Public transcript

Palantir CEO Alex Karp's this morning on CNBC

By: UninformedInvestors
Length: 20 min
Language: English
Text from: the video's captions
Transcribed: 13 October 2026

Read the full transcript Open on YouTube

Notes in English.

Summary

Palantir CEO Alex Karp joins CNBC's Squawk Box to discuss the company's expanded partnership with NVIDIA to build custom AI systems for the U.S. government. Karp argues that enterprises distrust frontier labs like OpenAI and Anthropic because they fear losing control of their data, model weights, and business

Key points

Palantir is expanding its partnership with NVIDIA to build custom AI systems for the U.S. government, and the stock is up almost 6% this week.
Karp says enterprises distrust frontier labs because they fear losing control of their data, model weights, and the alpha of their business.
Palantir's ontology application layer makes large language models safe, useful, and precise without touching underlying data or transferring IP.
Karp claims Palantir has more demand than it can supply and can see 15 to 18 billion dollars in free cash flow two years out.
He praises Dario Amodei as a historic figure who came from behind to become number one, while criticizing how frontier models are sold.
Karp says the relevant tech centers of the world are America, China, and Israel, and that China is a peer-to-peer adversary in AI.
He describes himself as the most publicly supportive CEO of Israel while also being, in private, probably its most effective critic.
Karp says Iran has been degraded and there are things not in the public space that would be comforting to people.

Questions it answers

00:06NVIDIA deal

What is the new Palantir-NVIDIA partnership about?

Palantir is expanding its partnership with NVIDIA to build custom AI systems for the U.S. government. Karp says the deal centers on giving customers control over their compute, models, data stack, and alpha.

Stock up almost 6% this week
Deal addresses who controls models and weights

01:23Ontology layer

Why does Palantir's ontology matter for AI deployment?

Karp explains that large language models need an application layer to be valuable in battlefield, regulated, or manufacturing contexts. Palantir's ontology makes LLMs safe and useful without touching underlying data or transferring IP.

Ontology prevents data caching and IP transfer
Everyone is now copying the ontology, per Karp

04:10Control and trust

Why do enterprises distrust frontier labs?

Karp says enterprises believe they are paying for tokens that create no value while frontier labs take their data and the alpha of their business. He argues customers want to own the means of production rather than transfer their alpha to a third party.

Enterprises are livid in private, per Karp
He challenges viewers to call CEOs privately to verify

05:58Open models

Can Palantir's approach compete with frontier models?

Karp claims the value comes from model plus application layer plus compute together. He says Palantir can get an open model to the point of a frontier model while the client controls the weights.

Value = model + application layer + compute
Open models from China and NVIDIA are world class

08:30AI bubble

Is the AI market oversold and heading for a bust?

Karp says the technology is genuinely changing the course of history and does not need to be overhyped. He argues models have been irresponsibly oversold, which he blames for backlash like wealth taxes.

Compute plus ontology plus model is changing history
Karp rejects overselling but not the underlying technology

10:30Palantir business

How is Palantir's own business performing?

Karp says Palantir has more demand than it can supply, and looking at the financials you can see 15 to 18 billion dollars in free cash flow two years out. He criticizes investors who judge enterprise tech by their own consumer usage.

More demand than supply
15 to 18 billion dollars in free cash flow visible two years out

13:50China and the AI race

How does Karp view the U.S.-China AI competition?

Karp says the relevant tech centers are America, China, and Israel, and calls China a peer-to-peer adversary. He says it is binary — either they win or we win — and it is not a foregone conclusion that America wins.

China is a peer-to-peer adversary
America has more creativity but China lacks internal political dysfunction

15:49Israel and Iran

What does Karp think about Israel and the Iran situation?

Karp calls himself the most publicly supportive CEO of Israel and probably its most effective private critic. He says Iran has been degraded and there are non-public things that would be comforting to people.

Israel is an important partner not always fully aligned with the U.S.
Karp believes Iran has been degraded

Notes

The NVIDIA Partnership

Palantir is expanding its partnership with NVIDIA to build custom AI systems for the U.S. government; the stock is up almost 6% this week.
Karp says the deal came together around technical issues: who controls the models, who controls the weights, and who controls the value of the business.
What aligns Palantir with NVIDIA is giving technical customers control over their compute, models, data stack, and alpha — they want to own the means of production.

Why Enterprises Distrust Frontier Labs

Karp says clients using LLMs on the battlefield in Ukraine, Israel, and elsewhere run on top of Palantir's ontology, and those clients have lost trust in frontier labs.
The ontology application layer makes large language models safe and useful: it doesn't touch underlying data, prevents the model from caching data, and doesn't transfer IP.
He claims enterprises are

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