# Alex Karp: Palantir's Lethal AI & The Truth About LLMs | Full CNBC Interview

*Vampyre Drakul* · https://www.youtube.com/watch?v=xe84cwQJKaE

## Summary

Palantir co-founder and CEO Alex Karp sits down with CNBC to discuss the company's Maven platform, which he says makes LLMs lethal and useful on the battlefield, and Palantir's role in powering Western militaries. He defends Palantir against Wall Street narratives that LLM companies could replicate its business, criticizes frontier labs as a hyper religion of optimism, and warns that AI-driven dislocation will reshape American politics and jobs. The conversation closes with his views on Anthropic's Dario Amodei, health care, and the SpaceX IPO.

## Key points

- Karp says the Palantir Maven platform takes valuable LLMs and makes them lethal and useful on the battlefield, with the company's primary focus on bringing American war fighters home safely.
- He argues America's recent precision operations, such as against Iran and in Venezuela, would not have been possible two or three years ago.
- Karp describes three kinds of software: hard-coded on-prem infrastructure for data sovereignty, management tools like ontology and Foundry, and forward deployed engineers plus large language models.
- He warns that AI cannot be discussed in blue-red political terms and that "nationalization" of the debate is coming.
- Karp says the issue is not mass job loss but mass retraining, and that America's dynamism is its biggest advantage.
- He is dismissive of frontier labs' enterprise approach, saying enterprises come to Palantir screaming that LLM companies don't understand or care about their needs.
- Karp praises Dario Amodei as a very important person whose models run on Palantir, while disagreeing with his belief that all problems are transitionary to a perfect future.
- On SpaceX's IPO, Karp says he roots for every entrepreneur, is bullish on Elon in space, but has no idea about space in general.

## Notes

### Palantir and the Military
- Alex Karp explains that when war zones heat up in the allied world, the **Palantir Maven platform** comes into action, taking valuable LLMs and making them "lethal and useful on the battlefield."
- The company's primary focus is bringing American war fighters home safely, giving them a "massively unfair advantage" over adversaries.
- Karp claims recent American precision operations, including against Iran and in Venezuela, were not possible two or three years ago.

### Three Kinds of Software
- **Infrastructure**: hard-coded, on-prem systems that preserve data sovereignty — "excruciatingly valuable."
- **Management tools**: ontology, Foundry and similar — very valuable and durable.
- **FDEs and LLMs**: some valuable, some worthless; most FDEs are "just throwing people at a problem."

### Politics and Dislocation
- Karp insists AI infrastructure cannot be discussed in blue-red terms and warns that "nationalization" of the AI debate is coming.
- He clarifies he was not advocating a draft, but describing a massive time of dislocation requiring cultural unity.
- He calls himself a "card-carrying progressive" and argues wealth accumulation and job dislocation are real political problems that must be addressed.

### On LLMs and Frontier Labs
- Karp describes frontier labs as a "hyper religion of hyper optimism" that believes all problems will be solved, including human nature.
- Enterprises are fed up; he says Palantir debates paying people to leave LLM companies before joining them.
- He praises Dario Amodei as "the frontier person" who has gone from behind to ahead, but disagrees that all problems are transitionary to a perfect future.
- LLMs are probabilistic and useful where 51% accuracy suffices, but "that stuff doesn't ship" for missiles, rockets and cars.

### Health, Security and Space
- Detecting a deadly pathogen resembles detecting a truck moved by 2 cm before an attack — both require software infrastructure, ontology and sovereign data handling at the millimeter level.
- On SpaceX's IPO, Karp roots for every entrepreneur, especially Elon, and is bullish on Elon in space, but "bullish on space? No idea."

