CrowdStrike Founder: "China Has Planted Sleeper Cells In Our Power Grid!" | Dmitri Alperovitch

Instead of just breaking into companies or governments and stealing secrets, stealing IP, they launched this new campaign that the US government calls Vault Typhoon that is designed to penetrate organizations and networks where you have nothing worth stealing. So, small water utilities, port facilities, electric substations and power plants, and they get into those networks and they do nothing.

They just sit and wait.

Sleeper cells waiting for orders to come in and destroy that network in the event of conflict or in an attempt to deter us from intervening in a conflict. Now, this is what the military would call preparation of the battlefield. And this is one of the many reasons, one of the many signs that I look at and say this is now different. They're preparing for an invasion of Taiwan. You won't be able to get any microelectronics components for your MacBooks, for your PCs, for your servers, for your cars, for your refrigerators. Like everything around us is now dependent on chips. If there is any sort of conflict or blockade of Taiwan, you have $5 trillion of GDP wiped out in the first year.

And I think that's probably low.

In the last 100 years, the two countries that have seriously prepared for the invasion of Taiwan, obviously China, particularly in the last 45 years or so, and the United States that during World War II, when Taiwan was still under Japanese occupation, as we were hopping across the islands in the Pacific on the way to mainland Japan, one of the plans that was being proposed was to invade Formosa, as Taiwan was called at the time. And the US military even developed a war plan in August of 1944 for this invasion. What happened in June of 1944, the greatest military amphibious invasion in human history, Operation Overlord, the landing of the beaches in Normandy. The interesting thing is the similarities between the two, right? Taiwan Strait about the width of the English Channel, landing on the beaches that are well defended and obviously the Normandy invasion you could say even gave us great deal of hope that this can be successful, gave us experience to orchestrate something like this and nevertheless this plan that was put together was called Operation Causeway. You can look it up online. It's available. The Chinese are studying the heck out of it. And that plan called for double the number of forces that were used in Normandy, right? 50,000 Allied troops cross English Channel on June 6th, 1944. Over 300,000 were going to be used for Operation Causeway. It was going to be a two-pincer maneuver from mainland China and from the Philippines, which the Chinese today do not have the luxury of having. And they expected over 45,000 casualties in a brutal two-month campaign to take the island in 1944, only against 60,000 entrenched Japanese forces that were on the island at the time. So, just to give you a sense for where things are today, there are over 100,000 active duty forces in the Taiwanese military who are actually pretty good. A lot of disparagement that you hear about Taiwanese military. Most of that has to do with the reserves of which there are over a million. Their training is quite lacking, but the active duty people are quite good.

Around 25 million. So, it's quite significant. And by the way, just to give you a sense of casualties on D-Day, Allies lost 2,000 people, right? So 45,000 casualties for Taiwan. And ultimately Roosevelt MacArthur looked at this and said, 'No, this is too risky, too costly. We're going to bypass Formosa, we're going to go for Okinawa, which is what we ultimately did.' So for the United States that has just pulled off the greatest amphibious invasion in human history obviously with allies to say this is difficult in 1944 before long range anti-ship missiles, radar that was, you know, much more effective than it was back then, you know, all the sort of satellite imagery, etc. This is now infinitely more complex, infinitely more dangerous for the Chinese to pull off and something that they've never done before, unlike us. In fact, they have literally no one in the Chinese military in the PLA that has any combat experience whatsoever. The last person that wore the uniform in combat in 1979, the last time they fought a war, which was with Vietnam, which they suffered quite a number of defeats in that conflict and had to withdraw. That last person was on the central military committee up until recently and was just purged by Xi Jinping. So there was literally no one who has any combat experience of any seriousness in the Chinese military today. And something that is even more difficult to do now than it was back in 1944. So for people that are sort of dismissing this as a foregone conclusion that China will take Taiwan, look at the map, go visit Taiwan. This is one of the worst places possible to invade.

