The War to Come

What happens when the tech companies realise governments can be their biggest customers?

It has become increasingly obvious that a significant shift has occurred in the relationship between Silicon Valley and the American government. These two forces have been at odds with each other for many years—caught in a never-ending argument over regulation and data protection. The tech companies typically see the government as a limiting force, slow-moving, wasteful, stifling. The government sees the tech companies as rapacious, underhanded, arrogant. It hardly matters that both sides are reliant on each other: the government has long guaranteed the environment in which the tech companies can thrive; the tech companies generate huge amounts of employment and taxable revenue (not to mention American cultural propaganda) which the government depends on. Nonetheless, there has been a mistrust between them that has led to bickering in the media and adversarial congressional hearings.

That situation has changed. The most visible example of the reversal has been the spectacle of the major tech company CEOs lining up together at Trump’s inauguration. As worrying as those images were—and I felt a genuine shiver of fear when I first saw them—the shift I’m concerned with here is something different. Where the FAANG CEOs are aligning themselves with the current administration because that is the prudent thing to do from a business perspective, others have been working to embed themselves within the day-to-day operations of government, particularly the military. We are right now witnessing the early fruits of a multi-decade strategy to disrupt both the established Silicon Valley model for what a tech company is and does, as well as the traditional habits of government procurement practice.

Software is Eating the World

When the venture capitalist Marc Andreessen published his 2011 op-ed in the Wall Street Journal, ‘Why Software is Eating the World’, he was speaking most effusively about the consumer internet—the world of apps, websites, and social media. “More and more major businesses and industries are being run on software and delivered as online services—from movies to agriculture to national defense,” he wrote. “Many of the winners are Silicon Valley-style entrepreneurial technology companies that are invading and overturning established industry structures. Over the next 10 years, I expect many more industries to be disrupted by software, with new world-beating Silicon Valley companies doing the disruption in more cases than not.”

It has now been almost fifteen years since Andreessen’s post. While some specific predictions have proved to be misguided¹, the general direction of travel has largely aligned with Andreessen’s vision. The major companies he highlights – Apple, Google, Facebook, Amazon, Netflix – are much more valuable and much more central to both the stock market and the average person’s daily life than they were fifteen years ago.

I want to focus on one part of Andreessen’s post: “national defense”. Andreesen leaves this area until last in his list of disrupted industries, and he is somewhat vague about what that disruption looks like. “The modern combat soldier is embedded in a web of software that provides intelligence, communications, logistics and weapons guidance,” he writes. “Software-powered drones launch airstrikes without putting human pilots at risk. Intelligence agencies do large-scale data mining with software to uncover and track potential terrorist plots.”

Andreessen does not name any companies responsible for bringing cutting-edge software to the world of military operations, but it if he had been asked, it seems likely he would have suggested Palantir.

Palantir was founded in 2003 by Peter Thiel, Alex Karp, Stephen Cohen, and Joe Lonsdale. Its purpose has largely been to provide cutting-edge surveillance and monitoring tools to intelligence agencies, police forces, and the military. Karp, the Palantir CEO, has echoed Andreessen’s thoughts on the ever-increasing primacy of software many times, and suggested in a recent video aimed at Palantir employees that we are living in a “software world.” He speaks often of a “software mindset”. I will admit that a full realisation of what Andreessen and Karp were really saying has only recently dawned on me. I had considered Andreessen’s point particularly as a kind of banal truism, obvious to everyone, and self-serving to boot. Andreessen makes his money investing in software companies (including Substack), so it makes sense that he would say the future is all about software companies and everyone ought to get on board. Karp is the CEO of a software company, so his whole business is selling people on the necessity of very expensive software. But the movement within Silicon Valley towards a new business relationship with government—the American government and others—has helped me to see the broader implications of their approach.

Building To Dominate

Palantir was founded in the aftermath of the September 11th attacks, and is an ideologically anti-terrorist organisation. I mean that its moral and political understanding of the world is one where The West is locked in battle with barbarian outsiders who want to impose a different (more primitive, more restrictive) way of life on us. For Karp and Thiel, who are the intellectual heart and public faces of Palantir, it is absolutely imperative that the forces of freedom, most notably American and European governments, have the tools they need to survive that battle, and to protect the precious and tenuous civilisation The West has built. Palantir has been designed to provide those tools. Or, as a recent advertisement for the company put it, “we build to dominate.”

