Palantir CEO Alex Karp is escalating his criticism of major AI labs, warning that businesses are losing control of their own intellectual property. During a recent interview, Karp argued that these labs are effectively trying to create dependency by pushing models that require enterprises to sacrifice their data sovereignty. He contends that firms should maintain their own models internally rather than ceding control to third-party providers.
Karp suggested that businesses are waking up to the risks of handing over institutional keys to these labs. He emphasized that the current structure favored by many frontier model makers leads to a loss of profitability and control. According to Karp, companies have begun to question why they should pay to support systems that do not yield direct business utility or protect their core assets.
While OpenAI and Anthropic have stated that customer data remains secure, Karp remains unconvinced. He believes the industry needs a different approach that prioritizes data privacy and operational independence. This stance aligns with Palantir’s broader mission to advocate for what he terms AI sovereignty, where organizations keep their data stack under their own management.
This debate has gained significant traction among other leaders in the technology sector. Some investors and executives have praised Karp for his focus on maintaining control over compute and data infrastructure. As organizations navigate the current AI landscape, the question of whether to rely on closed-source providers or maintain internal control remains a central point of contention in boardrooms everywhere.