Palantir CEO: AI companies have to be nationalized to cap their liability
Speakers
Alex Karp
— CEO, Palantir
Summary
Alex Karp, Palantir co-founder and CEO, discusses AI safety warnings, arguing the real issue is clients' IP being absorbed by closed AI models and the resulting liability risk. He says closed-model AI companies may need nationalization to cap liability, while open-weight models plus post-training and an application layer are cheaper and safer. Karp also highlights Palantir's role in military and commercial AI and its AI sovereignty initiative with NVIDIA.
- Alex Karp joins CNBC to discuss AI safety and Palantir's positioning.
- Karp says closed AI models can capture customers' IP and alpha without honest disclosure.
- He argues liability risk could force AI companies toward government nationalization.
- Karp says open-weight models with an application layer are cheaper, safer, and more controllable.
- He says Palantir is at the forefront of military AI and dominates the U.S. commercial application space.
- Karp mentions NVIDIA as a partner in the AI sovereignty revolution.
- He says he respects Dario but disputes the safety-versus-regulation framing.
Ideas
Palantir leads sovereign AI application layer
Karp argues Palantir is at the forefront of military and classified AI, built the first functional AI technology, dominates the U.S. commercial application space, and is central to the AI sovereignty revolution. He says clients are angry that closed-model providers capture their IP and alpha, while open-weight models plus post-training and an application layer are cheaper, safer, and allow institutions to control their future—positioning Palantir favorably.
