Money for Enemies

I do not believe we shall ever have a good money again before we take the thing out of the hands of government.

Friedrich Hayek, Denationalisation of Money, 1976

---

In May 2026, when this chapter was frozen, the Strait of Hormuz had been a war zone for ten weeks. Tanker traffic was largely blocked, ships had been seized on both sides, and a naval blockade had answered. The most strategically contested corridor on the planet was a place the ordinary machinery of settlement no longer reached. And underneath the blockade, some of the oil still moved, and some of it was still being paid for. A year earlier Reuters had reported, from four sources with direct knowledge, that Russian crude was already clearing toward Chinese and Indian refiners in bitcoin and stablecoins, and analysts tracked the same pattern growing in Iranian oil flows.¹ The buyers and sellers cannot settle safely in any currency whose issuer can revoke access, so they settle on the one that has no issuer. Iran’s April toll on tanker transit, a dollar a barrel payable in bitcoin, is the same fact from the other side: a sanctioned state, unable to collect in dollars, taking its fee in the one asset its adversaries cannot freeze.² The volume is small as a fraction of global oil, and not long ago it was zero. What it shows is not scale but availability: when the rails that normally carry a trade are closed by politics, there is now one that politics cannot close.

None of it appeared from nowhere. It is the operational consequence of an institutional shock four years earlier. The Russian central bank’s frozen reserves still total approximately three hundred billion dollars, four years after the freeze went in. No legal proceeding has returned them, and no proceeding has been seriously attempted, because the freeze is not, in legal form, a confiscation. It is an exclusion from the banking infrastructure that makes the reserves usable. The dollar instruments still exist. They still belong, on paper, to the Russian central bank. They are simply unreachable through any rail the Russian central bank can call into. The architecture that excluded them is the same architecture every other major sovereign has been using to settle international trade since 1971.

For most of the post-war period the question of how enemies settle did not require an answer. The dollar provided settlement. The banking infrastructure that cleared dollar transactions was deep and operationally neutral, in the sense that it did not ask which side of any conflict a counterparty was on. Iran sold oil. Russia sold gas. China bought Treasuries. The wires worked. The arrangement worked as long as one assumption held: that the United States, as the operator of the settlement layer, would exercise its operational authority with restraint. The neutrality was not a property of the dollar. It was a property of American forbearance. Sovereigns extended trust to that forbearance because no alternative could match the depth of dollar settlement, and because forbearance had, in fact, been the pattern. The system’s neutrality was political rather than architectural, contingent on a choice made consistently for long enough to look like a property of the system.

The 2022 freeze ended that era. The action itself had precedent; Iran, Venezuela, and Afghanistan had been treated similarly. The scale and the target were what made the implication unambiguous. Reserves held in dollar instruments are conditional on the holder’s political alignment with the United States. That had always been technically true. After 2022 it was operationally demonstrated, at scale, against a sovereign of the second rank. The neutrality was borrowed. The lender called it back. The architecture of the loan is now visible to everyone who held the note. Every central bank that is not unconditionally aligned with Washington had to look at its balance sheet and ask which line item, in which future scenario, was theirs.

---

The operational answer to that question is being assembled in real time by the actors with the most at stake. Four responses are visible in the public record. Each addresses part of the problem. None addresses all of it.

The first is repatriation. Germany pulled gold home from New York and Paris starting in 2013. The Netherlands, Austria, Poland, Hungary, India, and Turkey followed. After 2022 the trickle became a flood. The Bundesbank now holds more than half its gold in Frankfurt, where in 1990 it held almost none. Onshore physical gold restores custody sovereignty. Custody is no longer political. But settlement is. Gold sitting in Frankfurt cannot pay for Iranian oil. To move it, you need a counterparty willing to accept it, transit infrastructure that allows it to cross borders, insurance against loss, and permissions through chokepoints, every one of them revocable. Repatriated gold is a reserve asset a foreign custodian can no longer hold hostage; as a settlement asset it is no better in Frankfurt than it was in New York. The repatriation solves the wrong half of the problem.

The second is bilateral. China and Russia now clear roughly ninety-five percent of their trade in yuan and rubles; India pays for Russian oil in rubles, dirhams, and yuan in shifting proportions; Saudi Arabia takes yuan from China for some marginal volume of oil. These arrangements work for friendly counterparties and fail at exactly the moment they are most needed, when the bilateral itself becomes adversarial, because the trust they require is the trust the situation has, by construction, already removed.

