Palantir Technologies posted second-quarter profit of roughly $1 billion and raised its full-year revenue guidance, sending shares up 12% and defying a slide driven by concerns about the artificial intelligence software trade.
The company earned $1.07 billion in net income, or 41 cents per share, up from about $329 million, or 13 cents per share, a year earlier. Revenue climbed 93% to $1.94 billion, beating the $1.80 billion that analysts surveyed by LSEG had expected. Adjusted earnings per share of 41 cents topped the 35-cent estimate. Karp wrote that Palantir generated more profit in a single quarter than it did in total revenue in the same period the year before.
U.S. commercial revenue surged 149% to $764 million, and when accounting for compounding, has jumped 380% since 2024. The company now expects U.S. commercial revenue to exceed $3.42 billion in 2026, up from prior guidance of $3.22 billion. Remaining U.S. commercial deal value more than doubled to $6.24 billion. U.S. government revenue grew 90% to $809 million. The company is widely known for selling its software to the U.S. government and military, but its commercial revenue has accelerated.
CEO Alex Karp told CNBC that no business at Palantir’s scale has grown even half as much. “Forget consensus,” he said in an exclusive interview with CNBC’s Seema Mody. He added that the strong growth looks like it will continue for at least another 18 months.
Karp calls AI labs ‘Marxist’ and pushes open-weight models
In a letter to shareholders, Karp wrote that there are “Marxist overtones and undertones” to Palantir’s business. He said others, including many building large language models, “intend, knowingly or otherwise, to capture the means of production of their purported partners.” The CEO, who studied philosophy and earned a PhD in social theory, has repeatedly warned that AI frontier labs are too untrustworthy for enterprises.
On the quarterly conference call with Wall Street analysts, Karp expanded on the analogy. He questioned whether companies would buy into a future where their work helps adversaries win and a small group of people control the means of production. He criticized what he described as enterprises signing up for token-based arrangements that allow AI labs to migrate their intellectual property, know-how, and expertise into the labs’ own models, enabling those labs to build competitive businesses that do not require the original partners. He said the labs are doing this for what they believe are moral reasons, calling them superior and saying they deserve to colonize the enterprise.
The theory points to a significant list of companies that partnered with or paid for Anthropic and OpenAI while the AI labs launched similar businesses ranging from design tools to healthcare operations, legal services, and drug discovery. According to TechCrunch, none of these companies are economic villains or heroes any more than other for-profit companies are, and AI is growing so quickly that there is clearly room for all, as Palantir’s results show.
Raised outlook and calls for competition
Karp has been a vocal proponent of open-weight models and reducing reliance on the token model from frontier labs, particularly as Chinese tools have closed the gap. Last month, Karp and Palantir joined tech heavyweights in a letter urging the government not to restrict open-weight models. “We need competition if we’re going to keep model companies honest, which is the same thing as enterprise software,” Karp told CNBC. He said the way to win in America is to compete, and that open models will have to become as good as Chinese open models.
Palantir lifted its full-year revenue guidance to between $8.15 billion and $8.16 billion, up from a prior range of $7.65 billion to $7.66 billion. The company serves model-agnostic AI and analysis software to governments and enterprises, allowing organizations to control their data as well as their AI exhaust, including prompts, orchestration, and context. Before the report, shares had fallen 29% this year due to broadening concerns that the artificial intelligence software trade is running out of steam and growth could slow.