The Technological Republic

Alexander Karp had been the chief executive of Palantir Technologies for more than two decades when he and Nicholas Zamiska, his longtime collaborator and the company’s head of corporate affairs, sat down to write a book about what they believed had gone wrong with the relationship between American technology and the American state. The Technological Republic appeared in early 2025, at a moment when that relationship was being renegotiated in real time. Silicon Valley founders who had spent the 2010s announcing their refusal to work on Pentagon contracts were rediscovering defense as a market and a vocation. The Department of Defense was openly soliciting commercial software companies for everything from battlefield logistics to autonomous systems. And the war in Ukraine had made it impossible to pretend that the question of who builds the most capable weapons could be left to history. Karp and Zamiska wrote into that moment, and the book reads as both a diagnosis of how the West reached it and a manifesto for what its institutions should do next.

The central argument is that the United States and its allies have lost the civic and intellectual confidence that once allowed engineers, soldiers, and politicians to share a single national project. The authors trace this loss to a drift that began in the late twentieth century, accelerated through the 2000s, and reached a kind of cultural peak in the late 2010s when employees at Google walked out over a Pentagon contract for an artificial intelligence imaging project called Maven, forcing the company to abandon the work. To Karp and Zamiska this was not a one-off labour dispute but a symptom of something deeper: a generation of the most talented engineers in the world had been taught that the institutions of their own country were morally suspect, that defense work was disreputable, and that the highest calling of a software company was to optimise consumer attention or food delivery rather than to build the technological substrate of national power. The book argues that this is a historical anomaly. The Silicon Valley that produced the integrated circuit, the personal computer, the internet, and modern aerospace was built in close partnership with the American national security state — funded by it, staffed by people who had worked in it, organised around problems it had posed. Severing that connection was a choice, the authors argue, and a choice the West cannot afford.

To make the case, the book moves between intellectual history, business memoir, and political argument. Karp brings an unusual background to it. Before founding Palantir with Peter Thiel and others in 2003, he had earned a doctorate in neoclassical social theory in Germany, working in a tradition descended from Jürgen Habermas and the Frankfurt School. That training shows up throughout the book in the form of a Continental European sensibility toward questions of legitimacy, civic religion, and shared purpose — a register that is unusual in writing by an American tech executive. The book reads in places like a meditation on what the political philosopher Carl Schmitt called the friend-enemy distinction, retooled for an age in which the West’s enemies build hypersonic missiles and surveillance states while its own engineers debate the colour of an emoji. The voice is sharper than the typical chief-executive memoir. It is also, in places, more openly elegiac. Karp is repeatedly described in the press as a Bernie Sanders donor and a former hippie; he is not a partisan figure of the American right, even though his book has been embraced by it.

The structure of the book moves through several connected arguments rather than offering a clean chronological narrative. An opening section sets out what the authors call the slow disengagement of Silicon Valley from the work of the state. They argue that the original architects of American computing — figures associated with Bell Labs, Xerox PARC, Lockheed’s Skunk Works, and the early days of Stanford’s engineering school — would not recognise the values culture of contemporary big tech. The book reaches back to the Manhattan Project, to ARPA’s role in seeding the internet, to the Reagan-era Strategic Defense Initiative, and to the early intelligence-community customers who funded the first generation of in-memory analytical databases. These are not nostalgic detours; they are evidence for a thesis. The authors want the reader to see that the great epochs of American technological achievement came when scientists and engineers accepted that they were participants in a national project, not detached vendors of consumer goods.

A second strand of the argument turns to what Karp and Zamiska see as the cultural conditions that made the disengagement possible. Here the book becomes a critique of the contemporary American university, of what they regard as the hollowing out of the humanities, and of an academic culture that has lost the willingness to make strong claims about the moral foundations of liberal democracy. The authors describe a generation of graduates who arrive at technology companies fluent in the vocabulary of identity and harm but uncertain about whether their own civilisation is worth defending. Without a shared answer to that second question, they argue, you cannot ask engineers to work on weapons; you cannot ask anyone to make the trade-offs that engineering for the state requires. This part of the book has been the most controversial and has drawn the most pointed rebuttals from reviewers on the cultural left. It is also the part that gives the book its title. A technological republic, in Karp and Zamiska’s usage, is a polity in which technical capability and civic conviction reinforce one another — where the people who can build also know what they are building for.

The book’s middle stretches deal with what the authors call the engineering mindset, and here Karp draws openly on Palantir’s own history. The company’s early work for the United States intelligence community on counterterrorism software, its later work for the Department of Defense, and its eventual public listing become illustrations of a thesis about how serious software is actually built. The authors describe an environment that prefers small teams to large committees, that allows engineers to be embedded directly with the operators who will use the product, that treats requirements documents as starting points rather than contracts, and that accepts a high tolerance for friction with bureaucracy. They contrast this with the procurement style of the traditional defense primes, in which complexity and cost are accumulated through layers of review. The book argues that the United States military will not be able to field the systems it now needs — autonomous platforms, software-defined sensor networks, AI-enabled targeting — by relying on the procurement model that delivered the F-35. Something closer to the commercial software model has to take its place, and that, in turn, requires founders and investors who are willing to build companies that orient themselves around the state as customer.