The terrain is incredibly difficult. First of all, Taiwan Strait itself is some of the most treacherous waters on the planet. There is fog. There is typhoons. Very, very rough waters. Most of the time, it's not even possible to invade it. There are only a few times a year when an invasion, amphibious invasion might actually take place. Once you land, you've got very shallow waters. And the beaches themselves are very rocky, very narrow. They're surrounded by mountains from which you can fire control the beaches, put up artillery and decimate the invading force. And then again, you're up against mountains. There's nowhere to go. So, you have to get around those mountains. You have to go through tunnels, bridges that can be blown up over rivers to get to the city of Taipei. So, for you to actually get to the center of power of Taiwan is incredibly difficult even if you manage a landing. So that is why China has been thinking about this for decades, hasn't yet pulled it off, is not yet ready to pull it off, even though Xi Jinping has given an order to the military to be ready to invade Taiwan by 2027, which is next year. They're nowhere close to being ready today. And I think he is appreciating this and one of the reasons why you're seeing all these purges across the military, all these generals that he's purging is because he knows that they're not ready. There's a lot of corruption that's taking place, money being diverted. But it's also an incredibly difficult challenge to pull off and the stakes for him are existential. This is one of the things that I think most people don't appreciate. If you think about what a failure of invasion of Taiwan means for Xi Jinping and for China, it probably means that the Chinese fleet, Chinese navy is at the bottom of the ocean. Hundreds of thousands of casualties potentially, strikes on mainland China. This scenario is probably something that he personally cannot survive as leader of the Chinese Communist Party and the Chinese Communist Party may not survive if they suffer this enormous defeat. So for him to go and to give an order to go he has to have a great deal of certainty in the success which I don't think he has today.

Well, that remains to be seen of how the ending goes, right. So you could say, you know, Putin was seen pre-2022 as a leader that, you know, took Russia out of the miseries of 1990s and rebuilt it and rebuilt its military. But look today and how it's going is certainly not commensurate with how he wants to be seen. For Xi Jinping, I think he personally sees himself as another great leader of China. Someone who I think aspires to be greater than Mao. In fact, when you look at the power he's accumulated in the Chinese political system, it is truly unprecedented in modern times. Mao did not have this level of power that Xi Jinping does now except maybe for a brief moment during the cultural revolution. It is his decision, his alone. He's completely consolidated power across the foreign policy apparatus, across the military. So if he decides to go, there is no one that can stop him. And that is something that is very different. He's in fact a new emperor of China. But in terms of what he's achieved for China and how he'd be seen in the long arc of history, it could change on this entire.

And in fact, you know, I don't think he's got much of a record. The economy is not doing well. And growth is stagnating. In fact, there are some estimates that the growth in the United States, real growth, is higher than it is in China, right? And the United States economy is 25% greater in terms of nominal GDP. So just doing the math, they're never going to catch up unless something dramatic happens here or there.

He's done some things on making sure that the supply chain independence is greater today than it was before he came into office. Of course he put together this plan in 2015, Made in China 2025, where he said we have to achieve independence in semiconductors and aerospace and a bunch of other areas. They made progress. They did not get there by 2025, but he's still pouring money and effort into it. So, he's doing some things to prepare himself for confrontation with the United States and the broader West economically and obviously militarily. But, you know, in terms of achieving greatness for China, you know, their economy is greater today than it was before he took office in 2012. But I don't think you can attribute that to him versus the trajectory that has been in place really starting with Deng Xiaoping in the 1970s.

I don't think it's inevitable. He wants to do it. I'm convinced he wants to do it. I think he wants to do it on his watch, which makes him very different from other Chinese leaders. Right? Every Chinese leader since Mao had wanted to invade Taiwan and unify it. But for them, it was a problem that was far off that one day we know it's going to happen. It doesn't need to happen now. It doesn't need to happen on our watch. I think Xi Jinping came into office in 2012 and said, 'We've waited long enough.' And he even said publicly, we can't pass this problem off to future generations. It's gone long enough. We need to resolve it. And he's invested massively in accelerating the military modernization, the independence on the supply chains to withstand sanctions and other economic blockade options from the United States and its allies. And I think he's gearing up to do this on his watch. And I think that everyone's focused on 2027 because of that order he had given. I think the much more important date is actually 2032.

