Alex Karp: Most of the things Anthropic talks about in public are running on Palantir
Speakers
Alex Karp
— CEO, Palantir
Summary
CNBC's Sara Eisen interviews Palantir CEO Alex Karp about whether large language models like Anthropic can replicate Palantir's enterprise AI business. Karp forcefully dismisses the threat, arguing that LLMs are probabilistic and unsuitable for complex, high-stakes operations while Palantir's implementation layer is irreplaceable. He reveals that much of what frontier AI companies discuss publicly runs on Palantir and expresses strong confidence in his company's moat.
- Karp argues LLMs cannot replace Palantir in complex enterprise or defense settings
- He says most public AI demonstrations from companies like Anthropic run on Palantir
- Karp describes pure LLM firms as culturally and technically ill-suited for mission-critical work
- He views short-term investor skepticism about Palantir as uninformed and bets on implementation value
- The conversation touches on competitive dynamics with Anthropic, calling the relationship more nuanced than friend or foe
Ideas
Palantir's enterprise moat is intact.
Palantir's deep enterprise integration and implementation expertise create a durable competitive moat that large language models cannot replicate. LLMs are probabilistic and unsuitable for mission-critical tasks like manufacturing or defense, while Palantir is the backbone that actually runs the frontier AI work that companies like Anthropic showcase publicly. Short-term investor fears about LLM disruption are unfounded, and Karp is betting heavily that implementation will capture the value.
