Photograph Source: UK Government – Deputy Prime Minister Oliver Dowden attends AI Summit – CC BY 2.0

Palantir CEO Alex Karp’s description of the company’s quarterly earnings in August 2025 as “once in a generation” wasn’t far-fetched. It surpassed $1 billion in quarterly revenue for the first time, after only turning profitable in 2022. Its surge has made it an outlier: still too small by revenue to crack the Fortune 500, its roughly $400 billion market value by December (more than double what it was in January) places it among the 25 most valuable companies in the world.

Palantir’s ascent over the last couple of years comes after two decades of operating largely outside the public spotlight, at least compared to Big Tech giants or Elon Musk’s companies. Founded in 2003, it went public only in 2020 and has fewer than 4,000 employees. It runs no public advertising and sells no consumer-facing products, focusing entirely on developing software platforms for government agencies, companies, and nonprofits.

Part of Palantir’s success stems from what amounts to a monopoly. In his 2014 book Zero to One and follow-up essay “Competition Is for Losers,” co-founder Peter Thiel argued that the most successful companies are “creative monopolies,” and are so good at what they do that no real substitute exists. Citing Google’s search engine, Thiel wrote that monopoly profits let companies invest more in workers, products, and long-term innovation, while disrupting stagnant incumbents.

Understanding Palantir’s business monopoly is difficult because both the firm and many of its clients keep its platforms largely out of view. Named after the seeing stones in The Lord of the Rings, Palantir grew out of PayPal’s fraud detection project. Its operating systems now combine data from communications, supply chains, operations, and other sources to reveal patterns and support faster decision-making. Over time, Palantir has layered these tools into integrated management systems, with advisers often deployed to simplify adoption.

At the center is its “Ontology,” a central brain that links an organization’s data to show how everything fits together, “transforming raw data into valuable knowledge.” Originally a proto-AI system, it has been enhanced since 2023 with Palantir’s Artificial Intelligence Platform (AIP), which uses large language models, generative AI, and coordinated AI agents that work together. This setup reduces scaling constraints and enables a more “plug-and-play” approach that can be managed remotely.

Clients can build comparable systems in-house, but doing so is slow, expensive, and often difficult to maintain. As the first mover, Palantir can deploy advanced, customizable platforms in weeks, and the combination of performance, convenience, and high switching costs tends to lock customers in.

Developing the Domestic Engine

Palantir’s platforms have been used in the U.S. for decades, initially by agencies seeking advanced planning and security tools after 9/11. Early access to unique government datasets trained its systems on everything from modern databases to decades-old legacy software. Its tools can integrate radar, satellite imagery, photographs, communications records, human intelligence, and more. Palantir’s first customer was the CIA, whose venture capital arm, In-Q-Tel, bet early on its ability to organize massive datasets for counterterrorism, and is widely believed to have aided U.S. efforts to locate Osama bin Laden.

In 2008, Palantir launched Gotham, one of its flagship platforms. Using police reports, phone records, social media, and other data, Gotham can create detailed personal profiles and relational maps that help uncover hidden networks, patterns, and anomalies. Continuously updating, it can incorporate new layers of data and platforms as they emerge.

ICE, a Palantir customer since at least 2011, uses customized systems like FALCON and Investigative Case Management (ICM), and in April 2025 signed a new $30 million contract to track real-time migrant movements, prioritize targets, and accelerate deportation procedures. The software uses biometrics, travel and visa records, geolocation, and vehicle and phone data to flag individuals and streamline the enforcement process. While the contract prompted protests in multiple U.S. cities, Palantir’s lack of consumer-facing products makes it less vulnerable to public pressure campaigns.

Similar Palantir systems have been used across the country for various forms of predictive policing. Public backlash led some agencies to restrict or scale back their programs, but a few have quietly restarted.

Palantir has long cast itself as a disruptor. During an early 2025 earnings call, the company prepared for expansion into more government sectors, Karp said the company “loves disruption,” while CTO Shyam Sankar derided traditional “forever software projects” as “sacred cows of the deep state.” The language fit neatly with Elon Musk’s Department of Government Efficiency (DOGE), which, according to WIRED, in April, tapped Palantir to help build a “mega-API” to harmonize data across agencies such as the Internal Revenue Service.

Palantir previously successfully sued the U.S. Army in 2016 for shutting “the company’s commercial offering out of the competition.” Its push to break open defense contracting continues with its support for the Fostering Reform and Government Efficiency in Defense Act (FoRGED Act), aimed at reducing the dominance of incumbents like Boeing and Lockheed Martin.

Yet Palantir is increasingly becoming an incumbent itself. It secured a 10-year, $10 billion contract with the U.S. Army in August 2025, its largest ever, and will integrate its Army Vantage platform across the force. Previously, in June, “the U.S. Army created a special unit and swore in senior executives of the top four American AI companies, including Palantir, as senior advisers and lieutenant colonels in the Army Reserve,” stated an article in PassBlue.

[Content truncated due to length…]

From CounterPunch.org via this RSS feed