Why? Because 2027, he's up for re-election as a Chinese Communist Party leader. Despite the fact that it's not a democracy, it is an election. It is an election of the Communist Party elites. Now, he's consolidated power almost certainly, he'll get re-elected in 2027. And I think part of the reason he's given that order to be ready for the military for invasion of Taiwan by 2027 is not that he wants to invade in 2027, but that he wants to go in front of the party congress in the fall of that year next year and announce that Xi Jinping for the first time in our history has built an invasion force capable of doing this right and get credit for it. It's also a very symbolic year because it's the 100th anniversary of the establishment of the People's Liberation Army, the Chinese military. It's actually fascinating because the Chinese military predates the establishment of the People's Republic of China because it is one of the few countries where the military is not the military of the state. It's the military of the Communist Party. So the Communist Party gets established in 1920s in China, creates a military arm that becomes the PLA with the establishment of China in the 1940s. So establishment of the Communist China. So 2032 is going to be another re-election year. He gets a 5-year term. In 2032, he's 79. And we've just seen in the US what happens to leaders that try to get elected in their 80s. You know, mental issues, health concerns can intervene. Who knows what his hold on power will be at that time, particularly if the economy continues to not do well. There's a lot of sort of latent opposition to him and his policies within the party. No one is willing to raise their neck right now and because they're afraid of him but that could change in the next four or five years and I think he's thinking that you know if I'm going to do this I need to do this before 2032 so that is the date that I'm concerned about.

Absolutely. Look, AI is fundamental to everything, right? It's fundamental to economic growth. It's fundamental for military technologies. We're seeing this on the battlefield in Ukraine, even now in Iran. And the amazing thing about AI is that we have all the levers to control it unlike almost any other technology that I can think of in human history because we design all the chips. We have all the leading AI companies here in America. Yes, there are Chinese companies too that are trying to catch up but mostly they're catching up by distilling from our own companies which is theft and using our own chips to train their models using that stolen intellectual property. So we have unbelievable power to cut them off at their knees and dramatically restrict their progress. It's not that they won't be able to build any models. Obviously you can't stop everything but you can curtail their progress and even more importantly you could curtail their ability to do inference to run those models. Right now because of the shortage of chips, despite the fact that our export controls are full of holes like Swiss cheese even then they still have huge constraints on compute and they have to decide do we allocate this compute to running these models, do we allocate it to train these models. In fact when Kimi came out last month, within hours the Kimi cloud went down because all demand which is not even commensurate with the demand that Anthropic or OpenAI sees on any normal day and they couldn't handle it because they didn't have enough chips to run those models, right?

Give me a wait list. Yeah. So, this is a real issue. We can make it so much worse because one, we allow selling some chips to China like the H20s and H200s which you know you get in enough quantities, you can accumulate enough compute to train really good models. So, it's not just about the very best chips. And two, there is no restrictions today on remote access to compute. So, if you've got a data center in Malaysia that's full of Nvidia Rubins, let's say the new chip that Nvidia just announced, there's no restriction on Chinese companies training there. Why not?

Now, there are a bunch of bills in Congress right now, including a bill on this remote access to try to restrict it. But we've just been very haphazard and have not made it a policy to restrict China's ability to advance in AI. It is one of the greatest mistakes I think the United States has been making vis-a-vis China in the last 5 years because you're absolutely right it is everything it is military advantage, it is economic advantage, it is productivity and we need to do whatever it takes to win this race and you win the race in two ways right you run faster which we obviously need to continue to do to invest in this technology but you also kneecap the other guy.

Well, if you have the best models, and today we do have the best models, right? For all the concerns around Kimi and Qwen that you see today, we need to appreciate the Chinese are still 6 to 9 months behind and they've been perpetually six or nine months behind ever since the release of DeepSeek what almost a year and a half ago. And I'll give you one example, right? So the latest Kimi model that everyone's so concerned about is nowhere close to Mythos. Mythos was trained in February of this year.

Right. 6 months ago. So, and they're not even close to that. So.

They can't access Mythos. Although, Fable, you know, is a reduced version of Mythos that is widely available. And you know, obviously there are reports that US government has access to Mythos, other companies have access to Mythos. So, it is absolutely serving us advantage today.

They're not, and they should be, right? We've talked about how difficult this invasion is to orchestrate, but that requires a capable defense force. And unfortunately, the Taiwanese for the last really 47 years have not been focused on their own defense. It's starting to change, but we are actually the ones to blame for this situation, believe it or not.