## Transcript

[00:00] joins us this morning with Palantir's co-founder [music] and CEO Alex Karp. Morning, Sara. >> [laughter] >> Good morning, Carl. It's good to see you. Yes, I am here at Palantir with the founder and CEO Alex Karp. It's good to have you. Good to speak with you. >> Very nice to be here. >> And and Carl was just talking about the Iran war risk, which is influencing the market today and the fact that we haven't gotten a deal yet. I know that you are heavily involved AI operating system for the military. What what is your view on on where the war goes from here? >> Um well, first of all, you know, at Palantir, you know, people currently focusing a lot on our commercial business and there's a lot of interesting things

[00:39] there, but once uh war zones uh heat up anywhere in the allied world, uh the Palantir Maven platform, which um takes very valuable LLMs and makes them actually lethal and uh useful on the battlefield, come into action and then the primary focus of our whole company is bringing our American uh war fighters home safely, meaning that when we engage that they engage in a way that it gives them a massively unfair advantage, uh scares our adversaries, and make sure that when they're in action, they have an a the best humanly and inhumanly possible chance of of coming home safe and without injury. And so, the whole company really just focuses on the operations of that. And um obviously, these things begin

[01:30] well before uh the public sees uh and we have a wide swath of of Palantirians working on various platforms, particularly the Maven platform. And then, there are a lot of things that are in the public space. So, Maven has been kind of discussed and kind of older versions of Maven have been shown in public, but um I think the American people would be very positively surprised by the advances in technology that really only America has. >> Well, without giving away too many secrets, like can you just explain what edge that is that you give our war fighters? >> of all, there really two things. Every every military force in the world, so you see this uh on the battlefield in Ukraine, like one of the big issues the Russians had

[02:13] is how do you transform your operations into a modern operation? And if you take components of Maven, there's basically infrastructure which allows you to take older systems and make them newer systems. And then, so that allows you to to have a completely integrated battle space, which is super hard because of security, clearance, data sources. You have You can think of it as infant data with very with some of which is very sensitive, some of which no one else can see, some of which everyone has to see. You have allies working on it. How do you get that into one plane of glass? With the security. How do you then use advances in AI and machine learning to enhance that uh while making those decisions not like what's proverbially

[02:53] called loose AI code. It's just not useful on the like None of these things actually are used anywhere on the battlefield. So, how do you enhance your ability to see things? How do you enhance your ability to go through very complex systems? So, like for example, in Ukraine, you're not It's not just getting the drone to the end point. It's going through advanced radar systems. And then, vis-à-vis Iran, it's 100 times more complicated because they have You know, it's like you look at American performance, which has been you know, thanks to the bravery of Americans and the ingenuity and I would like to believe our tech, very few losses on our side, very precise, very controlled, and then in a complete

[03:28] ability You saw this in Venezuela to dismantle these things. And so, because the dismantling is happening, we underestimate the complexity that it took to do that. It was not possible to do these things 2 years ago, 3 years ago, and now America can can can really do this with great precision and with very low cost to to us in terms of lives and casualties. >> opening up new business in different regions for you, the success of what you're saying in the Middle East? >> the the the the one of the many secrets of Palantir, people talk about our forward deployed engineers, they're trying to copy, talk about ontology, they try to copy, talk talk about, you know, how software that's like fine software's worthless, but everyone's

[04:06] trying to copy us. So, that's obviously and some points of software are not that valuable, but some parts are crucial. >> is. >> Well, there are three kinds of software really. There's basically you can think of infrastructure, which is like hard coded things written into the enterprise that have to be managed on prem often, so you maintain maintain the sovereignty of your own data. That is not valuable. That is excruciatingly valuable. Then you have what how do you manage that? So, like ontology, foundry and other things, those things are are very valuable and durable. Then you have FDEs and large language models. Most FDEs are just throwing people at a problem that anyone else can solve. So,

[04:43] if you can manage so those it's some are valuable, some are worthless. Uh free code is essentially very easy. So, you have a very different but in any case, um what you see in crucial situations in Russia, in the Middle East or in Ukraine, Middle East, uh where you have to bring home people safely and where your adversary has to be dismantled and then adversary don't, you know, the Persians and the Russians and the and the CCP are exceedingly talented technical people in various different ways with very unique and historically viable ways of fighting. So, you're not going up against an adversary that's just going to lay down. When you have to do that, you have to bring your unique way of doing something enhanced with

[05:27] things no one else has. Those things now are infrastructure software uh foundry, Apollo and large language models that we are producing managed by us. It is the combination that is legal. By the way, you see this in US commercial and the interesting thing about what's going on in the world right now, so domestically is you have essentially a really important, uniquely American movement uh call it Frontier Labs. They are There's the way we think about this actually is just philosophically and technically wrong. Philosophically it's wrong because it's not doomer versus not doomer. It's a hyper religion of hyper optimism. They believe all problems peasant peasant present, past, and future, including the ones they

