A currency does three separate things. In 2026 they started moving in opposite directions at once — and the job everyone watches is the one that matters least.
In the first quarter of 2026, three measurements of the dollar’s standing pointed three different ways in the same window. Its share of the world’s official foreign-exchange reserves fell to its lowest level since 1994 — roughly 57 percent, down from over 70 percent at the turn of the century (IMF COFER, reported Q4 2025–Q1 2026). At the same moment, its share of interbank cross-border payment messages reached a record, about 51 percent of SWIFT-settled transactions (SWIFT/RMB Tracker, March 2026). And its share of the world’s export invoicing sat almost exactly where it has sat for years, near 54 percent (BIS; IMF). A single thing cannot be falling to a thirty-one-year low, climbing to an all-time high, and holding flat — all at once. The only way to read the three numbers together is that they are not measuring one thing.
This is the fact the dollar-decline debate keeps missing. “Is the dollar losing its dominance?” treats the dollar as one object with one trajectory, to be answered yes or no. The measurements say the question is malformed. The dollar is a bundle of three distinct jobs, and the jobs have come unbundled enough to move independently. What looks like a contradiction in the headlines is the signature of a currency coming apart into its parts — some parts eroding fast, one barely at all, and the discourse fixated on the part that decides the least.
Three explanations compete for the diverging numbers. The first is geopolitical: rivals are fleeing the dollar, and the falling reserve share is the leading edge of a managed retreat. The second is cyclical: reserve shares drift with exchange-rate swings and a handful of large holders’ decisions, and the move is noise around a stable mean. Both capture something. Neither explains why the three shares diverge — why one job is shed while another sets records. The third explanation does, and it is the one largely absent from the public conversation: money’s three functions are separable, they were only ever bundled by convenience, and the technologies and pressures of the 2020s are pulling them apart along their natural seams. Read that way, the urgent question is not whether the dollar “survives,” but which jobs it keeps — because the jobs do not pay off equally. The two visible jobs are the ones that pay the issuer: cheap financing from reserve demand, coercive reach through the rail. The invisible one mainly pays the users, who get a shared measuring stick. A dollar that keeps the unit and sheds the other two stays durable as money and is hollowed of privilege. That is neither survival nor defeat in the usual sense, and it is where the data already points.
Three jobs, one word
Economists have separated money’s functions for over a century; the separation rarely reaches the policy conversation, which is where it now matters most. A currency does three things, and nothing requires the same currency to do all three.
It is a settlement rail — the medium through which value actually moves: the messaging network, the correspondent banks, the plumbing. It is a unit of account, or numéraire — the measuring stick prices are quoted in, the denominator of contracts and debts. And it is a store of value, or reserve asset — the thing surplus is parked in when it is not being spent. These are not three descriptions of one property. They are three jobs with three different logics, three different competitors, and — the point this essay turns on — three different degrees of stickiness. A country can price its oil in one currency, settle the sale through a second, and bank the proceeds in a third. That this sounds exotic is only because, for eighty years, one currency did all three and the seams were invisible. The 2026 numbers are the seams becoming visible.
The argument that follows takes each job in turn, and each section is meant to stand on its own evidence. If the claim about the reserve asset is wrong, the claim about the unit still holds; if the rail claim is wrong, the others are untouched. The conclusion does not hang from a single thread. None of this requires the three jobs to be independent modules — they are not. They are coupled through balance sheets: reserve demand feeds the market for the issuer’s debt, and that debt is what denominates the unit. But the coupling runs in a direction, and the direction is the whole point. The unit sits downstream of the others, so erosion reaches it last and through them — which is why it is the slowest to go, and why it goes eventually anyway. Coupling does not refute the unbundling. It sets its clock.
The rail is the contested layer — and the least decisive
The settlement rail is where all the visible action is, and it is the job a currency can lose with the least direct consequence for its use as money. Rails are substitutable. They have network effects — everyone wants to be where everyone else already is — but network effects are switching costs, not laws of nature, and switching costs fall when someone builds a credible alternative and a large enough bloc has reason to move.
