Palantir Technologies (PLTR) hosted its AIPCon on Thursday, an invitation-only customer conference that has become an important way for the data analytics firm to showcase its AI software. The company has been a huge beneficiary of AI, with the stock up nearly 40% so far in 2024.
In an exclusive interview with Yahoo Finance’s Josh Lipton, Palantir CEO Alex Karp explains how the company’s AI work is providing value for both customers and investors. “The large language model is much more like a chemistry experiment. The outgrowth of which is something that is useful when refined. And the refinement of that for your enterprise happens in what we call our ontology, which is where we impose the logic of your business on the large language model in the security and intellectual logic of your business. And this is transformative. And what it means for investors and others is there is value in this market.”
On the 2024 election, Karp says that, in the past, it hasn’t made much of an impact on Palantir who controls the White House. But what has changed is how the US Department of Defense is spending its money. “The DOD wants to move quicker,” Karp says, in a bid to stay ahead of adversaries.
Karp argues there is “only one way to fight” deadly adversaries and that is “with a deadly set of weapons.” Karp explains, “My version of fighting is you’re so strong, you’re so willing to fight, they don’t fight you…. You want peace in the world? You have a lot of growth in America and you have really deadly weapon systems. And you tell your adversaries, ‘You know what? I just might use them. I’m not even gonna explain why I’ll use them. You start acting up, I will use them.'”
Video Transcript
I’m Josh Lipton and I am here now with Punier Ceo Alex Carp for this exclusive Yahoo finance interview here at A IP con.
This is the company’s customer conference.
Alex.
Thanks for taking the time to chat.
Very happy to be here.
So today there’s gonna be more than 250 of your customers and they’re gonna be here, Alex talking about A IP.
This is your flagship A I platform.
Let’s start there, Alex for, for many of you who are as familiar, what is a IP, how does it work and how are businesses leveraging it?
Right?
Iii I think what everybody watching this is familiar with is you have a massive he cycle around large language models.
And then when you try to use them in your enterprise, you find out that it’s more like self flagellation and it’s expensive with no output.
And what we learned in the context of war fighting primarily, but also across the uh 20 years of building uh software infrastructure were.
How do you manage an emergent natural resource called large language model uh models in a way where you actually get value, meaning you can transform your enterprise, you can change the margins, you can turn tech, non technical people into technical people on the battlefield.
You can do things that would have otherwise cost billions of dollars for millions of dollars.
Meaning being very precise in how you allocate troops, being precise in how you target people.
Um There’s a fundamental fallacy around large language models.
People could conflate uh actual data of an enterprise which is structured and can be taken apart and understood with a large language model, which is much more like an emergent property.
Something that is can be modeled used, but you need precision and all the value in the market is going to go to chips.
And what we call ontology, we have this ontology and the ontology allow you to take a large language model and use it, refine it and you then impose it on your enterprise in the logic of your enterprise, in the security model of your enterprise.
And what does that mean for the 7070 people are here presenting talking about why they’re using it, how they’re using it.
You gotta remember we started uh selling this product just over a year ago.
Uh We started with the claim that we knew a lot about the precursors of large language models and large language models.
And the general approach of just buying models is going to be essentially self pleasuring for an enterprise at the cost of the enterprise.
And no one believed us And now you see 70 people saying, hey, we’re using this for construction, we’re using it for hospitals.
Uh the people who are not speaking, but we’re using it on the battlefield.
We’re using it to compress margins.
We’re using it to build enterprises that we’re only able to build in Asia and America.
We’re making engineers, uh better engineers.
We’re making people are not engineers into engineers using or anology and a large language model.
And this works very, very quickly and it is substantially changing the health and vitality of every business here as opposed to the alternative, which is that you buy some large language model, you, you party with it basically.
And the next day you have a hangover and, and then, and so like and again, for, for people just looking at this, what does it mean for Palent here?
It means we are sitting on the only thing that actually creates uh a quantifiable transformational value in an enterprise.