Unlike most Silicon Valley start-ups, Palantir had identified their customer from the very start. Over the last twenty years, Palantir has led the way in making the Silicon Valley method of software production more acceptable and understandable to governments around the world by adapting its practices to suit the requirements of bureaucratic government procurement. The company long ago internalised Paul Graham’s famous suggestion to “do things that don’t scale”, at least until you find the things that do scale. And one very lucrative thing that doesn’t scale is manually sorting, cleaning, and presenting the data collected by the state on its own citizens, and on its military enemies. Palantir gained its foothold in the offices of government by promising to organise overwhelming floods of data, and then sold tools which could provide actionable visibility into that data. That’s the part that scales.

For instance, if you are government agency trying to find people who are supposedly in your country illegally, you might turn to Palantir to help create a live database of people you’re interested in tracking. And if you wanted to sort those people by hair colour, tattoos, visa status, or “criminal association”, you can probably make that happen. (Palantir might need to set up an ‘Ethics Education Program’ for their own employees in order to get it across the line though.) If Palantir can do this for you quickly and efficiently, you might start to consider them a “mature partner”, to whom you grant more contracts for larger and more long-lasting pieces of work. You might even sign a “strategic partnership” with them as you engage in genocidal destruction. This is what Palantir does: it uses small projects (‘sprints’) to demonstrate the capabilities of their software, which then invites their prospective customer to go ‘all-in’ on the technology they provide.

The Biggest Customer of All

The US military budget for 2025 will likely be in the region of $1tn. The biggest names in defence contracting today include Lockheed Martin, RTX (formerly known as Raytheon), Boeing, and General Dynamics. In 2023, those four companies alone received $127bn in contracts from the American government. All of those companies have been in existence, in some form or another, for a century or more. Their business, at least since the second world war, has been in fulfilling government contracts for military equipment. They are huge corporations, focused on providing the hardware that makes American war possible. The United States has been at war constantly for over 80 years, and they generally spend more than all their adversaries combined, so it’s a profitable business to be in.

It is also a huge opportunity for disruption. If a company like Palantir can claim even a fraction of the market for American defence contracts, the returns are potentially huge. But Palantir, like any company with Silicon Valley DNA, don’t want just a fraction of the market—they want as much of it as they can get, and they see the incumbents as ripe for unseating. Karp’s vision for the future of war is an extrapolation of what war is now: a long-distance engagement between technologically advanced superpowers. In that vision, tanks and troops are less important than the software you use to direct them, from surveillance all the way through to deployment. Palantir’s aim is to provide tools to The West that no-one else can match, because in their view whoever has the most effective tools will win.

As a business, however, they want to apply the zero marginal cost logic of modern software distribution to the traditionally resource-intensive environment of warfare. The opportunity here is almost global, because just about every government wants tools like those they provide. They want them for reasons of efficiency, but also for reasons of control—having spent the last twenty-five years being told that data is everything, governments of all stripes have internalised the idea that they need ever more information about their citizens. Palantir spend a long time and a lot of money developing the software they can now deploy for governments all around the world (and not just governments: corporations, banks—whoever needs intelligent data analytics services at scale), and the more contracts they sign, the better their margins get.

It has taken a long time to get to this point, but now that the doors to government are open, many companies are angling for their piece of the action. Elon Musk’s SpaceX has been a pioneer in this area, taking on much of the work that would once have been done in-house at NASA. Trae Stephens, co-founder of Anduril (a military contractor that infuses AI and mixed reality into defense tech) and a long-time associate of Peter Thiel, has gone back and forth between industry and government throughout his life, and has made a very lucrative career out of building companies specifically to receive defence contracts. These people are the tip of the iceberg, the frontier of a new generation of military contractors built for the 21st century. They look different to the old companies, but they are just as committed to war.

The Free World

On the surface, it seems ironic for arguably the most famous cabal of conservative libertarians in the world to become so focused on winning government contracts. But it does make sense when considered from the ambitious, monopolising point-of-view of Peter Thiel’s well-publicised and hugely influential ideology. The end-goal here is a total dependence, not of the company on the state, but of the state on the company. If the state is to exist at all, then it ought to rely on the most advanced technology and intelligence it can find. State actors are not set up to find and develop that technology, or the talent that can create it. Better then to leave everything to the fast-moving, well-capitalised leaders of private industry. In time, the few remaining competencies of the state will be fully hollowed out, and a brutal, opaque rationalism—as defined by a Spartan and self-selected elite—will take its place. And we will call it progress.

1. The statement that “companies like Shutterfly, Snapfish and Flickr have stepped into Kodak’s place,” now seems laughable. As does the idea that Groupon and Foursquare (both A16z portfolio companies) were ever challengers to Google’s dominance in advertising.