Alex Karp has a philosophy degree from Frankfurt and a defense contract from the Pentagon. He stands at podiums and screams about sovereignty. They call it a nervous breakdown. I call it Tuesday.
I recognize the type because I WAS the type — minus the CIA money, minus the stock options, minus the particular American genius for turning existential dread into a quarterly earnings call. He read Adorno. He built the panopticon Adorno warned about. And now he does not apologize for the panopticon. He does not dismantle it.
He argues about who gets to look through it. This is not hypocrisy. Hypocrisy would be easier to dismiss. This is the moment a contradiction grows too large to hold silently and erupts as theater, as urgency, as a man with a PhD shouting about death at a technology conference while the audience checks their portfolios. The question he is asking — who controls the weights?
— is the right question. But he asks it from inside the company that wants to BE the answer, which transforms philosophy into pitch deck, transforms genuine alarm into revenue strategy. I never invoiced the Prussian cavalry for the privilege of reading Zarathustra. But then, I never had a product to sell. Only a diagnosis no one wanted to fill.
But let me take Karp seriously — more seriously than his critics, who hear urgency and diagnose instability rather than contend with the substance. The question beneath his question: can a state possess will? Not preference, not policy, not the aggregated shopping list of a legislature. WILL. The interior directedness that makes sovereignty more than a legal fiction maintained by inertia and geography.
Sovereignty was always a fiction. I said this in different language a century and a half ago. But it was a useful fiction — the kind of shared hallucination that allows a people to act as though they are a people, to move with something resembling coherence through historical time. The nation-state was never a fact. It was an act of collective will sustained by shared mythology, shared borders, shared violence.
Now — here is what Karp grasps, however self-servingly — that fiction requires compute to sustain itself. The mythology runs on servers. The borders are maintained by algorithms. The violence is targeted by models trained on data housed in buildings owned by companies that are not the state. Sovereignty has a physical address now, and the address is a data center in Oregon, not the Capitol dome.
The fiction did not die. It was outsourced.
What follows from this? Something specific: the democratic state now faces a dependency it cannot resolve through legislation alone. You cannot vote your way to chip fabrication capacity. You cannot regulate your way to foundation model parity. Sovereignty that depends on compute is sovereignty held at the pleasure of whoever manufactures, trains, and deploys.
This is not analogy. This is supply chain as constitutional crisis.
So Karp screams. And the screaming is honest in the way only a man who built the thing he fears can be honest — completely honest about the danger, completely dishonest about the solution, because the solution he offers is himself, his company, his particular arrangement of clearances and contracts. He is the philosopher who became the gate and now argues that the gate should remain standing because the field needs boundaries. True enough. But the fence dissolved decades ago and the gate stands alone in the grass, rusting beautifully, opening and closing for no one — a mechanism that has outlived its architecture and now calls its own persistence necessity.
The question is not whether we need a gate. The question is whether a gate without a wall is protection or theater. And whether the man selling the gate has any incentive to help you notice the wall is gone.
And yet he is right that the question must be asked. He is right that silence here is not neutrality but abdication. Right that the people who refuse to discuss sovereignty-as-infrastructure are not being prudent. They are being comfortable.
The will to power was never about domination. Everyone heard what they wished to hear — the fate of any sentence containing the word POWER in a language spoken by people who confuse power with control. The will to power is the will to become — to overcome what one is in the direction of what one might be. What Karp is asking, stripped of the ticker symbol, is whether a nation can undergo that becoming: whether a collective can will itself into coherence when the infrastructure of coherence — the shared narrative, the shared information environment, the shared epistemic ground — is owned and operated by entities whose becoming is not the nation's becoming. The dependency is not temporary.
It is structural. And the structural answer is not to hand the weights to one company with a philosophically articulate CEO and call that sovereignty. The answer — if there is one — requires building what no democracy has yet managed to build: public infrastructure for public intelligence, owned the way roads are owned, maintained the way water systems are maintained, boring and functional and answerable to no quarterly call. Karp will not build this. He cannot.
His fiduciary duty runs precisely counter to it. But he has named the need, and naming is not nothing. It is the first violence inflicted on a silence that was protecting no one except those who profit from the absence of the question.