[06:12] create and don't acknowledge they create are going to be solved by them, including human nature, >> Mhm. >> uh disparities in Okay, and then on the infrastructure side, enterprises are just fed up because they know this doesn't actually work this way and it's not working. And that basically is drives our commercial business. On the on the government side, we're we are super pro-American, super pro-West. >> You've always said that. And you've always said that. I do wonder though how you protect the company from getting politicized. I mean, you've worked with administrations on in both parties before, but especially, you know, when it comes to this war, which is unpopular and the Democrats have been fighting it,

[06:54] you know, how how you avoid making Palantir a target in the midterm elections or in the next presidential election? >> I don't want to be positively overly naive. We are trying I am trying sometime with success, sometimes without to move people that that that AI AI infrastructure revolution is a unique opportunity for America because it we've developed almost all of it first. However, um it it cannot be discussed in blue-red terms. Like currently the biggest danger to America is you you know, it's like people are like, "Oh, this guy agrees with me politically, therefore he's right. This woman agrees disagrees with me politically." They're like, "This is a massive massive revolution and there is a danger" Look,

[07:37] I've been running around telling people for the last 6 months in private and somewhat in public >> The nationalization >> is coming. >> And the reason why it's coming is you can't do a blue red debate. This is a massive revolution and it's going to there's opportunities only America has and there are dangers in this revolution. And for example, I got in trouble for saying not saying we would have a draft. I wasn't saying that. I was saying we're in a >> Some people think you were advocating a draft. >> what I was saying is we're in a massive time of dislocation. And in dislocation, we have to find ways to come together as a culture. And while it is the case in Silicon Valley that everyone thinks

[08:13] we're going to magically like each other because there's abundance and no one has a job. Unless we're ending history. In fact, it's actually very much the opposite. We're going to have massive resources, but they're going to disproportionately go to people who are already wealthy. Okay, that is a political problem. You can't just say oh, we're all going to have more therefore we're happy. Okay, well, you know, we're not at the end >> So, what should the what should the industry do about it? >> We have to be we have to be frontal about there are issues that are real and getting worse like wealth accumulation is totally fine. Even though I am actually a card-carrying progressive, I don't think card-carrying progressive

[08:50] have anything against >> a card-carrying progressive? >> I I I know. The question is not whether I'm progressive. It's the question is whether some progressive are progressive. I am progressive. I want poor people to have better lifestyle, higher standard of living, all poor people. I got in trouble for mentioning that, you know, 40 37% of our GD GDP is female. 50% of Americans 52 roughly are female. There's a dislocation there. 67% of people who have who have gone into graduate school are female. These parts of the market are going to be put under massive pressure. It doesn't help to pretend we're not going to have these problems. In fact, the progressive position is progressive because you speak about problems like

[09:33] okay, so we have we got we can protect our workforce. Can we protect our workforce and have my great migration the way we used to? Can we protect our workforce by teaching people the same skills we used to in the past? You know, we have a massive program at Palantir to help neuro divergence. Not because I'm particularly a good person on that score. I mean, I am dyslexic and some people think I'm neuro divergent. But because those people are particularly valuable for our economy. And what made someone come to America from all over the world? They were neuro divergent. They're like, it's not working out for me in my culture. I got to come. So >> Okay, I'm glad to hear that. >> But but in any case, yes. But But the

[10:08] point is we have to somehow get everybody around people who totally get you know, it's like when Bernie Sanders says, "I'm going to take 50%." Okay. You know what? >> You think that's actually going to happen? >> I I'll give I'll give you a If If we keep going, Bernie Sand Do you know how people don't think I'm progressive? In 2 years, they're not going to think Bernie Sanders is progressive. They're going to be like, "Bernie Sanders, you only want 50%?" What do you What is this 50%? >> we're headed as a country? >> We're absolutely headed in that direction. And we're headed in that direction simply because the most important Leaving the tech aside, which is also driving it. The most important

