Public transcript

CNBC's interview with Palantir CEO Alex Karp June 10th 2026 (1080p HD)

By: Global Secrets
Length: 21 min
Language: English
Text from: the video's captions
Transcribed: 13 October 2026

Read the full transcript Open on YouTube

Notes in English.

Summary

CNBC's Sara Eisen interviews Palantir co-founder and CEO Alex Karp about the company's role in the Iran war, the Maven military platform, and why he believes large language model companies cannot replicate Palantir's enterprise work. Karp argues the AI infrastructure revolution cannot be debated in blue-red political terms and warns of coming nationalization pressures and workforce dislocation. He also discusses Anthropic's Dario Amodei, health care work on pathogen detection, and the SpaceX IPO.

Key points

Karp says the Maven platform makes LLMs lethal and useful on the battlefield to bring American warfighters home safely.
He claims recent precision operations with very few losses were not possible two or three years ago.
Karp warns that nationalization is coming and the AI revolution cannot be debated in blue-red political terms.
He says no enterprise customer is factually worried about LLM companies like Anthropic replicating Palantir.
Karp calls Dario Amodei a very important person but disagrees with his belief in a transitionary path to a perfect future.
He says pathogen detection work requires millimeter-level sovereign data controls and security accreditation.
Karp is rooting for Elon and calls the SpaceX IPO possibly the largest ever.
He describes himself as a card-carrying progressive concerned about AI wealth going disproportionately to the already wealthy.

Questions it answers

00:00Palantir's role in the war

What is Palantir's role in the Iran war and military operations?

Karp says the Maven platform takes valuable LLMs and makes them lethal and useful on the battlefield, with the company's primary focus being bringing American warfighters home safely with a massively unfair advantage. He says these operations begin well before the public sees them and that recent American precision and low casualties reflect technology that did not exist two or three years ago.

Karp says the Maven platform makes LLMs useful and lethal on the battlefield.
He claims recent precision operations were not possible two or three years ago.
He says the American public would be positively surprised by the technology advances.

02:11The edge Palantir provides

What edge does Palantir give warfighters?

Karp describes two components: infrastructure that turns older systems into newer ones for a completely integrated battlespace, and AI/ML that enhances the ability to see things and go through complex systems like advanced radar. He says Iran is 100 times more complicated than Ukraine, and that adversaries like the Russians, Persians and CCP are exceedingly talented and will not just lay down.

Maven infrastructure integrates older systems into a modern, integrated battlespace.
Iran is described as 100 times more complicated than Ukraine.
Karp calls the combination of infrastructure, Foundry, Apollo and LLMs lethal.

07:01Politics and nationalization

How does Karp keep Palantir from being politicized?

Karp says the AI infrastructure revolution cannot be discussed in blue-red terms and warns that nationalization is coming. He describes himself as a card-carrying progressive, argues wealth from AI will disproportionately go to those already wealthy, and says the country needs a common purpose or charismatic figures on the far right and left will win politically.

Karp warns the nationalization is coming.
He says AI-driven wealth will disproportionately go to the already wealthy.
He cites figures like 37% of GDP being female and 67% of graduate school entrants being female.

12:08Competition with LLM companies

Can Anthropic and other LLM companies replicate what Palantir does?

Karp says no one in enterprise is factually worried about it, and that LLMs are probabilistic and useful where 51% accuracy suffices, but do not ship for manufacturing, rockets or missiles. He calls the idea of replicating Palantir with a deploy a complete farce and says frontier labs' attitude that your problems will be solved tomorrow is largely religious.

Karp says no enterprise customer is factually worried about LLM companies replicating Palantir.
He says Palantir debates paying people to go to LLM companies first before coming to them.
He calls replicating Palantir with a deploy a complete farce.

15:58Dario Amodei and Anthropic

Is Anthropic a friend, foe or partner?

Karp says most of the things Anthropic talks about in public are running on Palantir, and calls Dario Amodei a very important person who has gone from way behind to ahead. He says they disagree philosophically—Amodei believes problems are transitionary to a perfect future while Karp wants heaven on earth now—and bets that implementation is where the value is for the next seven years.

