Alex Karp, CEO and cofounder of Palantir, has intensified his criticism of leading AI labs. He accused them of attempting to take over enterprise businesses by absorbing their data, intellectual property, and expertise into their models.
During Palantir’s second-quarter earnings call on Monday, Karp stated the labs believe they “deserve to colonize your enterprise.” He described enterprise companies as engaging in what he termed “tokenmaxxing”—prioritizing short-term experimentation over long-term control of critical data.
AI sovereignty as a selling point
Karp’s comments align with Palantir’s push for “AI sovereignty,” a concept the company presents as vital for enterprises to maintain control over their workflows and data. The argument is that businesses should be able to use or replace AI models without losing ownership of their competitive advantages.
In July, Karp called AI models “irresponsibly oversold.” He cited private discussions with business leaders who felt they were handing over their unique edge to AI labs while seeing little benefit. “Something has gone completely wrong,” he said. “The basic view among enterprises in this country is that they waste time with tokens, gain no value, and surrender their intellectual property.”
Palantir’s latest earnings report highlighted this theme. The company reported a 93% year-over-year revenue increase, with U.S. commercial revenue rising 149%. It also raised its full-year revenue outlook from $7.65 billion to $8.15 billion. Shares climbed about 15% in after-hours trading after the announcement.
Executives linked the company’s growth to customers’ interest in sovereign AI. Karp argued that businesses must keep control over their operations to prevent their competitive advantages from becoming training data for future AI models.
Related: Nile Cruise and Stay Holidays — What You Need to Know
A commercial stake in the debate
Palantir’s stance is not just theoretical. The company sells software and engineering services that integrate AI models with enterprise data rather than building its own advanced models. This approach has become central to its pitch to customers reluctant to surrender control to external providers.
“Palantir has been the clearest counterexample to the claim that enterprise AI doesn’t scale beyond pilots,” said Jacob Bourne, an analyst at Emarketer. He noted that the accelerating growth makes the argument harder to dismiss. The challenge, he added, is whether AI will eventually disrupt the software layer Palantir occupies—a concern that has affected its stock price this year.
Karp’s warnings about AI labs have grown more frequent as Palantir’s financial performance improves. The company’s argument centers on the idea that enterprises, not AI providers, should decide how their data is used. Whether this message gains broader traction remains uncertain.
For now, the focus on sovereignty is delivering results. The earnings report showed rising demand for its AI-driven platforms, even as the debate over data ownership in enterprise AI continues.
Karp’s language—”colonize,” “token self-pleasuring”—is meant to provoke. It highlights a growing tension as AI adoption expands. Businesses want to use the technology, but providers stand to benefit from their data, creating friction.
The discussion extends beyond corporate strategy. Regional banks have also introduced financial support packages to help businesses handle similar challenges.
Leave a Reply