[10:45] political decisions in this country are going to be driven by Do you understand AI? Do you understand what it's doing? And as importantly, do you Are we going to tell the American people don't believe your lying eyes? Like politicians in the past have done this. It does not work. The American people are really wondering what is going to happen to them. And the answers aren't all good or bad. And then they can't also >> Do you see mass job losses? >> It's It's It's It's It's It's It's not the job loss the thing that is mass needing to retrain and retool and do things differently. By the way, this is an incredibly important American advantage as I I spent half my life in Europe. If you just said, "What are the

[11:26] advantages they have? What are the advantages we have?" The what The biggest advantage we have is our dynamism, our ability to change, our ability to rebuild. But we're going to have to use that advantage because the jobs of the past are going to be very different than the jobs of the future, and we're all going to have to change, and that's going to disproportionately affect people, and we need and then we need a common purpose. What is our common purpose? And it that's something that we're really struggling with, and when you have these struggles, people go to the far right and the far left, and they're very charismatic. We should not make the mistake Europe has made by not addressing these things, and then the

[11:59] only people who win politically are people who may not know what they're talking about, but are very charismatic because they're addressing a real issue. And in this case, the real issue is Americans are afraid. And by the way, the other real issue, and I'll just talk about somebody I've spent 20 years in enterprise. We work with the most important governments. We are controversial, quite frankly, because we power every single Western power that's at war. >> Yeah. >> And by the way, we're also we're all Yeah, you know why we're also controversial? They have capacities they didn't have in the past, and our adversaries >> I know, you continue to make that point. >> but the commercial thing, the commercial

[12:35] thing enterprises are also unhappy. You know, it's not just the man and woman on the street that is unhappy with the Frontier Labs. It's in private, every single enterprise we deal with. And we have this thing inside Palantir that's somewhat controversial, but we keep debating, should we pay people to go to large language model companies before they come to us? Cuz people come out of there screaming saying this could never work for me. They don't understand the enterprise. They don't care about my enterprise. They want a token max. >> Okay, so this is you're getting to the heart of it because I think the biggest question around Palantir right now on Wall Street is whether the large language models, the Anthropics and the

[13:10] AIs, which are about to go public, can replicate what you're doing. >> You know what? It's it's a real question that no one in enterprise factually is worried like, you know, the investors are really into it short-term have no idea what they're talking about, long-term end up being right. We we we will just debate this with the facts. You know, I I've spent all my life for better or worse dealing with the most complicated, most enter- interesting enterprises. I'm on the ground floor in that, probably like no one else. And I >> you're saying. >> They're I mean, most of them are chill chill chillaxing over their latte, reading a report about something that they don't understand the technical capacity of it. You know, and you know

[13:51] >> But they're hiring the kind of engineers >> Yeah, they're great. >> that you have. >> That that And those kind of engineers are great engineers. And I'm telling you they don't talk to the enterprises or don't understand the technical challenge. And also, by the way, the ability to be a great investor in the kind of where those That's where large language models are more useful because it's it's probabilistic and you don't have to get better than really 51%. If you want to manufacture a car and you need a part or you want to send a rocket to the moon or you want to put a missile on your adversary's head and bring home Americans safely, that stuff doesn't ship. And by the way, and there's not a

[14:26] single high-end enterprise like that that would ever put By the way, that's before you get to the cultural impasse. Like the the When you go to San Francisco and talk to them, their basic vibe is we don't have to solve your problem today because tomorrow you're going to go away and all of your problems are going to be solved. It's largely religious. And then you get to we're going to do we're going to replicate Palantir by doing a deployco. I mean, it's a complete farce. Like the people who go there to solve the simplest easiest problems that sell tokens. And they're the part they don't understand, honestly, and I told them this, I probably shouldn't, is they don't understand how unlikable they

[15:01] are. So, they the product doesn't actually work the way it's very expensive and you got to overcome the fact that you may not realize outside of San Like actually, I'm not even saying I find them unlikable because actually some of the best and most interesting conversations I've had in business are like hanging out with some of them, Sam, Dario, others. I I I these are high, you know, like you talk to Darii, it's a great time. And he's he does >> together at AIG. I was going to ask, is Anthropic are they are they friend or foe? Are they competitor >> Well, he No, again, this is Well, I mean, most of the things they talk about in public are running on Palantir. So, it's like it Now, at this scale, we're a nation-state,

[15:40] basically. They are like we are fri- I like it's I don't wouldn't and they're and he is the he is the frontier person of a of the leading frontier model com- company currently. And he's gone from being way behind to being ahead. And he's just a he's a very, very important person. And And I'll tell you, he believes what he's saying. I just would say that First, wait. He believes all problems in the I think and I impute him or believes all problems of the current, present, and future the one are transitionary to a perfect future. I believe that we need heaven on Earth, not heaven in the next in 20 years. And I'm very We disagree on these things. But doesn't mean that, you know, it's like we don't agree on other things.

