Alex Karp’s crusade against frontier AI labs continues.
During Palantir’s second-quarter earnings call on Monday, the CEO and cofounder reiterated his view that leading AI labs could absorb enterprises’ data, intellectual property, and expertise into their models and ultimately use that to compete with their customers.
“They deserve to colonize your enterprise,” Karp said, characterizing what he believed was the labs’ attitude.
The CEO said that enterprise companies are participating in tokenmaxxing — or “token self-pleasuring” as Karp called it during the call — at the cost of handing over their data and expertise to outside AI providers.
The comments reflect Palantir’s ongoing campaign around “AI sovereignty” — its term for enterprises retaining control over their data and workflows while remaining free to use or replace different AI models.
Karp said in July that AI models had been “irresponsibly oversold,” as business leaders privately expressed concerns to him that they’re handing over their “alpha” — or competitive edge — to AI labs while seeing little value in return.
“Something has gone completely wrong,” Karp said at the time. “The basic view among enterprises in this country is ‘I’m going to chillax and waste my time with tokens, I’m going to get no value, and they’re going to get my IP.'”
Palantir made sovereign AI a recurring theme throughout its latest earnings report, which showed a blockbuster quarter. The company reported a 93% year-over-year revenue increase, largely driven by its US business. US commercial revenue in particular jumped 149%. Palantir raised its full-year revenue outlook from about $7.65 billion to roughly $8.15 billion.
Its stock was up about 15% in after-hours trading late Monday.
Executives linked the company’s latest acceleration to customers seeking sovereign AI. Karp said customers should retain “maximal control” over their operations and prevent their “competitive advantage” from becoming training data for future models.
There’s a clear commercial interest in Palantir drawing that distinction. Palantir is a data analytics firm that doesn’t build a frontier model and instead sells the software and engineering work that connects models to an enterprise’s data and operations.
“Palantir has been the clearest counterexample to the claim that enterprise AI doesn’t scale past pilots, and accelerating growth makes that harder to argue with,” Jacbon Bourne, analyst for Emarketer, wrote. “What it doesn’t settle is whether AI eventually collapses the software layer Palantir occupies, which has weighed on its stock price this year.”
