The Risk-Free Stack
Why full faith and credit is becoming a portfolio problem, and what it now means to own safety in a fragmenting world
By Shanaka Anslem Perera
June 3, 2026
For seventy years, the dollar and the Treasury were a single trade.
To own one was to own the other. A central bank that wanted the safety of the United States bought American debt. In buying the debt it held the currency. And the two privileges, the world’s money and the world’s safe asset, arrived welded together in one instrument at one yield.
That weld is breaking.
Not in a crisis headline. Not in a default. Not in a theatrical moment where one asset replaces another. It is breaking in the only place these things announce themselves early, the price the world will pay for safety above and beyond return.
The premium to hold dollars remains powerful. The premium to hold long-dated American debt has weakened severely, and at the long end has turned negative.
That is not a forecast. It is the finding of a December 2025 paper by Wenxin Du, Ritt Keerati, and Jesse Schreger, issued as Federal Reserve International Finance Discussion Paper 1427 and National Bureau of Economic Research Working Paper 35000. Separating the convenience of the currency from the convenience of the bond through deviations from covered interest parity, they document a decoupling. The dollar’s convenience remains strong. Treasury convenience has declined substantially and turned negative, most strongly at the medium and long maturities. The world still pays to transact in dollars. It is no longer willing to store decades of value in American duration on the old terms.
The implication is large, and the frame that maps it is not the one most allocators carry.
The risk-free asset was never one thing. It was a bundle of services that one instrument happened to deliver at once. That bundle is now splitting into a stack of layers, each supplied by a different instrument, each repricing on its own clock, each carrying its own custody, jurisdiction, settlement, and political risk.
The question that once had one answer, what is the risk-free asset, now has several, and they have stopped agreeing.
One boundary, drawn before anything else, because the entire argument depends on it.
That geopolitical risk is priced is not news. The International Monetary Fund priced it in the April 2025 Global Financial Stability Report, using news-based geopolitical-risk measures, sanctions variables, sovereign credit spreads, and asset returns to show that major shocks, military conflict above all, carry a persistent and measurable premium, and computing what it called a geopolitical-risk beta. The European Central Bank and the European Systemic Risk Board have built a fragmentation-monitoring framework. The academic literature already treats sensitivity to geopolitical risk as priced.
None of that is mine.
Anyone claiming to have discovered that politics moves prices is uninformed or selling.
The contribution here is not a new factor. It is the decomposition. Seeing the bundle come apart. Naming the layers that are separating. And pressing one question the existing work has not yet settled: whether the next useful measure of safety is structural rather than reactive.
Integration, not revelation.
Where the frame is proven, the article says so. Where it becomes a research program, it says so without flinching. For the reader with the IMF report open in another window, that line is what earns the rest.
The bundle, then.
And how it breaks.
What Safety Used to Mean
For most of the postwar era, one instrument, the marketable debt of the United States Treasury, performed five jobs for the global system so completely that almost no one thought to separate them.
It was the reserve, the asset a central bank held to anchor its currency and store national wealth in the world’s unit of account.
It was the collateral, the universal pledge of the repo market and an eligible Level 1 liquid asset in bank liquidity rules.
It was the cash, the Treasury bill where corporations, funds, money-market vehicles, and dollar institutions parked operating liquidity.
It was the store of value, the long bond pensions, insurers, and sovereign funds bought to hold safety across decades.
And beneath all four, it was the settlement, the asset whose ownership transferred through a clearing system the world assumed would always remain open.
The genius of the arrangement, the thing a French finance minister once called America’s exorbitant privilege, was that the five came bundled at one price. You did not choose. You bought the Treasury and received reserve status, collateral utility, cash equivalence, duration, and settlement finality in one trade, from one issuer, at one yield.
The convenience yield measured what the arrangement was worth.
For decades it was worth enormously.
The bundle is now unbundling. The services are separating. Different instruments are taking different layers. And the instrument that once supplied all five is visibly strong in some and visibly weakening in others, which is exactly the signature of a bundled good decomposing and repricing piece by piece.
Take the layers one at a time.