Right? Because in 1979 when Jimmy Carter decided to recognize China, mainland China, and remove recognition from Taiwan and remove all of the US military forces that were on the island, over 10,000 forces that were there, we've basically cut off all relations with the Taiwanese military and everyone else did the same. So for the last 47 years, they've basically been stuck on another planet, completely isolated in a bubble. And now we're coming to them and saying, 'Why don't you have a modern military doctrine, modern training? Why aren't you thinking about this in, you know, the way that warfare is now fought these days with focus on unmanned vehicles and long range strike and all the things you're seeing in Ukraine?' Well, because they have no opportunities to learn from those lessons. They can't go to Ukraine. They can't send their military people to Ukraine. They don't have diplomatic relations with Ukraine. They can't send them to US military academies or to Sandhurst in the UK. They're completely isolated and you know we are the reason for that because we decided to remove recognition from it and now we're asking them to do all these things to get better and they're starting to change. You know I'm going to be in Taiwan watching their military exercise shortly. It is much more impressive today than it was even a few years ago. But they still have such a long way to go because they're starting basically from scratch.

Well, wars are always unpredictable. So for anyone that thinks this is going to be a short war, you know, history is littered with examples when that proves not to be the case, including for great powers, including for large powers that think that they can subdue a tiny actor. You know, look at the Iran conflict this year and the challenges we're facing just trying to open the Strait of Hormuz against an actor who has no navy, no air force, right, and is still able to humble us in some ways. So, this is doable, but Taiwan needs to do a lot more to defend itself and focus more on modern military doctrine. So, they're still focused on old technologies. They're investing heavily in M1 Abrams tanks. There's not going to be a tank battle in Taiwan. You know, if you go visit Taiwan, it's an urban jungle. So you're not going to be fighting these massive, you know, Kursk-style battles of tanks on the island. And by the way, if China actually lands enough forces on the island for you to have a tank battle, it's over because you can't resupply Taiwan. There's not going to be an insurgency on Taiwan. You know, for an insurgency to take place, you need a resupply route. There is no resupplying of Taiwan. The Chinese are going to be very brutal in suppressing the population just like the Russians are in occupied Ukraine. There's virtually no insurgency in occupied Ukraine. Not because the Ukrainians don't want to resist the Russians, but because they know that the Russians are going to come into a house as a suspected insurgent, rape the wife, steal the children, and torture the husband to death, right? And they do this routinely. So, there's no appetite to be an insurgent in those types of circumstances. I don't think the Chinese are going to be any different.

I think the key thing to the strategy is to sink the fleet. You know, if you sink the Chinese fleet and they can't get over a million troops that they need to get on the island to pacify, it's over.

They have, it's the largest military on the planet. So, they have millions of troops under arms. But to get them there, you need enormous lift capacity, which they don't even have in the Chinese Navy. They need to use civilian vessels to augment it. And that's why they've been giving these requirements to container vessels and these roll-on roll-off ships that are used for commercial purposes like to transport cars for example. They've been giving them military specifications so that they can in the event of conflict borrow the ships for this transit but they still don't have enough. They're building rapidly so it's not something to be complacent about. But you know for people that see a tiny exercise with China encircling Taiwan with a few dozen warships and saying this is preparation for invasion. My answer is relax. You will know when it's an invasion because this is going to be the biggest military operation in human history.

Well, not all of us. In fact, you know, I was one of the early ones that said back in December of 2021 that this is going to happen. Got criticized by people saying I'm a warmonger and this is never going to happen.

In fact, I remember I was at Munich Security Conference the weekend before the war broke out talking to senior people in German government, German intelligence who are saying this is never going to happen the weekend before. So, there were plenty of people that were incredibly complacent about this. But you're right, to me this was completely obvious months beforehand. We, it's going to be even more clear with Taiwan. Obviously, there's no question that they want it. There's no question that Xi Jinping believes that he has the rationale for invasion because you know to him Taiwan is just a Chinese territory and it's an internal matter. But you will see mobilization of such an extent even just in the open source just through satellite imagery that's going to show up that everyone's going to know months in advance that this is coming.

Yeah. So you know there's all these questions will they blockade Taiwan. I don't believe that a blockade alone can win the victory. In fact there's no modern history of successes of blockades. You can look at the siege of Stalingrad during World War II, the siege of Sarajevo in 1990s. Cities don't even surrender to blockades much less countries of 25 million people. But in terms of phase one to an invasion, they absolutely have to blockade it to prevent resupply from the US.

Well, let me just say this. If it doesn't, Taiwan will fall.

