Palantir Technologies CEO Alex Karp on Monday reiterated his criticism of leading artificial intelligence labs, arguing that enterprises risk losing control of their intellectual property and competitive advantages by engaging with frontier AI models. The remarks came during Palantir’s second-quarter earnings call and a subsequent interview with CNBC, where Karp framed the issue as a threat to business autonomy.
Palantir reported a 93% year-over-year revenue increase for Q2, driven largely by its U.S. commercial business, which saw a 149% surge in revenue. The company also raised its full-year revenue outlook from approximately $7.65 billion to $8.15 billion. Palantir’s stock rose 15% in after-hours trading following the earnings announcement.
During the earnings call, Karp described what he characterized as the attitude of frontier AI labs toward enterprises, stating that they believe they ‘deserve to colonize your enterprise.’ He argued that businesses participating in AI adoption—particularly through token-based interactions—are effectively ‘tokenmaxxing’ or engaging in ‘token self-pleasuring,’ a term he used to describe actions that prioritize superficial engagement over meaningful outcomes. Karp claimed these practices allow AI labs to absorb customer data and expertise, which could later be used to compete against those same enterprises.
In the CNBC interview, Karp expanded on these concerns, stating that enterprises should avoid being ‘drug addicted’ to AI models controlled by external labs. He warned that businesses risk a future where they ‘own nothing,’ lack profitability, and lose control over their operations. Karp advocated for ‘sovereign AI’, a model where enterprises retain full control over their data and AI workflows by using in-house or third-party applications that do not require sharing proprietary information with model developers.
Karp’s comments targeted companies like Anthropic and OpenAI, which have stated that customer data is not used to train their models unless explicitly opted in by the user. Despite these assurances, Karp dismissed them as insufficient, asserting that major enterprises—including government clients—remain skeptical of the value proposition and concerned about data security.
Palantir’s push for sovereign AI aligns with its broader business strategy, which emphasizes enterprise control over AI deployment. The company’s financial performance suggests strong demand for its approach, particularly among U.S. commercial clients. Karp emphasized that customers should retain ‘maximal control’ over their operations to prevent their ‘competitive advantage’ from becoming training data for future models.