Throughout, the authors return to the role of Palantir itself. The book is candid that the company’s stance has not been costless. Karp and Zamiska describe the recurring waves of criticism that have followed Palantir’s contracts with United States Immigration and Customs Enforcement and with the Department of Defense, the boycotts and walkouts that other technology firms have used as warnings, and the personal cost of refusing to back down. They argue that the alternative — a Silicon Valley that quietly refuses the most difficult and consequential work — is far worse. They are openly proud of Palantir’s role supplying Ukraine with software for battlefield analytics and command-and-control coordination after the Russian invasion in February 2022, and they treat that work as evidence that a technology company can choose a side without dissolving into a defense contractor. The book is at its most concrete in this register: when the authors are describing actual systems, actual customers, and actual decisions, the argument tightens.

The argument widens again when it turns to China. Karp and Zamiska treat the People’s Republic as the central organising fact of the twenty-first-century technological competition. They argue that China has built a model in which the state and its technology firms operate as a single instrument of national power, in which there is no question of whether engineers will work on military or surveillance applications, and in which the long time horizons of the Chinese Communist Party allow for sustained investment in capabilities that American venture capital, on its own, would not finance. The authors do not romanticise this model — they are clear-eyed that the cost of the Chinese arrangement is borne by the country’s own citizens — but they argue that the West cannot beat it by pretending the competition does not exist. The book pushes back hard against the idea that the responsible position is to call for a global moratorium on autonomous weapons. The authors observe that adversaries will build what they will build, and that the question facing democracies is not whether to develop these systems but whether to develop them under conditions of democratic oversight or to cede the field. This is one of the more important political claims of the book, and it has shaped the way the argument has been received in Washington and in European capitals.

The reception of The Technological Republic has been intense and divided. Within the defense-technology community, where authors such as Christian Brose, Raj Shah, and Chris Brose’s collaborators have spent years arguing for a faster, software-led model of military modernisation, the book has been received as a popular synthesis of arguments that had previously been read mostly in policy journals. Defense-oriented venture capitalists and the new generation of dual-use founders — at companies such as Anduril, Shield AI, and Saronic — have largely endorsed it; some have used its language to recruit. Conservative reviewers have welcomed the book’s critique of campus culture and its insistence on the moral legitimacy of national defense. Reviewers on the cultural left have been sharper, arguing that the authors mistake their own commercial interest for a civilisational diagnosis, that the book’s brisk dismissal of the concerns raised by Google’s Maven walkout undersells the genuine ethical questions involved in building target-identification software, and that Karp’s billionaire status undermines his claim to be diagnosing the sickness of the meritocratic order. There has also been pushback from inside the academy, where some critics have argued that the authors’ picture of contemporary humanities scholarship is a caricature. The book is not above these critiques. It is most persuasive when it is making historical and operational arguments about how things actually get built; it is most vulnerable when it diagnoses the inner life of people the authors have not met.

In the broader landscape of recent writing on artificial intelligence in war, the book sits in a specific niche. It is less interested in the granular doctrine of human-machine teaming than Paul Scharre’s work; less interested in the bureaucratic battle for Pentagon procurement reform than Christian Brose’s; less focused on the geopolitics of the semiconductor supply chain than Chris Miller’s. What it offers instead is a high-level argument about the cultural and political preconditions for any of those efforts to succeed. A reader who wants to understand why the contemporary American technology sector is reorganising itself around defense — why companies that would not have existed a decade ago are now valued in the tens of billions of dollars, why an unprecedented number of Stanford and MIT graduates are taking jobs at firms working on autonomous systems, why the European Commission has begun to talk seriously about an industrial defense base — will find this book a useful map of the underlying argument. It pairs well with Scharre’s work on the autonomous-weapons debate, with Brose’s case for procurement reform, and with the operator-side writing of figures such as Stanley McChrystal on adapting military organisations to fast-moving technological environments.

Some elements of The Technological Republic will age better than others. The book’s specific cultural diagnoses are tied to a particular moment in American academic and corporate politics, and that moment is already shifting; in a decade, the section on campus culture may read as a period piece. The arguments about Silicon Valley’s drift, the procurement gap, the need for a closer alignment between technical talent and national mission, and the case against treating autonomous systems as uniquely abhorrent are likely to remain in circulation much longer. So is the book’s basic claim that capability and conviction are inseparable — that you cannot build the systems a democracy needs to defend itself if you have spent a generation telling its most talented engineers that their country is not worth defending. Whether or not readers accept the politics, the book closes an argument that defense-technology insiders had been making for years and brings it into the mainstream of the conversation about the West, its competitors, and the wars it may yet have to fight.