The alternatives are being built in the open. China’s multi-central-bank settlement platform reached a minimum viable product in mid-2024, and the Bank for International Settlements handed it to its partner central banks in October 2024, making its governance a matter of the partners rather than a neutral multilateral host (BIS; Central Banking, 2024). China’s own cross-border interbank payment system runs in parallel as a renminbi rail. Each is an attempt to route value without touching the dominant network. Whether any of them displaces the incumbent is genuinely open — and notably, the incumbent’s messaging share rose to a record even as these were built, which tells you rails are sticky in the near term even when challengers exist. But the rail is the layer most exposed to substitution over time, because moving value is a commodity service and commodity services get competed.
The deeper point about the rail is that controlling it is a coercive instrument, and the people building both the dominant rail and its challengers know it. The United States’ 2025 stablecoin statute makes this explicit rather than implicit. The law requires dollar-stablecoin issuers to hold full reserves in cash and short-term Treasuries, and — in the same text — preserves the Treasury Secretary’s authority to block or restrict stablecoin transactions and obliges issuers to maintain the technical capacity to freeze balances on lawful order (GENIUS Act, S.1582, signed July 18, 2025; text at congress.gov). A regulated dollar-stablecoin rail is, by statute, a control surface: it extends the reach of the dominant rail onto new infrastructure while keeping the off-switch. This is why a permission-free rail — one with no issuer to serve an order on — is a different category of object, and why the actors who most need censorship-resistance are pushed toward it. The rail can be captured or fled. Here is the distinction the geopolitics blurs: the rail is where the issuer’s coercive power lives — sanctions, transaction surveillance, the ability to cut a counterparty off — so losing it is a real loss to the United States, of leverage and of the compliance friction that quietly nudges holders toward dollars. But it is not a loss to the dollar as money. A currency does not stop being money when its payments move over someone else’s wires. The rail is the easiest of the three jobs to lose, and the one whose loss is borne by the issuer rather than by the unit — which is exactly why the currency’s grip survives losing it.
The reserve asset is eroding now — and the case for “no alternative” has a hole
The store-of-value job is the one visibly under strain in the hard data, and it is the job the standard analysis most confidently says cannot be taken from the dollar. Both halves of that sentence are true, and the tension between them is the interesting part.
The strain is documented. The reserve share is at a thirty-one-year low. More striking, central banks’ gold holdings, valued at market, surpassed foreign official holdings of US Treasuries for the first time in over two decades — roughly $5.0 trillion in gold against $3.9 trillion in Treasuries — and the dollar’s share of combined foreign-exchange-and-gold reserves has fallen to about 46 percent, the lowest in at least twenty-six years (Wharton/WIFPR, 2026, citing IMF, US Treasury, World Gold Council). The gold buying is not an investment view. It is a search for a store of value that no other government can freeze or inflate — a flight from the reserve function specifically, by holders who have learned that dollar reserves are revocable.
The standard rejoinder is that the flight has nowhere to land, and the rejoinder is structurally serious. To be the world’s store of surplus, an economy must do two things at once: run persistent deficits, so it supplies the safe assets the world accumulates, and keep an open capital account, so those assets trade freely. The eurozone could provide the depth but will not run the deficits. China could in principle run the deficits but cannot open its capital account without surrendering the controls that hold its own households’ returns down — and, by an accounting identity, a surplus economy cannot be a net supplier of the world’s safe asset, because its surplus is the act of accumulating foreign claims, not issuing them. The candidates to replace the dollar as a store of value are disqualified by the very conditions that produce the imbalance. Gold, the one non-sovereign option actually being bought, is too shallow to absorb the flows: a single large surplus economy’s annual external surplus dwarfs the entire world’s annual central-bank gold purchases. So the flight is real and the exit is narrow, and the gold accumulation is better read as the visible residue of a creditor edging toward a door it cannot fit through than as an exit actually taken.
