Palantir has become one of the most recognizable technology companies working with the U.S.(nové okno) government. Its software is used by military and intelligence organizations, while its work with law enforcement and immigration agencies has made it a focus of privacy and civil liberties debates around government surveillance.
Speaking broadly, Palantir’s software(nové okno) helps organizations connect and make sense of large amounts of existing data. That capability can be useful for businesses and governments alike, but its applications become more consequential when the data involves military operations, policing, or immigration enforcement.
Palantir’s influence extends beyond the company itself. A growing group of defense and surveillance startups has emerged from the same network of former Palantir employees, investors, and executives. Some build weapons. Others build the software that connects sensors, drones, databases, and people. Several work directly with Palantir or with the same large defense contractors that have supplied the U.S. military for decades.
You could call this network the “Palantir Mafia,” borrowing the language once used for the PayPal alumni who went on to reshape Silicon Valley. Within it is a smaller group with a particularly strong connection to military and surveillance technology.
The companies are different, but the pattern is striking. People leave Palantir, carry experience from government and military projects into new businesses, and often find the capital and relationships needed to sell those products back to governments.
The result is a new layer of the military-industrial complex, built around software, data, artificial intelligence, and venture capital.
Meet the companies spreading Palantir’s model
While all of the companies in the “Palantir Mafia” perform different jobs, their connections create a recognizable pattern. Former Palantir employees are carrying knowledge, relationships, and ways of working into new defense companies, while investors and established contractors help those companies reach government customers.
Epirus
Epirus, founded by Joe Lonsdale (who also co-founded Palantir), makes Leonidas, a high-power microwave system designed to disable drones. The system can be deployed in different configurations, including on vehicles, and it has been integrated with General Dynamics Land Systems’ Stryker combat vehicle.
Epirus has also joined Palantir’s Warp Speed program, giving the companies another direct connection.
Valinor
Valinor is newer and less flashy. Founded by former Palantir executive Julie Bush, the company describes itself as tackling “unsexy” problems in defense. Its products include systems for drones and other military infrastructure, and its software can connect with Palantir’s platforms. The name itself is a nod to the works of J. R. R. Tolkien which also inspired the names of Palantir and Anduril.
Valthos
Valthos takes a different approach. The company develops AI systems intended to detect and prevent biological weapon attacks. Its co-founder and CEO, Kathleen McMahon, was previously the Head of Life Sciences at Palantir. Valthos reflects a broader concern among some technology executives that advances in AI and biotechnology could make biological threats easier to create.
Anduril
Then there is Anduril, the biggest and best-known company in this group. Founded in 2017 by three former Palantir employees, including Trae Stephens, who is also a partner at Founders Fund. Its products include surveillance and combat drones, autonomous systems, and Lattice, a software platform that uses sensor data to identify and track objects.
Anduril’s growth shows how far this model has moved. The company was valued at $61 billion after a 2026 funding round, putting it in a very different financial category from most defense startups.
Its technology has also crossed between military and domestic security. Lattice has been used in border surveillance, showing how systems developed for national security can have applications closer to home.
Kela
Kela, an Israeli defense technology company, follows a similar software-first approach. It provides a platform that can connect drones, sensors, and weapons from different manufacturers into a single system. Its co-founder and president, Hamutal Meridor, previously ran Palantir’s Israeli operation. Kela has also secured work connected to the Israeli military.
Rivet Industries
Rivet Industries was founded by David Marra who previously worked at Microsoft on a military mixed-reality headset program. Marra later then spent a year at Palantir. Rivet now develops wearable systems for soldiers. In 2025, the Army awarded Rivet a $195 million Soldier Borne Mission Command contract(nové okno), and the company works closely with Palantir. (nové okno)
Peregrine Technologies
Peregrine Technologies was founded in 2018 by Nick Noone, who previously ran Palantir’s U.S. Special Operations business. At Palantir, Noone worked with data to identify ISIS members in Syria. He later took that experience to a small police department in California, initially building Peregrine’s technology for free before turning it into a commercial product.
Peregrine now provides software to hundreds of agencies. Its platform brings together information that police departments already have access to, making it easier to search and visualize data from different sources.
From foreign battlefields to local police
Palantir’s own history includes a predictive policing project in New Orleans. The system used information such as arrest records, social media posts, police contacts, and family relationships to generate assessments about people’s likelihood of being involved in violence. The project was later examined by researchers and civil liberties organizations concerned about the lack of transparency and the risks of assigning people police risk scores. The Harvard analysis of predictive policing(nové okno) provides additional context on the New Orleans system.