The third is commodity-direct: settle in oil, grain, or metals, priced in something outside any sovereign’s accounting. Energy importers and exporters can clear this way, sometimes do, and have for centuries when the political layer broke down. But it does not scale. A semiconductor manufacturer cannot pay a wheat farmer in bushels of wheat, and a settlement layer has to be fungible across trades where physical commodities are not. It is a workaround for specific corridors.

The fourth, the one policy papers favor, is a CBDC bridge: a multilateral settlement layer run by a coalition of central banks, of which the BIS-piloted mBridge (China, Hong Kong, Thailand, the United Arab Emirates) is the visible case. The structural problem is the dollar’s, only smaller. Whichever coalition operates the bridge holds the same authority over it that the United States holds over dollar clearing, and any participant who falls out with the coalition faces the same exclusion. It is borrowed neutrality with different lenders, and the lenders are, on average, less restrained than the one being replaced.

What is left, when the four responses are stacked together, is a settlement layer that is none of three things. Transferable without physical logistics: the asset moves between counterparties who do not share a transit corridor, without requesting permission from anyone whose interest is in preventing it from arriving. Final without third-party permission: once the transaction settles, no party, including the operator of the ledger, can reverse it. Verifiable without trusting any counterparty’s records: each side confirms the transaction independently, in a form both can read and neither can alter. Gold has the first property only on paper, in clearing accounts a third party operates, and physical gold only across borders that allow it to cross; the Russian-Iranian corridor cannot move gold through Western airspace. Bank wires fail the second by construction: they can be reversed, SWIFT messages can be retracted, clearing-house sales can be unwound by court order. Every traditional rail fails the third, because the records live with parties whose cooperation is the very thing in question. The properties cannot be assembled by combining traditional instruments. Something else has to provide them.

Bitcoin is the first asset in monetary history with all three at once. Transfer is a protocol-level operation no jurisdiction can prevent from confirming. Finality is proof-of-work: six confirmations deep, a transaction is final in a sense bank settlement is not. Reversing them means rewriting the chain against the accumulated work of the global hashrate, an expenditure so large that finality is probabilistic in principle and prohibitive in practice. Not physically impossible. Economically absurd, and growing more absurd with every block. Verifiability is by construction: both parties read the same chain, and neither party operates it. The properties come from the protocol rather than from policy, and they exist whether or not any sovereign approves. A sovereign that disapproves can criminalize on-ramps, pressure exchanges, and prosecute developers in jurisdictions that allow it. The Tornado Cash and Samourai prosecutions did exactly that, and reached people. What no prosecution has reached is the settlement layer itself. The reachable surfaces are the people and the services. The protocol is not a person and is not a service. It is a specification, and the specification runs on machines that do not know whose country they are in.

The deeper consequence is not that Bitcoin provides another settlement asset. It is that Bitcoin changes the shape of risk in cross-bloc trade. Under traditional settlement, counterparty risk accumulates. Every dollar of trade between sovereigns who do not fully trust each other builds exposure that can be seized in a future political rupture. Russia’s three hundred billion is the canonical example. The principle is general. Any reserve holding, any in-transit shipment, any clearing account is a hostage to the relationship’s continuation. The longer the trade relationship runs, the more accumulated value sits in forms the counterparty’s sovereign can reach. Trade between adversaries under traditional settlement is a structure that gets more dangerous to both sides the more it succeeds.

Under Bitcoin settlement, each transaction is atomic. The settlement happens, the funds arrive, the trade is complete. There is no accumulated exposure to reach. A future political rupture costs the next trade, not the last thousand. The hostage is bounded to the transaction in flight rather than to the cumulative relationship. This is a different shape of risk than the international system has previously offered. It does not eliminate political risk. Sovereigns can still close on-ramps and pressure local conversion. What it eliminates is the accumulation of political risk over the life of a relationship. Each trade stands alone. Each is settled or not, and once settled, settled.

For trade where the parties cannot rely on each other’s banking systems, this changes the calculus of whether to trade at all. The current alternative for many such trades is no trade: the political risk of building exposure is too high to accept. Bitcoin makes the trade insurable, and its marginal value is highest exactly where existing systems work least, in the highest-friction, highest-mistrust corridors where the alternative is not a different settlement method but no transaction at all.

