. They know that he's a sample of one. And they really have the same kind of admiration that you would actually-- that people would typically be accustomed to feeling in the tech industry when they admire founders, because they kind of know that he's just such a leader and also just so unique, unlike anyone else. I mean, it's just like when people in the tech industry talked about Steve Jobs. He's kind of like this unique person that had a unique sense of attributes. And so that has actually been really, really striking. But people also want to work with America. I mean, people fundamentally like our country. And I think it's because subconsciously, I really do think it's because people know that we are positive sun thinkers. And so they inherently know that that is a less threatening worldview that working with America does not mean that it is zero sun. And so I think that's really a place to work on it. I'll make that my last question. Some people predicted that this sort of era of this administration would be in all forms kind of a retreat from globalization and weakening of our alliances. But in a PACSELCA represents a great global partnership and strengthening alliances. One of the things you talked about is how different countries bring their own specialization and superpower to the table and we're stronger together. One, you just share some examples of what that looks like. Yeah, so I would characterize, you know, you're right. Some people said it's the end of globalization, et cetera. I would obviously counter that. It's not the end of globalization per se. Globalization was a banner that was really used to describe a period of time where we engaged in unilateral trade concessions that were completely one sided with the rest of the world. And when people say globalization, what does that mean? That means a world where we were running massive trade deficits where we basically exported almost nothing while importing everything from the rest of the world. That was globalization. And so what we're doing is actually reengineering our trade architecture to have a world where we continue trading, but our trading architecture is on fundamentally more favorable terms that are fair and more reciprocal for American workers and American builders. That doesn't mean that we're against partnering people. We started Paxillica, the president started the board of peace. We actually welcome partnering with a lot of other countries who share our interests and our world view. So the idea that somehow revisiting a lot of the globalization era arrangements is a retreat from the world. I think is obviously completely false and very partisan. We're engaged in almost every place on Earth. And we just do it in a way that is fundamentally more favorable to our national security interests. Yeah, that's a great way to wrap. Jacob, thanks so much for coming on the podcast. Thanks for listening to this episode of the A16Z podcast. If you like this episode, be sure to like, comment, subscribe, leave us a rating or a review, and share it with your friends and family. For more episodes, go to YouTube, Apple Podcast, and Spotify. Follow us on X, and A16Z, and subscribe to our substack at a16z.substack.com. Thanks again for listening, and I'll see you with the next episode. As a reminder, the content here is for informational purposes only. Should not be taken as legal business, tax, or investment advice, or be used to evaluate any investment or security, and is not directed at any investors or potential investors in any A16Z fund. Please note that A16Z and its affiliates may also maintain investments in the company's disgust in this podcast. For more details, including a link to our investments, please see A16Z.com/disclosures.