And the beauty of our position here is obviously we have a policy of ambiguity. We don't say one way or another whether we would defend Taiwan. And for China, the dilemma is that even if they think there's a 1% chance that the US might be in, they have to treat us as a 100%. Yes. Right. If they were certain that we would not get involved, they would invade tomorrow because the Taiwanese are so unready for this. But because they can't count us out, they have to prepare for a fight with the United States and not just the United States, but Japan who views this as existential. Australia very likely because they're also incredibly concerned about Chinese expansionism in the region. That's a very very tough proposition.

Well, we wouldn't say go ahead ever.

But we may not fight and you know this is a very tough call, right? A war with China. You know, we've never fought a great power directly. We've never fought a nuclear power. Right. So the idea that a conflict with China would stay conventional without escalating to a nuclear level is not a guarantee. So this is a very very big decision for whoever is president at that time. Hopefully it never happens. But you know I think it's at best a 50/50 proposition.

No 50/50 that the US would get involved.

No. Again, I don't think there's an inevitability. Xi Jinping wants to do it. But if he can get himself certain of success, he won't do it. And we have to do whatever it takes to create more dilemmas for him and create more doubt that he can be successful.

In Taiwan.

Yeah. So we have a policy of not obviously putting overt military capabilities on Taiwan as part of the deal with China that we made in the 1970s when we withdrew forces. China would see it as a red line. What that means is a question because if they can't invade today, you know, what a red line means in practice is a good question, but it will probably be a break in diplomatic relations, economic relations with China if we were to do that.

Well, I've always been focused on geopolitics. I actually studied in school. But my two passions obviously I started a cybersecurity company, CrowdStrike. Really came together in 2010. And it actually is what led me to start CrowdStrike because I was in another cybersecurity company McAfee which is one of the giants of the field at the time and I get a call.

No. Oh, that predated me by a couple of decades. So, heard lots of stories but never met him.

But in January 2010, I get a call from this little company you may have heard of called Google that wanted to cooperate with me on an investigation into a hack of their systems that ended up being from China. And it was the first time when the world once it got announced really started to understand that cyber is not just some criminals in a basement stealing credit cards that this is nation state activity, militaries, intelligence services doing this. Google was the first one and then once I finished that investigation I named it Aurora. I asked the question of is this a one-off or is this happening all over the place and we just don't know about it. And over the course of the next year and a half ended up discovering that this was happening on a massive scale. And I coined this phrase at the time that there are two types of companies, those that know that they've been hacked and those that don't yet know, but they've all been hacked. And that this is the greatest transfer of wealth in history that's taken place in terms of the IP theft committed in cyber by China. And most companies are not even aware of it. And that directly led to the founding of CrowdStrike to try to stop it. But it also sort of intertwined these two passions of mine of technology of cyber with geopolitics because I realized that this is not just about China breaking in and stealing some IP for you know pain from a chemical manufacturer or a defense contractor. This is about domination. This is about undermining the collective West's economic power. This is about building up their own military for fighting against us. And the stakes were much bigger than just cyber. And over time, I realized that cyber is actually one of those canary in the coal mines that can predict geopolitical conflicts. And one of the things that got me convinced very early on in 2021 that Russia was going to invade was actually what I was seeing in cyber domain, which was very aggressive tactics in Ukraine by the Russians focused on things that would only be useful in an invasion scenario. So, you know, nation states use cyber because it's a covert, deniable tool, but still advances their national security priorities. And if you're paying attention to what they're doing, you can tell what their overall strategy is outside of cyber.

China is by far the most aggressive actor. They have hundreds of thousands of people focused on cyber intrusion activity. It's a whole ecosystem. They have the PLA, their military doing this. In fact, they stood up cyber command in 2015 to replicate what we've done with our own cyber command to do this on an industrial scale. They have their MSS, Ministry of State Security, their version basically kind of a combination of FBI and CIA equivalent doing this. And then they have a huge number of independent contractors within the Chinese ecosystem. They're basically private companies that are doing contracts hacking for hire for the Chinese government. So the scale of this is just absolutely enormous.

Well that was the goal. So this activity started people believe probably late 90s early 2000s long before anyone was really paying attention. So it's hard to go back and understand exactly how early it was but certainly by the early 2000s this was all going on at full speed and only kind of snowballed from there.

