Palantir Raises Guidance After Revenue Soars 93% and U.S. Commercial Doubles to $764 Million
The maker of Foundry and AIP closed the second quarter with $1.94 billion in revenue, $764 million from U.S. commercial, and raised annual guidance to $8.16 billion after Alex Karp described demand as otherworldly.
Palantir Technologies closed the second quarter of 2026 with revenue of $1.94 billion, a 93% year-over-year increase, and raised its full-year revenue guidance to a range of $8.15 to $8.16 billion, equating to 82% growth compared to 2025. The figures were released on the evening of August 4 and exceeded Bloomberg's consensus, which had estimated $7.7 billion for the year.
The surge came from U.S. commercial sales. This segment, which includes corporate clients outside the federal government, grew by 149%, reaching $764 million in the quarter. The company raised its annual projection for this segment from $3.22 billion to above $3.42 billion. For a company historically associated with defense and intelligence contracts, this is the profile reversal that analysts have been demanding since the shift in the Foundation cycle.
What Alex Karp Calls AI Sovereignty
CEO Alex Karp described demand on the call as "otherworldly" and attributed much of the surge to what Palantir calls "AI sovereignty": companies wanting to run models on proprietary data without handing it over to a third-party model provider. This argument applies to AIP, the Artificial Intelligence Platform that Palantir integrated with Foundry, aiming to capture clients who have previously piloted LLMs in the cloud and now want to bring the workload in-house.
The quarter's Rule of 40, a combined measure of growth and operating margin widely used in SaaS, stood at 155%. Comparable-sized companies in enterprise software rarely exceed 60, and the best early-stage companies range from 70 to 80. It is worth noting that the base is still small for a company whose market cap approaches $400 billion: 155% Rule of 40 on $1.94 billion of quarterly revenue does not support the multiple that the market pays today, and sell-side analysts have been pointing this out as well.
The Counterpoint That Karp Ignores on Stage
Loop Capital, which carried the most optimistic recommendation on the street with a price target of $240, published a note acknowledging the outperformance but warned that the concentration of growth in a few large contracts makes revenue vulnerable to renewal delays. Rishi Jaluria from RBC is the known skeptical voice: he maintains that the discount for running Foundry with AIP in regulated on-premise environments explains the growth rates and not a structural advantage of the model. None of the two recommend selling, but both view the price as too expensive for any hiccup.
Insights for Consultancies and the Brazilian Market
Accenture, Deloitte, KPMG, and PwC have built implementation practices around Foundry throughout 2024 and 2025. The leap in U.S. commercial sales has a direct effect on these firms' pipelines because every dollar of Palantir licensing has historically pulled between $2 and $4 in deployment, integration, and change management services. For the technology partner at these firms, Palantir's quarter serves as a green light to bolster AIP squads and negotiate fatter SOWs.
In the UK, where the NHS signed a federated data platform contract with Palantir in 2023, the argument for AI sovereignty is likely to accelerate similar decisions in European ministries that resist sending sensitive data to American hyperscalers. There have already been reports of discussions with the German health service regarding equivalent architecture.
In Brazil, Palantir has a smaller presence concentrated in government, defense, and a portion of energy. The local commercial trigger depends less on Karp's narrative and more on what Petrobras, Vale, and a handful of banks decide regarding unified data platforms capable of running governance agents. A spectacular American quarter aids the sales pitch in São Paulo, but it alone does not close contracts in Brazil: here, the bar still goes through Databricks, Snowflake, and the combination of SAP plus BTP for ERP-centric clients.
The useful insight for the CIO is not to bet on Palantir. It is to ask whether the internal architecture, regardless of the provider, can honestly say the same thing Karp claims: that critical data runs models without leaving the owner's perimeter. As long as the answer is no, the incumbent is exposed to those who can answer yes.