Yes, it is not understood well, because everybody understands the problem incorrectly.
Yes, it is going to transform America and our clients are leaving the way and, and by the way, it’s like to actually show things that are not understood.
You have to actually show them the whole purpose of this conference is I can tell you how an ontology works.
It’s actually quite simple.
You have the logic of the business, including the security logic of the business and you have something that it approximates new knowledge in a new form called a large language model.
And it allows you to take the value of the new form, the raw resource of a large language model, largely powered by chips and put it into a precise organization.
In the logic of the business.
I can tell you that where I can show you 70 businesses say my business is stronger, healthier and much better than any other business in my sector.
And so what is the bottom line for investors?
Listing Alex?
How would you characterize what is commercial demand like right now?
For a IP the way I would explain it to my most important investors which are individual investors.
There’s two parts of the market that are creating value.
People will pay for chips and ontology and you understand the chip side now and now you’re going to understand on and what does that mean for your business?
It means you can actually use this raw resource and process it into something that actually works.
And if you want to see exactly how what the people presenting, you’re going to see people in construction, using it to build buildings quicker, cheaper, more accurately, you’re going to see people in the hospital industry saying how could I possibly ever distribute my patients ethically fairly and commercially relevantly?
How could I distribute resources across?
How can I manage my my company as not abstract units but as a portfolio.
These are these things are happening within days.
What I’m interested in the Alex is because you, you can talk a lot of CEO S will say and they’ll sit on your finds.
They’re interested in A I, but actually they’re not willing to pull the trigger yet because they have these questions about privacy security vs but, but you’re saying we’re crushing it.
We’re closing.
What I would tell any of them is how are you doing it?
Well, there’s a technical answer which I gave you the precursor of.
But again, do you wanna know how we’re doing it or do you wanna enjoy it being done?
Both?
Great.
Well, first, you, but no, actually, first you wanna see it works like the, the, the, the central issue I think, I mean, we have hundreds and hundreds of CEO S and by the way, the thing about the 70 people here presenting is, you know, they’re presenting to other people.
I’m not paying them to present.
The reason they’re presenting is they’re like, wow, I didn’t really believe this could work and now it’s working really well.
So the most important thing, by the way, for a normal user, someone who’s gonna pay is does it work?
Then the second question is how does it work?
Does it scale?
What is the commercial model you really have to establish it works?
I would stand by the thing that it, in this case is people think of large language models as the value in of itself.
What they’re going to find is the large language model is much more like a chemistry experiment.
The outgrowth of which is a something that is useful when refined and the refinement of that for your enterprise happens in what we call our ontology, which is where we impose the logic of your business on the large language model in the secure and intellectual logic of your business.
And this is transformative and what it means for investors and others is there is value in this market.
People, you can identify where the value is very easily.
Are people paying for it?
Will they pay for it?
And what will there’s a prey to optimality to a lot of this stuff and that prey to optimality happens when people say, oh I am actually getting value and that’s why people are here.
Let me ask you, you suggest I wanna broaden the conversation because you kind of suggested this, just give your take on the moment we’re in because there’s some Alex, you’ve heard who say, listen, this is a historic moment, Internet mobile and now A I NVIDIA Jensen Wong said the next industrial revolution has begun.
Do you agree with that?
Do you agree with Jensen?
Um I think you’re gonna have very divergent for some people, this is the next generation and for others who I think they’re doing very similar things, it’s gonna fail.
Who defines the winners?
Who will be the winners well, come talk to our 70 customers, they think they’re winning.
And it’s like, you know, you look, it’s a lot of this goes back to basic ideas of how you build things.
Why you gotta talk to people who are getting value and ask them, how are you getting value?
Why are you getting value?
Are you paying more or less for that value than you think you should?
And that’s exactly what you’re gonna see today.
You’re gonna see people saying I’m getting a lot of value.
It went much quicker than I thought.
I’m, I’m outstripping what I thought I could do.