Something changed around 2020 and that is that instead of just breaking into companies or governments and stealing secrets, stealing IP, they also started doing something else. They're still stealing on an industrial scale, but they launched this new campaign that the US government calls Vault Typhoon that is designed to penetrate organizations and networks where you have nothing worth stealing. So, small water utilities, port facilities, electric substations and power plants, and they get into those networks and they do nothing.

They just sit and wait.

Sleeper cells waiting for orders to come in and destroy that network in the event of conflict or in an attempt to deter us from intervening in a conflict. So very very different from what they've been doing for 20 years which was just theft theft theft steal as much as possible to help your domestic industry to help your military. Now this is what the military would call preparation of the battlefield. And this is one of the many reasons, one of the many signs that I look at and say this is now different. They're preparing for an invasion of Taiwan.

Yeah, it's exactly what it is, right? So at the end of the day, it's essentially malware, malicious software that they install. One of the things that they and other actors have been doing in the last seven or eight years now is actually moving away from developing custom code to commit their malicious activity to using code that's already on the system. So turns out, you know, in a Windows operating system, there's a lot of software already that can get you 99% of the way there in terms of maintaining persistence on a system, doing what you need to transfer files in and out. So if you don't have to bring in foreign code that could be easily detected by security systems, you can use what's already there, sort of living off the land is what the technique is called. That is much better for stealthiness and remaining as essentially an insider. So if they can steal passwords, credentials from a legitimate user and pretend to be them, detecting that is much much harder.

They have been certainly not at the same time. Yeah.

You know, companies detect it, they kick them out, a year later they may come back. But you know virtually every company that has anything worthwhile to the Chinese has been targeted by them.

Yeah, different. Right. So obviously we have the NSA, National Security Agency that is the primary signals intelligence agency in the US and signals intelligence now means not just let's listen to radio waves and collect things from satellites. It means let's go into computers through hacking and collect data. So we're doing this at a significant scale, but we're doing it for national security purposes. So, we're breaking into government networks of foreign countries that are adversaries to collect intelligence to ultimately brief the president and policy makers on what's happening. What we don't do is we don't break into companies, steal their IP and then go to our domestic companies and say, 'Here you go.' We don't do that. I've argued particularly in recent years that we should because at this point we failed to completely deter China from doing this. We failed to set the norms. So if we can't beat them, let's join them because guess what? We're so much better than they are. So if we were to do it at scale to them, that would be a huge wakeup call because we would do it much more aggressively and much better.

Yeah. Well, yeah. So to do sort of the preparation of the battlefield to them, the Vault Typhoon style activity, you know, we don't know if US government is doing that or not. Hopefully they are. Because that would make sense.

I don't think they're a brother. So, look, you know, in my book World on the Brink, I write that this is a new cold war. And actually before I started writing the book, I thought that this cold war, second cold war is very different from the first one. The more I spent time researching the book and writing it, the more I realized it's almost exactly the same. If you look at the confrontation with the Soviet Union during the first cold war and the confrontation with China, the similarities are remarkable. You have a global competition for supremacy playing out in every single corner of the world, right? It's diplomatic. It's economic. It's geopolitical. It's military. It's not just in Asia. It's in Africa. It's in Latin America. It's in Russia and Europe. It's everywhere. You have an arms race. Both countries building up their militaries. President Trump announced this year that he wants a $1.5 trillion budget. Why? Not to fight Iran, for China, right? China obviously has been going through an enormous military transformation the last 40 years. You have a nuclear arms race for the very first time. The Chinese under Xi Jinping have started to build up their nuclear arsenal. So since they got the bomb in the 1960s, they had a stable number of about 300 warheads which is plenty to destroy the United States and they felt that that was sufficient. Xi Jinping comes into office in 2012 says no we want parity with the United States. We want 1500 warheads. They're now approximately at 600. So they've more than doubled under his watch. And the US intelligence community believes that they're going to match our numbers in the 2030s. And they're building more silos, more ballistic missiles. You don't need that many missiles to destroy the United States, but you know, just like in the first cold war, you got to match missile for missile, warhead for warhead. The exact same dynamic is playing out. You have a tech war, you have an economic war that's taking place with tariffs obviously for the last really since the first Trump administration. And you have an ideological struggle and you have a race for allies and bases both in the region and across the world. So on almost every level this is the exact same confrontation. Now there is a difference in that you have this economic interdependency that exists by the way there was not zero economic interdependency with the Soviet Union. In fact, there was a lot of trade and you know the Soviet Union particularly in the 1960s and 70s started to produce a lot of gas and oil that was being sold to the west in exchange they were getting food. So the scale is different but the similarities are still there but of course now we're in the midst of decoupling. So those economic interdependencies are falling by the wayside on both sides as China wants independence and we are waking up to the same threat with critical minerals with semiconductor dependencies and other microelectronics dependencies on China.