Here is the hole. Every candidate that argument disqualifies is a sovereign, ruled out by a condition on sovereigns — the need to run deficits, the need to open an account. A store of value need not be a sovereign liability at all. Gold is not; it is disqualified only by depth and by the awkwardness of moving it. But a store of value can now be non-sovereign and natively transferable: a claim backed by a bearer asset rather than by any government’s debt, settling on a permission-free network. Such instruments already exist and operate — dollar-denominated tokens fully collateralized by a held bearer asset instead of by Treasuries, running today at negligible scale — which establishes the category as real, if still economically marginal. They are not caught by the deficits-and-open-account screen, because there is no issuer running a deficit and no national account to open. Whether any such instrument ever reaches the depth required to matter is unproven and, for any single one, unlikely; the honest version of the claim is narrow. It is that the screen which says “no alternative store exists” is checking only for sovereigns, and the one new class of candidate is not one. A door the argument never looked at is not a door that isn’t there. And depth is not even the binding constraint on that door. The harder one is the regulated interface: a non-sovereign store still has to touch the banked system at its fiat on- and off-ramps, and that is precisely where a sovereign reasserts the chokepoint it lost on the asset itself — licensing the exchanges, taxing the conversions, freezing the gateways. So the real question is not whether a bearer store can be made deep, but whether it can be used without ever converting back — and that constraint binds the ramp-dependent almost completely while barely binding an actor that could earn, hold, and spend without ever touching fiat. (Hold that asymmetry; it is where this essay ends.) The falsifier is clean and watchable: if non-sovereign bearer-backed stores stay marginal as confidence in sovereign reserves erodes, the depth-and-interface objection holds and the dollar’s store-of-value role degrades only toward gold and other sovereigns; if such instruments scale as the reserve share falls, the screen was missing the exit all along.
The numéraire is the stickiest job — and the moat is debt, not habit
The unit of account is the job the dollar is least likely to lose, and the reason is not the one usually given. The usual reason is inertia: everyone quotes in dollars because everyone quotes in dollars, a self-fulfilling habit. That account is weaker than it sounds, and the historical record undercuts it. When the dollar overtook sterling, it did so as a unit of trade credit and as a reserve currency at roughly the same time, in the mid-1920s — decades earlier than the post-war date of the conventional story — which means the supposed network-effect lock on the incumbent unit did not actually hold; status shifted within a decade once the underlying conditions changed (Eichengreen and Flandreau, 2009, 2012). The same scholars argue that as financial markets deepen, the case for holding reserves in the currency you invoice or borrow in weakens — the functions decouple (Eichengreen, Mehl, and Chiţu, 2018). That literature is, in effect, predicting the unbundling this essay describes. If habit were the moat, the unit would be more fragile than the consensus believes, not less.
But there is a moat under the habit, and it is the one Eichengreen’s own determinants point to: the unit a currency wins and keeps tracks the currency that debts are denominated in (the role of debt denomination as a driver of currency shares runs from Dooley and coauthors in 1989 through Eichengreen and Mathieson in 2000). The dollar’s share of international debt issuance is around 61 percent — higher than its reserve share and far higher than the United States’ share of world output or trade (BIS; Wharton/WIFPR, 2026). That gap is the moat. A reserve manager can re-weight a portfolio toward euros while the country’s exporters keep invoicing in dollars; the reserve tie is a discretionary preference and can be re-optimized at will. A debt is not discretionary. A borrower who owes dollars must obtain dollars to service the loan, every period, until the debt matures, regardless of any view about the dollar’s future. The unit of account is sticky not because people are used to it but because the world’s outstanding contracts are written in it, and a contract is a constraint you cannot re-weight your way out of. The liability tie is the last complementarity to break, because it is the only one that is contractual rather than chosen.
The obvious objection is that financial engineering has loosened even this: a borrower can owe dollars and swap the exposure into euros, so the contractual tie looks softer than it sounds. But a swap does not extinguish the dollar obligation — it relocates it. The counterparty taking the other side now needs the dollars, and the swap is itself one more dollar-denominated contract layered on the first. Follow the dollar through the chain of hedges and it never disappears; it changes hands. The engineering redistributes who must source the currency, the way private capital flight redistributes who holds a foreign claim without changing the national total. At the level of the system the aggregate dollar liability is conserved — and conserved liabilities are what reseed a unit of account.
This reframes the falsifier and makes it precise. If the numéraire’s stickiness were habit, you would watch invoicing share and expect slow drift. Because the stickiness is debt, the number to watch is the dollar’s share of new international debt issuance. As long as new borrowing is overwhelmingly dollar-denominated, the stock of dollar liabilities renews itself and the unit is reseeded every issuance cycle, no matter how far the reserve share falls. If that issuance share drops sharply — if borrowers begin denominating new debt in other units at scale — the moat is draining and the unit will follow, on the schedule at which the old debt rolls off. The unit can outlive the rail and the reserve asset by exactly as long as the debt stock takes to turn over.