The Long End Lost Its Exemption
Start where the pressure is undeniable and the cause cannot be argued away, the long-duration store of value.
The cause is fiscal arithmetic, and the arithmetic is not subtle.
The Congressional Budget Office, in its Budget and Economic Outlook covering 2026 to 2036, projects a federal deficit of 1.9 trillion dollars this fiscal year, equal to 5.8 percent of output, widening to 3.1 trillion, or 6.7 percent, by 2036. Debt held by the public climbs from 101 percent of GDP to 120 percent, above the 106 percent record set after the Second World War.
The deficit is also changing composition in the worst possible direction.
Net interest, the cost of carrying the debt, rises from 3.3 percent of output in 2026 to 4.6 percent in 2036, at which point it consumes nearly one-fifth of all federal spending. The Office states that net interest exceeds 3.2 percent of GDP in every year of the projection, its highest recorded level since at least 1940.
That baseline itself now carries legal uncertainty, because the Supreme Court’s February 2026 ruling striking down the administration’s emergency tariffs put a material part of expected tariff revenue into dispute. The fiscal point does not depend on the exact tariff adjustment. The debt-service channel is already doing the work.
This is the engine.
A government running structural deficits while its interest bill compounds must sell ever more debt into a world whose appetite for that one instrument is finite. The rating agencies have logged it. Moody’s cut the United States to Aa1 from Aaa in May 2025, leaving all three major agencies without the United States at the top rating, citing rising debt, persistent deficits, and growing interest costs. S&P Global Ratings moved in 2011, Fitch in 2023, and a fourth agency, Scope, cut in October 2025.
Du, Keerati, and Schreger supply the wire from arithmetic to price. The erosion of Treasury convenience is a supply story. The relative flood of American paper against other developed sovereigns is draining the premium. The dollar’s convenience remains strong, but the bond’s convenience has weakened and turned negative at the medium and long maturities. The mirror image sits in Europe, where the ECB’s June 2026 review finds the German Bund’s convenience yield rising as scarcity is rewarded in a world starved of highly rated euro safety.
That distinction matters.
The world is not abandoning the dollar. It is repricing the duration risk of the sovereign that issues it.
That is the fracture most portfolios are not built to see. Dollar scarcity can coexist with Treasury duration indigestion.
So the first layer separates. The long end is becoming an ordinary risk asset, priced by supply, term premium, and fiscal credibility rather than by an unquestioned flight to safety. This is not a default. It is a repricing.
The thing that once defined safety across decades is, at the long maturities, no longer treated as safe in the old way.
The Front End Got a Captive Bid
Now a layer moving the other way, where the simple story that the world is losing faith in American debt breaks down.
While the long end weakens, the short end is being reinforced.
The source is the least likely corner of modern finance: dollar-backed stablecoins.
Privately issued tokens promising redemption at par have become a structural buyer of Treasury bills. In July 2025 Congress wrote that demand into statute.
The GENIUS Act became law on July 18, 2025. It restricts issuance to permitted issuers, requires reserves of at least one to one in cash and specified liquid assets, including short-dated government obligations and qualifying repo, tightly restricts the reuse of those reserves, and mandates monthly public disclosure of reserve composition.
The result is not a crypto escape from the state. It is a regulated front-end Treasury distribution channel.
Every dollar into a compliant payment stablecoin is pushed toward a narrow collateral universe. The front end gets an engineered buyer. The long end does not.
The size is real, and it must be told without inflation, because this is precisely the figure that gets exaggerated and then demolished.
The Bank for International Settlements, in Ahmed and Aldasoro’s working paper on stablecoins and safe asset prices, revised in February 2026, uses daily data from 2021 to 2025 and finds that a two-standard-deviation inflow, about three and a half billion dollars, lowers three-month Treasury bill yields by two and a half to three and a half basis points within ten days. The effect is state-dependent, statistically insignificant when bills are plentiful and rising to five to eight basis points when bills are scarce. It is concentrated at the short end, with limited to no spillover to longer tenors. The largest issuer, Tether, contributes the most, followed by Circle. By the issuers’ own reserve reports, the major dollar-backed tokens held more than 270 billion dollars by the end of 2025, about 153 billion of it in Treasury bills, and bought roughly 33 billion of bills over the year.