Well, we actually, by the way, I forgot about this. We actually do have a space race, a moon race. We're trying to get to the moon before the Chinese do, right? If that is not sort of the iconic picture of the first cold war, I don't know what else is, but winning the AI race is arguably way more important than getting to the moon, the first time or the second time because, you know, we got to the moon, we planted the flag, we shot some golf balls, and we left. It helped with some of our space industry development but it wasn't critical to winning the cold war and the cold war did not end obviously in 1969 when we got to the moon. But winning the race is essential for economic dominance and military dominance. There's no question in my mind about that.

We have to win. We have to as much as possible curtail Chinese advancements. You know I said they're 6 to 9 months behind. That's way too short of a time period, right? I want this to be years, not six to nine months.

Well, I actually think that's an advantage. So, you know, despite the fact that I'm a big China hawk, I'm not one of those people that says, 'Let's ban the Chinese models.' In fact, I use them personally. Now I never use them on Chinese clouds. I never want them to have my data, but I can download them to my systems and use them.

Well, that's that I think is a problem. If you want to share all your data with China, you know, you have it coming.

Exactly. So, so that's a huge problem. I don't like people doing that. But if you want to use it on a US cloud provider on Amazon or Microsoft that offers those models, I don't think there's anything wrong with that. And again like this is people are very surprised when they hear me say this because I'm a big China hawk. But the hypothetical I give them is let's say that the Chinese models there were no open weights Chinese models. It was all closed weights just like Anthropic and OpenAI and the US government let's say went in and did what China usually does which is break into those companies steal those models and then publishes it for the world to use. Tell me how that would benefit China. I don't think anyone would make an argument that it would benefit China just like if they did that to Anthropic or OpenAI would say what a disaster. And here we don't have to do that. They're doing it to themselves, right? Particularly if they're not getting any benefit from it, right? I'm not in favor of paying them. I'm not in favor of giving them data. But if they're giving the world these open-weight models that we can use cheaper and to use them in circumstances where you can't use OpenAI or Anthropic, if you want to load it on a drone, for example, it's too big of a model and they won't let you do it. Great let's do that. I don't think that they will continue doing it for long because even China cannot afford to spend hundreds of billions of dollars on chips and infrastructure to give it away right at some point the dynamics of the market will be such that they'll go close weights. In fact you've seen this in the US where Meta went close weights, Qwen one of their models went close weights in China. I would be surprised if in a couple years there are any open-weight models from leading Chinese AI companies because it's just not sustainable.

Well, I don't think you're giving it.

Well, yeah, that I think we should try to curtail and stop. And I think we should actually sanction the Chinese companies for nothing else. If we have evidence that they're stealing IP from Anthropic and OpenAI and Google, we should absolutely sanction them for doing that, for this massive illicit distillation that they're engaged in. But look, at the end of the day, the reason why I'm not concerned today about this in terms of impact to American companies is you look at Anthropic, it's the fastest growing company in the history of humanity.

While Chinese progress was taking place. So clearly they're not hurting. You know, you could argue maybe they'd be at $100 billion if it weren't for them. But, you know, I'm not shedding tears over the growth rates. So if that starts to change, you know, I may adjust my views. But for now, it seems like there's plenty of space for both American closed-weight companies, Chinese open-weight companies. And look, I'd love for there to be more American open-weight companies as well. I think that would be great for innovation. It's a very difficult economic model, I think, to sustain to basically say we're going to spend tens of billions of dollars, maybe hundreds of billions of dollars building something and giving it away for free.

Yeah. Well, they won't invade in '27. I'm very confident of that. Maybe by 2032, but hopefully never.

I don't think it'll be over. I think there will, you know, AI progress will continue. So, this is not something where you've gotten a great model and you say we're done.