Karp says most of what Anthropic discusses publicly runs on Palantir.
He calls himself and Palantir effectively a nation state at their scale.
He says implementation is where the value is for the next seven years.

17:23Health care and pathogen detection

Why is Palantir working on health care?

Karp says low-probability events like a deadly pathogen entering the country are underestimated, and detecting one looks similar to detecting that a truck moved two centimeters before an attack. He explains this requires software infrastructure, ontology, LLMs, and millimeter-level sovereign data controls, plus accreditation to work on secure networks that companies without prior experience cannot get.

Karp says detecting a deadly pathogen resembles detecting a moved truck before an attack.
Every micro part of a data set must be its own sovereign data set.
Accreditation to work on secure networks requires having done it before.

19:16SpaceX and the IPO

What does Karp think about SpaceX going public?

Karp says he is optimistic and rooting for every entrepreneur including Elon, noting Palantir partners in some areas and probably does not compete in others. He calls it possibly the largest IPO ever with super cool technology, and when pressed says he is bullish on Elon in space but has no idea about SpaceX itself.

Karp says he is rooting for Elon and every entrepreneur.
He calls it one of the largest IPOs ever with cool technology.
He is bullish on Elon in space but has no idea on SpaceX itself.

Notes

Palantir's Military Role

Karp says the Maven platform takes valuable LLMs and makes them lethal and useful on the battlefield.
The company's primary focus is bringing American warfighters home safely with a massively unfair advantage.
He claims recent American operations showed very few losses, high precision, and capabilities that did not exist two or three years ago.
Iran is described as 100 times more complicated than Ukraine because of advanced systems adversaries have.

The Technology Edge

Maven has infrastructure that turns older systems into newer ones, creating a completely integrated battlespace.
Integrating sensitive, classified and shared data into one pane of glass with proper security is super hard.
Karp describes three kinds of software: on-prem infrastructure (excruciatingly valuable for data sovereignty), management tools like ontology and Foundry (very valuable and durable), and PhDs and LLMs (some valuable, some worthless).
The lethal combination is infrastructure, software, Foundry, Apollo and large language models managed by Palantir.

Politics and Nationalization

Karp argues the AI infrastructure revolution cannot be discussed in blue-red terms.
He warns the nationalization is coming and that AI wealth will disproportionately go to people who are already wealthy.
He cites that 37% of GDP is female, roughly 52% of Americans are female, and 67% of graduate school entrants are female, saying these parts of the market face massive pressure.
He describes himself as a card-carrying progressive and says Palantir has a program to help neurodivergent people because they are particularly valuable.
He predicts that in two years people will say Bernie Sanders only wants 50%.

Competition with LLM Companies

Karp says no enterprise customer is factually worried about LLM companies replicating Palantir.
LLMs are probabilistic and useful where 51% accuracy suffices, but do not ship for manufacturing cars, rockets or missiles.
He calls replicating Palantir with a deploy a complete farce and says frontier labs' optimism is largely religious.
Palantir internally debates whether to pay people to go to LLM companies first before coming to them.

Anthropic and Dario Amodei

Karp says most of what Anthropic discusses publicly is running on Palantir, and calls Palantir effectively a nation state at its scale.
He calls Amodei a very important person who has gone from way behind to ahead, but disagrees with his belief that all problems are transitionary to a perfect future.
Karp bets that implementation is where the value is for the next seven years.

Health Care and Security

Detecting a deadly pathogen resembles detecting that a truck moved two centimeters before an attack.
Every micro part of a data set must be its own sovereign data set, managed at the millimeter level.
Accreditation to work on secure networks requires having actually done it before.

SpaceX

Karp is rooting for Elon and calls the IPO possibly the largest ever with super cool technology.
He is bullish on Elon in space but says he has no idea about SpaceX itself.

Text from the video's own captions. Summary and notes written by AI from the transcript, so check anything important against the recording.