The standard objection to this argument focuses on whether Bitcoin can replace the dollar as a reserve currency. The framing misses what is being claimed. Reserve status is about storing value over time, and gold has that role essentially locked. Central banks have purchased roughly a thousand tonnes of gold per year for several years running. They have not purchased Bitcoin at any comparable rate. They are voting with reserves, and the vote is not for Bitcoin. Anyone arguing Bitcoin replaces gold in the reserve role is arguing against the revealed preference of the actors whose preferences settle the question. Reserve status and settlement function are separable. Bretton Woods separated them: gold was the reserve anchor, the dollar the settlement medium. The post-1971 system collapsed them into the dollar, which worked while the dollar’s neutrality was credible. A genuinely fragmented world separates them again, with different occupants. Gold for reserves, the role it has held for five thousand years and continues to hold by every measurable signal of sovereign behavior. Bitcoin for settlement between parties who cannot trust each other’s banking systems, the role nothing else can do as a property of its construction.

This is not a defeat for the architecture but the realistic shape of its function. The dollar will remain dominant, and most trade will continue to clear in dollars and the rails that support them. The point is not displacement. It is that this specific functional niche, permissionless final settlement between distrustful parties, has exactly one architecturally coherent occupant, and that occupant exists, runs continuously, and is being adopted at the margin by exactly the actors the niche describes.

The institutional consensus prices Bitcoin as a risk-on asset. BlackRock’s spot ETF, launched in January 2024 and now the largest in its class, placed it inside the standard alternatives sleeve, modeled against tech-equity baskets and judged by drawdown thresholds it will not consistently meet. Every major allocator who has come into Bitcoin in the past two years came through a rail that books it next to the NASDAQ and asks the standard portfolio questions: expected return, volatility-adjusted contribution against a comparable basket. Neither has a good answer for an asset with no cash flows, which is why the price is volatile and the institutional posture stays cautious.

The framework is asking the wrong questions because it has the wrong category. Risk assets are priced by their cash flows and their place in a portfolio. Infrastructure assets are priced by network effects and by the size of the addressable market for the function they perform. Bitcoin has no cash flows. The question is not what return it produces. The question is what fraction of the world’s settlement-between-distrustful-parties needs the architecture, and at what price the float supports that throughput. The first question has no good answer. The second has an answer measured in trillions of dollars per year of trade the existing rails will not safely carry. The day the market starts asking the second question is the day the asset is repriced.

The conditions for this settlement role to mature are paradoxically those of the slow erosion the previous part described. A rapid dollar collapse would trigger emergency capital controls, exchange shutdowns, and the criminalization of crypto exits: the system defending itself with maximum force when it feels most threatened. Stable dollar dominance would leave no opening for the alternative to develop. Neither extreme is the path the architecture needs. Slow erosion is. Each year of gradual fragmentation extends the network’s track record, matures the surrounding infrastructure, normalizes ownership across generations who did not grow up assuming dollar permanence, and builds the operational competence (custody, derivatives, clarity in friendly jurisdictions) that turns a protocol into a functional rail. None of it requires a crisis, only time, which the slow erosion provides.

The clock is therefore running in the right direction without anyone needing to predict catastrophe. The trend reversing (dollar weaponization receding, blocs reintegrating, cross-bloc trust rebuilding) would require affirmative political choices no current actor seems positioned to make; the default just keeps going. This is the inverse of the maximalist Bitcoin argument. Maximalism needs collapse to win. The settlement-finality argument needs only continuation, the trend already running continuing to run at the pace it is already running. The architecture does not have to triumph. It has to be available when the existing rails fail, which it is, every block, regardless of policy.

---

A note of honesty. I do not know how the geopolitics resolves. I do not know which sovereigns end up enemies in 2035, or which corridors of trade break under which sanctions regime. What I know is what the architecture provides. A settlement layer that does not require the parties to trust each other or any third party. That property is rare in the history of money, and it is the property the situation increasingly requires. The match between what is provided and what is required is not a forecast. It is a present-tense observation.

The oil is the proof. It clears between counterparties who have no bank between them, on a chain neither side operates, because no other rail connects them. The protocol behind that settlement has run for sixteen years without interruption. The fraction of global trade routed this way is small; in 2020 it was zero. The architecture is what is left when the rails fail.