Peregrine says it does not operate a predictive policing system of that kind. Instead, the company says it helps agencies organize and visualize existing information rather than adding new surveillance sources.
The problem is that “existing information” can still be extensive. Once those systems are connected, the question is no longer simply what a company collects itself. It becomes a question of what information a police department can bring together and what conclusions it can draw from it.
That is one of the most important parts of the Palantir model. A company does not necessarily have to collect every piece of information itself to make a large amount of information more useful.
Bankrolled by the Founders Fund
There is another connection running through several of these companies: Founders Fund.
Co-founded by Peter Thiel, the venture capital firm backed Palantir and has invested in or supported several companies in this wider network. Trae Stephens is particularly important because he sits at the intersection of several parts of it. He worked at Palantir, became a Founders Fund partner, and helped co-found Anduril.
That makes the network less mysterious. You don’t need a formal organization with membership cards or a central office. Shared investors, former colleagues, and existing government relationships can create many of the same effects.
Founders Fund has also been unusually explicit about its investment philosophy. Its manifesto argues that successful entrepreneurs often have what it calls a “near-messianic attitude(nové okno)” and believe their company is essential to making the world a better place. It also argues for independent, contrarian thinking rather than simply following the rest of the venture capital industry.
Defense has traditionally been difficult territory for venture capital. Government procurement can be slow, contracts are complicated, and established contractors already have decades of relationships and infrastructure.
The newer defense companies are approaching the problem differently. They bring venture-backed software culture, rapid product development, and consumer technology ideas into a market that has historically been dominated by large contractors.
The money follows the opportunity. So does the talent.
The startups aren’t replacing the old defense giants
It would be easy to tell this story as a simple battle between Silicon Valley startups and traditional defense contractors.
Epirus has integrated its Leonidas system with General Dynamics’ Stryker vehicle, and General Dynamics Land Systems has invested in Epirus. Rivet names Northrop Grumman as a partner. Anduril has integrated Lockheed Martin radar into its battlefield software.
The relationship is closer to an expanding ecosystem than a hostile takeover. That matters because the growth of defense technology startups does not necessarily shrink the role of companies such as General Dynamics, Northrop Grumman, and Lockheed Martin. Instead, startups can provide new software, sensors, weapons, and interfaces that become part of existing military systems.
The distinction also changes how we should think about disruption. In consumer technology, a startup can sometimes replace an incumbent by building a better product and taking its customers. Defense procurement works differently. A new company can win a contract while still relying on established manufacturers, military infrastructure, and government procurement systems.
The result can be a larger market in which startups and traditional contractors both benefit.
The scale of that market is already visible. In 2025, the U.S. Army awarded more than $350 million for work on virtual displays for soldiers, with money going to both Anduril and Rivet. The two companies can compete for contracts while also building products that fit into the same broader military technology ecosystem.
The companies don’t need to destroy the old military-industrial complex to change it. They can become another layer within it.
Who decides what this new military industry becomes?
Governments want better intelligence, faster software, autonomous systems, counter-drone weapons, and tools that can connect huge amounts of information. In some cases, those technologies can give soldiers or investigators information they could not easily obtain before.
But greater technical capability also changes the balance of power.
When a company develops software that can connect databases, sensors, cameras, weapons, and government agencies, the technical decision about what the system can do becomes a political decision about what governments are able to do.
That is why the story of Palantir’s former employees matters beyond a list of startups. The network shows how quickly expertise can move between military intelligence, private technology companies, domestic policing, and surveillance.
It also shows how venture capital can influence what gets built. Dwight Eisenhower warned about the military-industrial complex in his 1961 farewell address(nové okno). He argued that the country needed to guard against the acquisition of “unwarranted influence” and warned that the growth of military and industrial power could endanger democratic processes. He ended with an appeal for an “alert and knowledgeable citizenry.” (nové okno)
More than six decades later, technology has changed but the basic question has not.
Who has the power to decide how military and surveillance systems are built, where they are deployed, what data they can use, and how their capabilities expand? This becomes harder when the systems are developed by private companies, financed by investors, and sold through government contracts that most citizens never see.
The companies emerging from Palantir’s orbit are not a single organization, and they do not all have the same products or goals. Some are building defensive systems. Some are developing tools for policing. Some are working on ways to connect military hardware and software.
What connects them is a network of people, capital, technology, and government relationships.
That network is becoming an increasingly important part of how the United States and its allies approach defense and surveillance.
The question is not whether these technologies will shape the future. They already are. The question is who gets to shape them, and who gets to decide where the power they create should stop.
