The prevailing instinct, as the MAGA coalition begins to visibly fracture over Iran, the economy, and the Epstein files, is to read every parallel move by the tech-defense industrial complex as a hedge — as strategic positioning against the possibility that the cult of personality cannot hold. The instinct is wrong. It flatters the electoral cycle and obscures what is actually happening. The sharper, darker reading is this: the surveillance-defense infrastructure that the current administration is accelerating does not depend on that administration for its continuation. It never has. And the actors building that infrastructure — Palantir chief among them — are not hedging. They are confident.

01The coalition is not the cult

Start with what is actually eroding. Moira Donegan’s Guardian column of April 21 makes a defensible case: Trump’s 2024 coalition is fracturing, his approval is plummeting, his working-class support is crumbling, and his fixation on culture-war grievances is a colossal misreading of voters who just want prices to come down. The polling supports the column.

G. Elliott Morris’s breakdown identifies where the bleeding is located: not the base, but the non-MAGA Republicans, the Harris-skeptics, and the voters who were simply fed up with the economy and wanted the other party in charge. Those are the people actually leaving. The cult is not.

This distinction matters. “MAGA is teetering” flattens two things that are moving in opposite directions:

• The coalition that elected Trump in 2024 — approximately four distinct groups stitched together — is dissolving at the edges.
• The cult that organized the coalition — the 30–35% base fused to the personality of the leader — is still locked in.

Any thesis that mistakes the coalition’s fracture for the cult’s collapse will mis-predict what comes next. The cult will remain even if the Republicans lose the midterms. The cult will remain even if Trump himself exits the stage. And more importantly, the infrastructure that the cult has been used to accelerate does not depend on the cult to operate.

02The manifesto is not new

On April 19, Palantir posted to X a 22-point summary of CEO Alex Karp’s book The Technological Republic, co-written with Nicholas Zamiska and published in 2025. The points were immediately labeled a technofascist blueprint by critics, and the labeling is not hysterical. The summary calls for AI-enabled U.S. military supremacy, a moral duty of tech companies to participate in defense, the necessity of hard power to replace the “soaring rhetoric” previously used to defend free and democratic societies, a denunciation of “regressive” cultures and DEI initiatives, an embrace of religion in public life, and a critique of what it calls the “psychologization of modern politics.”

Bellingcat founder Eliot Higgins flagged the point most worth holding onto: these are not abstract philosophy floating in space. They are the public ideology of a company whose revenue depends on the politics it is advocating — a company that sells operational software to defense, intelligence, immigration, and police agencies.

But note the timeline carefully. The book was published a year ago. The operating ideas have been internal to Palantir for longer than that. The manifesto is not new material. It is re-circulation. And the question the thesis has to answer is: why now, and why in public?

03Palantir does not need MAGA

Here is where the “hedge against MAGA’s collapse” reading breaks down.

Palantir’s revenue does not depend on MAGA. It depends on the state security apparatus — which is bipartisan, structural, and older than the current administration. ICE’s Palantir contracts predate Trump’s second term. The company’s government contracts have compounded across Obama, Trump I, Biden, and Trump II. Q4 2025 revenue exceeded $1.4 billion, driven largely by government contracts described by the company’s own analyst coverage as less sensitive to public social debates than consumer-facing businesses.

Institutional investors are not hedging either. In the most recent reporting period:

The insider/institutional divergence is its own story — insiders cashing out into a rising institutional bid is rarely a signal of operational distress. But the broader point stands: this is not a company positioning itself for a MAGA-less future. This is a company whose business case is indifferent to which party is in office, because the surveillance-defense consensus has been the bipartisan American consensus since the War on Terror. MAGA accelerated the infrastructure. The infrastructure was being built before MAGA existed. It will be available to whoever wins November, and whoever wins 2028.

04The quiet part, said out loud

If Palantir is not hedging, what is the manifesto doing?

The simpler, darker reading: it is a declaration of confidence. Karp no longer feels the need to disguise the operating ideology.

The 22 points are not controversial inside the defense-intelligence procurement ecosystem — inside that ecosystem, they are consensus. What is new is that they are now being said aloud, at scale, on a public platform, in a week when the broader MAGA coalition is visibly fracturing. The timing is not we need to stake a claim before the cult falls. The timing is we are confident enough that our position is secure that we can put the operating manual in public view.

That reading is worse than the “hedge” reading. A hedge implies vulnerability — a bet placed against a possible loss. Confidence implies the bet has already paid. The infrastructure is installed. The contracts are signed. The political cost of saying the ideology out loud has been calculated and judged tolerable.

05What contestation requires

The MAGA coalition will either fracture into midterm losses or absorb its current dysfunction and regroup. Either outcome is downstream of the deeper fact: the technofascist infrastructure that the current administration is accelerating is not contingent on the current administration. It was bipartisan before Trump. It will be bipartisan after him. Karp’s re-circulated manifesto makes the operating ideology of that infrastructure explicit — not because the infrastructure is vulnerable, but because it is not.

The question for anyone trying to contest this is not how do we win the next election. The question is how do we contest an infrastructure that electoral outcomes do not touch. The answer has to start with refusing to collapse the two questions into each other. The cult’s fate and the infrastructure’s fate are not the same story. Treating them as the same story is how the second story goes unwritten.

The groundwork is already installed. The operating manual has been posted in public. The question now is whether any of it can be named clearly enough, often enough, and early enough to be contested before it becomes the weather.