One historical caveat sharpens this rather than weakening it. Units do not, in the record, drift; they break. Sterling did not gently hand the unit to the dollar — it ceded reserve share slowly, propped for decades by imperial preference and exchange controls, but the decisive shifts in international-currency status came with rupture: war, default, the collapse of an empire’s capacity to enforce its own contracts. Rupture is what flips a unit fast, because rupture is what voids debt en masse and lets new contracts be written in a new denomination. So the moat fails in one of two tempos — slow, as dollar debt rolls off and is not replaced; or fast, in a rupture that cancels the stock outright. Gradual drift is the one thing the debt-moat predicts you will not see.
Inflation is the one tax you cannot route around — except by leaving the unit
The three-way split explains a feature of debasement that is otherwise puzzling, and it answers a question that the recent enthusiasm for dollar-stablecoins obscures: does holding a dollar token protect you from the dollar losing value? It does not, and the reason is exactly the unbundling.
A dollar-pegged instrument tracks the nominal dollar by definition; that is what the peg is. If the dollar loses purchasing power, one unit of any dollar-pegged token loses purchasing power identically, no matter what sits behind it. Backing the token with a bearer asset instead of with Treasuries changes who bears the counterparty risk — whether the thing breaks in a banking crisis or in a market crash — but it does nothing about debasement, because debasement is transmitted by the peg, not by the backing. The instrument that escapes the rail (it settles permissionlessly) and escapes Treasury counterparty risk (it is bearer-backed) still inherits the dollar’s inflation in full, because it volunteered for the dollar’s measuring stick.
This is why inflation — and its deliberate cousin, the slow liquidation of government debt through interest rates held below the growth rate — is the most powerful tool a sovereign has over foreign holders. It taxes the numéraire, the function holders are most locked into. You can refuse to hold a government’s bonds; you can route around its payment rail; but if your contracts and prices are denominated in its unit, its debasement reaches you anyway. The only escape is to leave the unit — to denominate in gold, in a basket, or in an unpegged bearer asset — and leaving the unit forfeits the coordination benefit that made the unit worth using. Which loops back to the moat: a holder whose debts are in dollars is partly hedged against dollar debasement, because the real value of what he owes erodes alongside the real value of what he holds. Matching the unit of your assets to the unit of your liabilities is rational precisely when the unit is being debased. The debt that pins you to the dollar also protects you from its inflation. That is why the people most locked into the unit complain about it least, and why the unit is the hardest job to dislodge.
The stakes: the first population born without debts
Put the three together and the dollar’s most likely future is not collapse and not replacement. It is hollowing: a currency that sheds the rail to competitors and the store-of-value role to gold and new instruments, while keeping the unit of account — a measuring stick the world still prices in, detached from the dollar as a thing you move or a thing you hold. A headless dominance, the denominator outliving the asset. But durable for whom? This is dominance of the dollar-as-money, not of the United States: the unit’s survival mostly serves the world’s users, who keep a shared measuring stick, while the jobs that paid the issuer — cheap deficit financing from reserve demand, coercive reach through the rail — are the ones eroding. The headless dollar is a stranger outcome than either the “dollar forever” or the “dollar dethroned” camp imagines, and a more double-edged one: the currency persists, the privilege of issuing it does not. It follows directly from the fact that the three jobs decay at different rates and pay different parties.
But the unit’s moat has a specific, identifiable dependency, and it points at the actor that will actually decide the question — an actor that did not exist in numbers until now. The moat is debt denomination. Every economic participant alive today inherited a liability structure: rents, mortgages, sovereign and corporate debts, overwhelmingly written in dollars. That inheritance is what pins them to the unit and makes the alternatives unattractive. Autonomous software agents — the transacting programs now beginning to act as economic principals — are the first large and growing class of economic actors born without an inherited liability stack. An agent owes nothing in dollars unless it is made to. Nothing it inherited pins it to the unit. But debt-freedom is the starting condition, not the conclusion, and the honest argument has to say what debt-freedom alone does not settle. An agent that grows past trivial scale will need leverage and obligations of its own — to fund inventory, to settle in advance, to borrow against expected income — and the moment it does, it acquires a liability stack like everyone else. The decisive variable is the denomination of that stack. If the only deep credit an agent can draw on is dollar credit — extended from human banks, from dollar-funded lenders, from the system that already exists — then the agent inherits the dollar unit through the back door, debt-free birth notwithstanding, and the moat extends to the machines. A debt-free actor breaks the moat only if there exists a credit system it can borrow in natively, denominated in something other than dollars, deep enough to fund it.