I’m better, much better than the people.
It happened efficiently and I’m probably, I don’t think they’ll say this, but I’m paying less than I think I should.
And that’s why I’m very happy about it.
Let me get your t too.
When we talk about A I, we got to talk geopolitics.
I’m very curious to get your answer to this question.
Who is in pole position?
Us or China?
Well, we have a very different problem in the US than any other country in the world.
We are in pole position, but we obviously have to move much, much quicker.
Um You see on the battlefield and again, without mentioning where these things are used, exactly what I’m talking about.
Large language model plus ontology is can be wielded as a very deadly weapon and create enormous efficiencies for especially small armies against large armies.
We have to be myopically focused in this culture of, we are in a pre war situation with our adversaries and we can no longer afford the theologies about what works on the battlefield.
We have to go very deep and look at what’s happening on battlefields and do a diagnostic of it.
It’s much less critical, but it’s the same thing that businesses have to do what is working, what is not, not a theological thing of what would I like to believe e is working and what works on the battlefield where our central advantage is as the United States of America, the leader of the West and the most important country in the world is technology venture backed companies and especially software and software driven weapon systems.
And how, how do you get the software to control the hardware so that we can, we can get out of the business of trying to compete with people who are better at hardware and get into the business of producing something they do not produce as well comes down to execution.
The problem we have in the West actually is again, is prey to optimality.
The, the the products that actually scare our enemies are, are just not products that employ mill millions of people.
They, but they still are deadly.
They are software driven hardware systems.
Moving to software driven autonomous hardware systems, one of the biggest lessons of war and one of the reasons this is so crazy.
Valuable is the Russians and all of our adversaries have now learned the value of electronic warfare.
By the way, again, these things are exactly isomorphic to commercial.
You have enormous new headwinds, the commercial, your competition has developed something that renders your product strategy irrelevant.
You have to rebuild it with a new playbook.
This is exactly what happens on the battlefield.
You cannot use a normal missile or normal tank on the battlefield anymore because the electric warfare, electronic warfare allows you to suppress the GPS signal.
And that means these things have to be controlled uh with software to go around map where the electronic, where the suppression happens.
And then in the final mile at some point, autonomously, completely different ways of doing A I completely different news ways of doing it, all of which are an advantage to America.
If we implement just like in American industry, we have something called a chip industry that is the world’s leading in this area and, or at least the, the, the final product and we have something called soft or ontology which actually makes all of this crazy valuable.
Do you implement it?
Do you go talk to people are doing it or do you read the powerpoint written by someone who has no skin in the game?
And, and, and the answer to that determines whether America in the West, which I view as a superior way of living because it’s fairer better meritocratic leads to free speech allows for wealth creation.
Does that culture actually dominate and win?
And we, by the way, have to get out of the apologizing for our culture thing.
It’s like we are a better way of living and we have to be willing to fight for it on the economic front with growth.
That’s why this conference is so important.
We need growth.
What makes America different than our adversaries right now is not just Western system, we have growth.
Where is that growth coming from?
Technology?
What is that technology?
It’s compute plus ontology basically.
And that can change the very structure of our culture from manufacturing to weapon systems to how we organize things to our health care distribution can have a completely different way and better way of living.
Let me ask you this, you, you’re talking about A I and we’re talking about the benefits, the advantage of the use cases.
I am also interested.
I was to get your take on when you think about the risks of this technology.
Let me put one forward.
We, we have an election coming.
How concerned are you as an American executive but also just an American citizen about the very adversaries we’re talking about here deploying A I for nefarious purposes rampant to sow confusion and division and there’s only one way to fight uh uh uh uh a deadly adversary.
What is that with a deadly set of weapons?
So like if you, you know, if you want to fight our adversaries and by the way, my version of fighting is you’re so strong, you’re so willing to fight, they don’t fight you.
So it was like, I recently have gotten in trouble for saying I’m a peace activist.
I am a peace activist.