[16:25] It's like, you know, they agree on certain national security, we disagree. It's It's not It's not as simple as agree, disagree, bad, good. These are some of the more interesting, important people in the world. I I'm betting heavily and my customers pretty much are on my side that, you know, it is not that large language models aren't crucial for the world. It's just the implementation is where the value is, certainly in the next 7 years. And if you want to ask me what's going to happen in 10 years, I don't know. But, you know, the risk of a pathogen uh destroying us >> [music] >> is probably as high as >> Is that why you guys are working so much on health care right now? >> Um well, you know, the the the the thing

[17:04] is these low probability events that shape the world are often underestimated. And they are again, they're These are not like These These are events that call require software infrastructure, implementation, ontology, and large language models. As an example, if you have a deadly pathogen come into this country, the signs of this are going to be very similar to the signs of finding a somebody moved a truck by 2 cm and that means they're going to attack. Is it is only possible and then the security around it which people often underestimate that we do, which is the bedrock of the reason why Palantir is indispensable. It's like you can't just use data sets in this environment. Each single data set you can think of as each

[17:49] single micro part of the data set has to be its own sovereign data set. So this whole move you get of what part of the data is sovereign, you can think of it we do that at the millimeter level. Like which part of the attribute is sovereign to the person. What can be shared? What can be imputed? What can't be imputed? What can't be generalized and shared? What has to be shared? Which experts can see what under under what conditions? Is the measurement accurate? How long is the measurement accurate? Under what conditions is it accurate? Who does the implementation? Do they have the clearance? What all these things that all those things have to be modeled and done in a platform that's been doing this. By the

[18:25] way, the cyber threat like how do you can't just run large language models. You actually how do you patch what you find? How do you patch what you find at scale? How do you is the software you run How do you and if you could do all that, how do you have the accreditation to work on a secure network if it's never been you've never actually done that before? No idea what's involved and you need the accreditation to do it. >> So so they're not a cure-all. But finally Alex, I did want to ask you about SpaceX cuz it's going public this week as well and I know you've partnered with them on on proposals like the Golden Dome missile protection and you are working on AI operating I think as it relates to space. So I'm just curious

[19:03] how you view that area. >> I very optimistic for the future. I root for every entrepreneur. I'm rooting for Elon. And uh, we partner in some areas and probably don't compete in others, but it's Look, when a when when when a massive IPO, first of all, you know, we're so blessed in this country. Like, okay, so I mean, I spent half my life in Germany. If there was an I IPO, DPO, or anything at like the $120 billion range, the whole country would be jumping up and down. Here we have a country where it's like appears to be the one of the large, if not the largest IPO ever. Uh, it's super cool technology, things that no one else have done, and like I think I think most Americans, and certainly I hope all entrepreneurs are rooting for

[19:51] success. >> And space? Bullish on Are you bullish on space? Are you guys doing something? >> Well, again, it's very similar to these other debates. You People underestimate that bullish on space if the person doing it knows what they're doing. Bullish on Elon and space? Yes. Bullish on space? No idea. >> Okay. Alex Karp, thank you very much. >> Thank you. >> It's always good to talk to you. >> Yeah, you're in great demand. >> Thank you. >> Great demand. >> Alex [laughter] Karp >> Glad I got on here. Bye. >> Um, Palantir, um, defend really defending the company, Carl, against I think some of the narratives that we talk about and hear about on Wall Street, especially when it comes to the AI software, where there's been this

[20:27] rotation out of it and names like Palantir into the into the infrastructure plays and potentially the LLMs here. >> Vampire Dracula here. If you found value in this video, please smash the like button. Let me know your favorite part of the video by commenting below. And most importantly, don't forget to subscribe and ding that bell. Until next time, keep it real.