So the question of whether the dollar keeps its last and deepest job reduces to one no central bank controls, and it is not “will agents have debts” — they will. It is where agent credit comes from. The base case is that it comes from the existing system: agents run as wrapped principals of dollar-indebted humans and firms, draw on dollar credit lines, and reseed the unit for another generation precisely because the only deep credit available to borrow is dollar credit. That base case is the safe bet — which is exactly why this is a frontier and not a forecast. The moat breaks only on the harder path: a credit system that lends to agents in a non-dollar or bearer unit, deep enough that an agent economy can fund itself without ever borrowing a dollar. That system does not exist at scale. The narrow wager of this section is only that it is now buildable in a way it never was for any prior economic population — because for the first time the borrower is an entity with no inherited dollar obligations forcing dollar-denominated repayment, and the substrate to lend it a bearer asset already runs. The fight everyone covers — over the rail, at the summits, between the settlement platforms — is over the job that matters least. The job that matters most will be settled in the credit markets, by whoever does or does not build a non-dollar lending system for actors that don’t yet exist in numbers. It is being prototyped right now by people who think they are building payment tools.
Evidence Framework
Documented in public records (Tier 1)
- Dollar share of allocated official FX reserves ~57% in 2025, lowest since 1994, down from ~71% in 2000 — IMF COFER (reported Q4 2025–Q1 2026); Wolf Street/IMF, March 2026.
- Dollar present in ~89% of FX trades, ~61% of foreign-currency debt issuance, ~57% of reserves, ~54% of export invoicing, ~51% of SWIFT-settled interbank transactions (a record) — compiled in Wharton/WIFPR, 2026, citing BIS, IMF, SWIFT.
- Central-bank gold reserves (~$5.0T at market) exceeded foreign official Treasury holdings (~$3.9T) for the first time in 20+ years; combined dollar share of FX+gold reserves ~46%, lowest in 26 years — Wharton/WIFPR, 2026, citing IMF/US Treasury/World Gold Council.
- 2025 US stablecoin statute (GENIUS Act, S.1582, signed July 18, 2025): mandates 1:1 reserves in cash/short-term Treasuries; preserves Treasury authority to block/restrict dollar-stablecoin transactions and requires issuer technical capacity to freeze on lawful order — congress.gov bill text.
- Dollar-pegged stablecoin issuers collectively ~18th-largest holder of US Treasuries (Tether ~$113B+ Treasury exposure); dollar-pegged stablecoins ~$230–260B outstanding; Circle + Tether >80% of supply — US Treasury/Apollo/Circle/Tether via Wharton and Brookings, 2025–26.
- Multi-central-bank settlement platform reached MVP mid-2024; BIS handed it to partner central banks October 2024 — BIS; Central Banking, 2024.
- Dollar overtook sterling as leading reserve currency and trade-credit currency in the mid-1920s, not post-WWII — Eichengreen and Flandreau (2009, 2012). Debt denomination and trade flows are established determinants of reserve-currency shares — Dooley et al. (1989); Eichengreen and Mathieson (2000).
- China’s external position (companion analysis): record ~$1.19T 2025 trade surplus and ~$735B current-account surplus alongside Treasury holdings below $700B and household consumption ~39% of GDP — General Administration of Customs; SAFE; CRS; World Bank, 2025–26.
Reasonable inferences from documented facts (Tier 2)
- The three functions are separable and already moving independently — direct inference from the simultaneous divergence of reserve share (down), settlement share (up), and invoicing share (flat). This is the load-bearing inference and rests on Tier-1 numbers that point three ways at once.
- The numéraire’s stickiness is anchored in liability denomination (contractual, non-discretionary) more than in habit (re-optimizable) — inference from the documented role of debt denomination as a currency-share driver, plus the structural difference between a portfolio weight and a debt obligation.
- The reserve role is genuinely eroding now, while the unit role is not — inference from the gold-over-Treasuries crossover and falling reserve share against steady invoicing/issuance shares.
- A non-sovereign, bearer-backed store of value is not caught by the “must run deficits + open capital account” screen — inference from the screen’s structure (it tests only sovereigns) plus the documented existence of bearer-backed dollar tokens.