Yeah. So, you're talking about recursive self-improvement, RSI, where AI sort of improves itself. The challenge with RSI is you still have the constraint of compute and until AI starts, you know, putting out its own chips, which is a little more difficult, you know, maybe one day we get there with robotics but today at least it seems like we're pretty far off from that. You're not going to be able to really get the full benefits of RSI if for no other reason that you're still going to be computed on inference on being able to run those models. So that's still important and that's where Taiwan comes in, right? Because TSMC is an American producer of all these AI chips whether it's Nvidia, AMD, Anthropic's chips, Maia from Meta and Microsoft. And while TSMC has some production in America and is actually under the Trump administration committed to spending $250 billion to increase that significantly, you're still going to have a huge dependence on Taiwan. And it's not just the AI chips themselves, it's high bandwidth memory that also gets produced, a lot of it on Taiwan. There's some in the US, there's some in Korea as well. But Taiwan is just such a center ecosystem of semiconductor production. That if Taiwan goes, if there's a war over Taiwan, a lot of that production will be destroyed just in the nature of conflict, right? These chip facilities are very, very sensitive to earthquakes, to any sort of disruption. Power outages can destroy the entire line of chips. That will create vast economic devastation, trillions of dollars of impact not just to the AI industry but to all of computing, right? Today, you know, Apple is raising prices on MacBooks because you can't get memory. You won't be able to get any microelectronics components for your MacBooks, for your PCs, for your servers, for your cars, for your refrigerators. Like everything around us is now dependent on chips.

So the estimates are about 90% of advanced chips go through Taiwan. Over 40% of all chips period go through Taiwan. So sometimes we overly focus on the advanced chips. They're very, very important but they're only useful really for a few purposes. They're useful for AI chips. They're useful for, you know, iPhones and latest MacBooks and some military applications. But everything else is these what I call foundational chips. So even if you look at this iPad you have in front of you, there are basically three advanced chips in that iPad. There's a processor from Apple, there's a memory probably from Micron or someone like that. And then there's a bandwidth modem probably from Qualcomm if it has one like an iPhone. But there's about 150, 160 foundational chips in there. Everything from this camera to the Wi-Fi to the GPS to the accelerometer to the power supply to the battery. All of those things without which you don't have an iPad or an iPhone, right? They're incredibly vital. The analogy I often use is foundational chips are sort of like steel and aluminum, right? Yeah, we have new materials like carbon fiber. You can build a Boeing Dreamliner out of them, but even a Boeing Dreamliner, that's a carbon fiber plane. 30% of that is steel and aluminum. You don't have steel and aluminum, there's no Boeing Dreamliner or anything, right? Those foundational materials are vital for everything that is around us and foundational chips are that for everything. Even you think about AI servers. They need power. They need networking. Those are foundational chips.

Yeah. So, we passed the CHIPS Act in 2022 that allocated about $52 billion in grants to companies. The Biden administration waited way too long to give that money out. You know, bill passed in 2022. Most of the money didn't go out till end of the Biden administration in 2024. And by then, you know, I was at a conference with Pat Gelsinger, CEO of Intel, in 2023, and he told me, Dmitri, if I don't get that money now, this quarter, Intel may not survive. He didn't get the money that quarter. He didn't get it the next quarter. He didn't get it until the end of the Biden administration. And Intel ended up being a disaster until the Trump administration came in and obviously took an equity stake, and they're doing better now. But it was a huge problem, massive wasted opportunity. The Trump administration actually took a very different approach where they said we don't want to give money to those companies. We'll take equity stakes and we'll give them other advantages as a result of those equity stakes. But we're also with foreign partners like TSMC and Samsung. We use the power of tariffs to browbeat them into investing more in the United States. So the Biden administration said to TSMC, we'll give you billions of dollars. Please, please, please come to Arizona. And they did.

They are operational. They're producing chips today. Trump administration said, Taiwan, we're going to raise tariffs massively on you unless you invest more. And TSMC committed to investing $250 billion in Arizona, committing to build six fabs there over the next five years or so. If that all comes to fruition, we will basically produce half of the chips we need in America. You're still going to have huge dependencies on Taiwan. Part of the problem with relying on TSMC in Arizona is that most of the expertise is still back home in Taiwan, the process engineers and so forth. So without them, you could have very, very expensive warehouses basically with equipment that, you know, you can operate existing production lines but maybe not optimize them. So it's a fallacy to think that just because we've increased production in the US, we no longer need Taiwan. Taiwan will be relevant for the semiconductor industry for the foreseeable future.