David Sacks said Palantir (NASDAQ:PLTR) CEO Alex Karp was “exactly right” in warning enterprises against handing proprietary data to frontier AI labs.
Speaking on the All-In Podcast episode aired Friday, Sacks said the real meaning of AI safety for enterprises is control over their compute, models, data stack and proprietary knowledge, or “alpha.” He argued that frontier labs risk absorbing clients’ know how and eventually turning it into competing products, leaving companies exposed if they do not own their AI infrastructure outright.
On Thursday, Sacks had made the same argument in a long post on X ahead of the podcast, framing it as a rejection of abstract alignment research in favor of practical enterprise control.
The South African-American technology entrepreneur and venture capitalist also argued that Anthropic and OpenAI have effectively formed a “duopoly” in the AI model layer. He warned that Anthropic’s push for stricter safety regulations could strengthen that dominance rather than limit it, leaving enterprises with fewer competitive alternatives.
The discussion on the All-In Podcast episode followed Karp’s CNBC appearance, dismissed by some media as a “crashout” but praised by Sacks as insightful.
Benzinga’s Take: The alignment between Sacks, Karp and industry analysts reflects a broader shift in enterprise AI spending, as companies increasingly weigh open source and on-premises models against proprietary frontier labs when deciding how to deploy AI without giving up control of their data and competitive advantage.
Disclaimer: This content was partially produced with the help of AI tools and was reviewed and published by Benzinga editors.
© 2026 Benzinga.com. Benzinga does not provide investment advice. All rights reserved.
To add Benzinga News as your preferred source on Google, click here.