I’m a peace activist because I want enormous growth in America and I want real deadly weapon system.
You want peace in the world, you have a lot of growth in America and you have really deadly weapon systems.
And you tell your adversaries, you know what I just might use them.
I’m not even going to explain why I’ll use them.
You start acting up, I will use them.
You don’t act up i through strength, peace through strength and a little bit of unpredictability.
You know, we are in a world where our adversaries are founders, they have the founder playbook, which is I’m gonna, I’m gonna be aggressive and win where I can.
We have a playbook written probably at Harvard and Yale and Stanford Institutions which are now symptomatic with not producing, producing structures where they don’t see discrimination when in front of you producing structures that and ideologies that are literally built for the decline of our country.
We need to get out of the business of we’re declining.
We’re apologizing, we’re not going to use force III I we needed and, and the peace activists are actually us and the pe people who claim to.
He’s activists are actually sowing destruction disorder discrimination.
It’s an exact inversion of what they claim to be.
And we, and we, but those of us on our side who are pro growth, pro America, pro our economy transfers and pro having the best systems in the world.
We need to organize, build things.
And then we also need to have exact understanding of what works, what doesn’t and implement it very quickly.
I I mentioned the election.
I want to stick with that real quickly.
Does it matter to Palent here?
Who wins?
Biden or Trump?
Does that impact any potential business you’re going to have with the US government?
Look, we’ve been in, we’ve dealt, we’ve been in business for 20 years.
I have not seen an impact on our business, depending who’s president, I am interested.
Listen, we talk about these conflicts around the world right now.
Running hot.
Russia, Ukraine, Israel Hamas has that changed how the dod Dod thinks about spending out and what they want to spend on.
Uh The dod is changing rapidly, how it’s been so.
Well, there’s much more of a focus you saw that we won this big maven contract.
Um These things are a natural growth, uh outgrowth of, we need to fight better.
We need to fight quicker.
We need to fight on the terms where America has superiority.
Of course, we can move quicker.
Of course, the dod wants to move quicker.
Our society wants to move quicker.
I think unfortunately, we are going to end up moving even quicker because our adversaries are, are, are likely to continue their nefarious acts which Russia, China, Iran, I mean, all three are working together.
You know, one of the things I, I just, I just II I travel around the world and basically wherever you go, people are like, well, I want you to, I want, I want you to be strong on this one and appease the other people who believe in appeasement, whether it’s Iran, China, Russia, pick your favorite, the one you’re sympathetic with.
They don’t understand.
We have adversaries there working together despite what problems you may have.
Some people don’t like the war in Ukraine.
Some people don’t like Israel.
Some people want an appeasement strategy with China.
It doesn’t work.
They are working together against us and we have to work together against them.
Last question you mentioned, Maven, I wanted to get the dod recently disclosed and this is a five year $480 million contract with Pal here.
Just walk.
What is that contract?
What is it, what does it mean for your company?
I never know what we’re allowed to say about these very sensitive contracts, but I, I’ll tell you what it means for the nation.
This country is focused on using A I to have a structural advantage in how we deploy and understand the battlefield, deploy our assets and understand the battlefield.
And P here plays a crucial role in that, in that arena.
We’re gonna continue to play a crucial role in the arena.
And by the way, we’re making it possible through our software to allow other software and hardware companies to work off of our platform.
So this is not just a win for Pollinger though it is, and I’m very proud of it.
This is a win quite frankly for the tech ecosystem that powers this nation.
How big is that ecosystem now, which has grown a lot since we first started talking, the the the venture capital world has completely flipped.
When we talked the last time by and large, the tech company co community was adversarial to palant here on this issue of supporting our nation.
Now, I think the tech community is the most, most supportive part of our economy in, in, in protecting our nation.
And why did that, why did that flip, by the way?
Hey, you know, they stopped hating the player or the game, Alex Carp.
I know you are very busy today.
We appreciate your time.
Thank you.
Take care.
Signup bonus from