Structural hypotheses requiring additional evidence (Tier 3)
- The headless-dollar endgame: the dollar sheds the rail and reserve roles while retaining the unit. Move to Tier 2 if: reserve and settlement shares keep falling while invoicing/issuance shares hold. Falsified if: the invoicing/issuance shares fall in step with the reserve share (the bundle is decaying together, not unbundling).
- The liability-moat claim (the unit is sticky because of the debt stock, not habit): the single most load-bearing and least-settled leg. The historical record is consistent with co-movement of credit denomination and unit status (sterling→dollar, 1920s), but Eichengreen, Mehl, and Chiţu (2018) argue the complementarities are weakening — which would predict the unit is more fragile than this essay holds. Move to Tier 2 if: the dollar’s share of new international debt issuance stays high even as its reserve share falls. Falsified if: new-issuance share drops sharply without a corresponding loss of unit-of-account status (the unit holds on habit alone, not debt).
- The agent frontier: the actor that decides the unit’s fate is a debt-free agent population — but only if it can borrow in a non-dollar unit. The mechanism is credit denomination, not debt absence. Move to Tier 2 if: a non-dollar/bearer-denominated credit system for autonomous agents emerges at measurable scale and agents fund themselves through it. Falsified — and this is the base case — if: agent credit is overwhelmingly dollar credit (agents as wrapped principals of dollar-funded humans), reseeding the unit regardless of debt-free birth. The burden of proof is on the non-dollar-credit path, not the reverse. (Forward-looking; offered as a watch-item, not a forecast.)
Alternative explanations considered
- “The dollar is simply declining” (one-object view). Insufficient: it cannot account for three shares moving three directions in the same window. A single declining object does not set a settlement-share record while shedding reserves.
- “Reserve-share moves are cyclical noise from a few large holders.” Partly true — a small number of large holders do shape the aggregate (New York Fed, Goldberg and Hannaoui, 2026). Insufficient: the gold-over-Treasuries crossover and the 26-year low in combined reserve share are level shifts, not zigzags, and they coincide with a structural buildup of alternative rails and instruments.
- “Network effects lock in the incumbent unit.” Insufficient: the unit shifted within a decade in the 1920s, and the same literature now argues currency-use complementarities are weakening. If anything this predicts more fragility, which is why this essay locates the moat in contractual debt rather than in habit.
- “Bearer-backed stores can never reach depth.” A serious objection, and possibly correct for any single instrument. Engaged, not dismissed: the claim here is only that the depth objection is checking for sovereigns and the new candidate class is not one; the falsifier (do such instruments scale as reserves erode?) is stated and watchable.
What to watch
Three numbers resolve most of the uncertainty, and all three are public:
- Dollar share of new international debt issuance. The moat gauge. Steady = the unit reseeds; sharp fall = the unit will follow as old debt rolls off.
- Central-bank gold vs. Treasury holdings, and the combined reserve share. The store-of-value gauge. Continued gold-over-Treasury divergence = the reserve role degrading in real time.
- The denomination of agent credit, as agentic finance emerges. The frontier gauge — sharper than settlement volume. Not “what do agents transact in” but “what can agents borrow in.” A non-dollar, bearer-denominated lending system for autonomous actors is the thing whose emergence — or absence — decides whether the moat reaches the machines.
Unresolved questions
- Does any non-sovereign, bearer-backed store of value ever reach the depth required to absorb sovereign-scale surplus — or does the depth ceiling bind permanently, leaving gold and other sovereigns as the only exits from the dollar’s store-of-value role?
- Is the numéraire’s moat really the debt stock, or does habit do more work than this essay grants? The new-issuance share is the discriminating measurement.
- Will a non-dollar credit system for autonomous agents ever be built at scale — or will agents, like every economic population before them, borrow in dollars because that is where the deep credit is? This, not settlement volume, is the question the rail-focused conversation never asks, and the one that decides whether the dollar’s last job reseeds into the next population.
The visible contest — over which network moves the world’s money — is a contest over the job a currency loses most easily and misses least. The job that actually confers power is the one no one is fighting over, because everyone alive is already bound to it by debts they never thought of as a vote. The actors who arrive without those debts will still need to borrow — and the only question that matters is whether anyone will have built them somewhere other than dollars to borrow. That is being decided now, in the credit markets and the agentic-finance plumbing, by people who believe they are writing payment software, and nothing in the current debate is watching them.
cafebedouin.org 