Karp's gospel, and what it tells us about tech capital's political theology

On the morning of 1 July 2026, Palantir's co-founder and CEO took to CNBC with a sentence that doubled as a confession and a banner. America must reach "gospel" status in the eyes of its allies, he argued, because the country had stopped believing in its own software, its own hardware, its own engineering supremacy. The framing was jarring precisely because it was unscripted: a defence-software chief, sitting in the country's flagship business channel, using theology as procurement language.

The more interesting story is not whether the United States has, in fact, fallen behind adversaries in shipbuilding or fighter production or pilot training (these are empirical questions, and the record on them is mixed). The more interesting story is the structural one. A founder of one of the most heavily contracted AI and data platforms in the Western national-security apparatus has concluded that the language of explicit political theology is the appropriate register for selling his product to the public. That is not a normal thing for a chief executive to say. It tells us something about where defence-software capital now sits in the American political imagination.

The gospel problem

Karp's rhetorical move depends on a slippage that has become routine in Silicon Valley's political commentary. He treats the technological contest with Beijing as if it were primarily a contest of belief. If only America believed in itself the way the adversary believes in itself, the argument runs, the production numbers would follow. The slippage is revealing because it concedes, almost by accident, that the production numbers alone are not currently moving the needle of public consent. The pitch is no longer "we ship more hulls." The pitch is "we need a story."

This is a familiar move in the history of industrial mobilisation. Wartime economies have always needed narrative glue. What is new is that the glue is now being mixed, on a business channel, by the chief executive of a publicly traded defence-software house whose products sort and display targeting data for Western militaries. The procurement logic and the political theology are being made in the same voice, in the same interview, with the same vocabulary.

The platform that wants a state

Palantir's commercial trajectory is the slow conversion of behavioural data into a sovereign function. The firm has, over the past decade, moved from controversial data-mining work into the heart of Western logistics, force-posture analytics, and increasingly into the operating layers of allied militaries. It is the kind of firm whose existence presupposes not just a customer but a state that buys from it in some continuous, intimate way.

A firm in that position has every incentive to want a state with strong opinions about itself. A state that believes in its own software is a state that pays to expand its own software. A state uncertain of itself shops around. Karp's theology is therefore not pious; it is also commercial. To call this a "gospel" is to do two things at once: to frame the contest as civilisational (which silences dissent on cost and method), and to lock in the political coalition that funds his firm's expansion. The defence-software sector has discovered the electoral utility of civilisarian language.

A wider pattern in tech capital

Karp is not the only voice. The past two years have brought a steady accumulation of public commentary from senior figures in AI, chips, and platforms arguing that the United States must adopt their preferred geopolitical posture more aggressively, and that internal dissent is a form of strategic naïveté. The pattern is consistent enough to deserve the name: tech capital is becoming openly partisan in its geopolitical demands, in part because its products have become load-bearing for the state's primary function (war-fighting, surveillance, sanctioning, logistics).

This is not the old defence-industry revolving door, where retired generals quietly wrote Congress. The new arrangement is louder and more public. It happens on cable, on podcasts, and increasingly on the platforms themselves (the same ones that decide who sees what about a war). When the firms that mediate the information environment and the firms that mediate targeting data use the same theological register to argue for a particular foreign policy, the boundary between commerce and state doctrine thins.

What this changes

The risks are not abstract. When a chief executive can frame defence procurement as a question of national belief, the political cost of opposing his firm's contracts rises. Dissent starts to look like heresy. Cost-benefit criticism of any particular platform gets parsed as insufficient faith. The categories of engineering (does the tool work?) and categories of state (does the state deserve its tool?) get braided together in ways that are very hard for a casual viewer to separate in real time.

The more uncomfortable corollary is that the adversaries Karp names have their own gospel economy and their own defence-software vendors, increasingly interoperable across the BRICS+ and Shanghai Cooperation orbits. Per reporting circulating on open channels this week, Moscow is engineering discounted crude structures to extend its logistical reach into client states that include New Delhi, while coordinated messaging in Sana'a and Tehran celebrates the political theology of resistance. The contest is, in fact, a contest of doctrines. The mistake is to imagine that declaring ours more loudly will by itself win it.

The thing to watch over the rest of this fiscal year is whether Karp's framing migrates from cable into procurement language, that is, whether the next defence-software contract reads less like an RFP and more like a creed. If it does, the slide from engineering to theology will be complete, and the public will not have been asked.