The grammar of deterrence

Posted on: 21 December 2025

749 billion dollars. The first human being in history to cross the 700 billion threshold. On Friday, the Delaware Supreme Court reinstated a compensation package that a lower court had twice voided, calling it "unfathomable".

The standard commentary oscillates between two predictable poles: moral indignation at the obscenity of wealth concentration, or uncritical celebration of entrepreneurial genius. Both reactions miss what is structurally interesting.

The mechanism worth observing is not how much Musk owns. It is how he operates.

The unused bomb

There is a logic that military strategists understand well: the nuclear weapon is more powerful when it remains unused. Actual deployment exhausts its power. It destroys, certainly, but then it is finished. Demonstrable capability, by contrast, multiplies indefinitely. It produces effects without being consumed.

Musk appears to have internalised this grammar and applied it systematically.

He did not buy Twitter to control public discourse. He bought it to demonstrate that he could control it if he wished. The acquisition itself was the message. He does not enter the American governmental orbit to actually govern. He enters it to make visible, to anyone paying attention, that the boundaries between private and public spheres are permeable to those who accumulate sufficient leverage.

This is the structural difference between conquest and deterrence. Between exercising power and accumulating optionality.

The absence of antibodies

Democratic governance systems know how to defend against those who act. They have developed robust antibodies through centuries of conflict: regulation, antitrust legislation, parliamentary inquiries, media pressure, public opinion mobilisation. These are mechanisms designed to react to concrete actions.

But they have no antibodies against those who could act and systematically choose not to.

What exactly do you react against? A hypothetical intention? An unexpressed capability? Someone who merely demonstrates, repeatedly, that they could do something without ever actually doing it?

Here lies the systemic fragility this episode makes visible.

Our frameworks for controlling power were constructed on the assumption that power must be exercised to be threatening. They do not contemplate actors who accumulate what we might call pure optionality: the capacity to act without the obligation to act. Every option kept open represents asymmetric convexity. You retain all potential upside without paying the costs of execution.

The shareholder vote

There is a procedural detail in this affair that merits attention.

Tesla shareholders voted in favour of this compensation package twice. In 2018 when it was proposed, and again in 2024 after it had been annulled. Chancellor Kathaleen McCormick of the Court of Chancery decided she must protect them from themselves. She ruled that even though they had voted knowingly, the process was flawed, the package was unfair, their choice could not stand.

The Delaware Supreme Court overturned this, calling the rescission "improper and inequitable" to Musk.

The American legal system is attempting to apply conceptual categories designed for an era when wealth was fundamentally passive. You owned assets, extracted rents, the relationship between ownership and control was relatively linear. Mechanisms for protecting minority shareholders assumed that a common interest could be objectively defined.

But when shareholders vote as members of a community of believers rather than as investors maximising returns, which categories apply? And who decides what is in their "true" interest when they themselves have expressed a clear preference?

The nature of new wealth

The traditional wealth of the great magnates was essentially static. You controlled railways, steel, oil. Physical assets with predictable returns. The power derived from them was significant but circumscribed.

This wealth is qualitatively different. It controls global communication infrastructure through X. It controls access to space through SpaceX, with approximately 42 per cent of a company valued at 800 billion. It controls a significant share of global electric transport. Potentially, through xAI, it could control artificial general intelligence capabilities.

It is not the quantity of money that changes the equation. It is the concentration of operational capabilities in a single decision node. It is the difference between owning and being able to do.

The Rockefeller precedent

John D. Rockefeller was not "challenging" antitrust legislation. Antitrust legislation did not exist. The Sherman Act of 1890 had to be invented after someone had empirically demonstrated, simply by operating at the maximum of permitted possibilities, what happens when there are no limits.

Rockefeller was not a reformer seeking to change the rules. He was a rational operator exploiting available spaces. The system had to evolve new antibodies in response.

Perhaps we are observing something structurally analogous.

Not necessarily a cinematic villain seeking world domination. That narrative frame is reassuring because it allows us to identify an enemy and mobilise against them. Reality is probably more complex and less dramatic.

What we may have before us is an industrial scale stress test of systems that were never designed to withstand this level of concentration of operational capability. A test that reveals where the fragility points lie, which implicit assumptions no longer hold, which control mechanisms have become obsolete.

The question that remains

The interesting question is not whether Musk is good or bad. That is a question for talk shows, useful for generating engagement but sterile for understanding what is happening.

The structurally relevant question is: which new governance mechanisms will need to be invented after someone has demonstrated, simply by existing and operating, where the gaps in the current system lie?

And who will design them? With what legitimacy? Over what timeframe?

Because if history teaches anything, it is that systems do not evolve through foresight. They evolve through necessity, after someone has empirically demonstrated the limits of those that exist.