A modest force at the margin, not a flood.
But the direction is the opposite of the long-end story. The cash layer of the dollar system is deepening on private digital demand at the exact moment the duration layer erodes on fiscal supply.
Bill and bond, once treated as the same asset at different maturities, are being pulled apart by different forces.
Unbundling, inside the Treasury curve itself.
There is a harder irony here, and it carries the rest of the argument.
Stablecoins reinforce the dollar. They do not free the holder from the dollar state. The same Act that mandates the reserve backing places issuers inside a regulatory perimeter and requires that payment-stablecoin systems be able to comply with lawful orders, including orders to seize, freeze, burn, or prevent the transfer of specified tokens.
The digital dollar is not anti-sovereign. It is a programmable extension of sovereign money through private rails.
That makes it a formidable instrument of dollar power, and a weak instrument for escaping that power.
A central bank worried that its dollar reserves can be frozen does not cure the fear by holding a dollar claim that is easier to freeze.
Which is the thread to the next layer.
Reserve Insurance Is Migrating, Not Collapsing
The reserve layer, the assets official institutions hold to anchor currencies and store national wealth, is where the politics of safety turns explicit.
The event was the immobilization of roughly 300 billion dollars of Russian central-bank reserves by the Group of Seven in early 2022. For the first time in the modern reserve era, the official sector watched the core safe assets of a major state turn unusable, not through default, but because the custodian and the clearing system, on political command, denied access.
The asset still existed.
Its owner simply could not touch it.
That distinction, between credit and usability under adversarial conditions, moved from footnote to center.
Start with the case against overstatement, because it is strong.
Dollar dominance is not collapsing. The dollar remains the largest reserve currency by a wide margin, near 57 percent of allocated reserves in the most recent official data, and the institutions that compile that data attribute much of the latest movement to exchange-rate valuation rather than active selling. The dollar still sits on one side of the overwhelming majority of currency transactions. A Federal Reserve study finds that roughly three quarters of official safe-asset holdings belong to governments militarily aligned with the United States, with little reason and less room to run.
Both are true at once.
The dollar can remain dominant at the funding and transaction layer while the official sector diversifies the insurance layer. Different layers, different instruments.
Watch the insurance layer move.
Central banks bought 863 tonnes of gold in 2025, the World Gold Council reports, below the thousand-plus tonnes a year of 2022 through 2024 but far above the 473 tonnes a year of the decade before the freeze. Buying reaccelerated into 2026, 244 tonnes in the first quarter, up 17 percent on the prior quarter and 3 percent on the year, with Poland and Uzbekistan among the leading buyers. The Council’s 2025 survey of central banks, its largest ever at 73 respondents, found that 95 percent expect global official gold reserves to rise over the coming year and a record 43 percent expect to add to their own.
Discipline matters more than drama.
By the end of 2025, the ECB reports, gold reached 27 percent of global official reserves at market value, passing both the euro at 15 percent and United States Treasuries at 22 percent. Striking, and on its own misleading, because it is mostly price. Recompute at end-2023 gold prices, stripping the rally, and Treasuries lead comfortably. Gold is being accumulated steadily and for a clear reason. It has not displaced Treasuries as the core liquidity asset of the official sector, and to say it has is to mistake a price chart for a migration.
Gold also carries a limit its enthusiasts rarely face.
Gold solves seizure better than it solves convertibility. Russia is the proof. It had cut its dollar holdings and stacked gold for years, and when the freeze came that gold sat in Russian vaults, safer from seizure but largely useless for buying what Russia needed in the currencies of the states sanctioning it, except where a counterparty would barter.
Usable control is not one property.
An asset can be seizure-resistant and hard to spend, or easy to spend and exposed to seizure. The reserve manager’s problem is to hold the combination. No single instrument holds it.
Unbundling, at the level of nations.
Beneath the reserve-share headlines runs a quieter shift in who buys American debt.
The Treasury International Capital data for March 2026 showed a net foreign inflow of 150.7 billion dollars, but composition was the point. Foreign private investors bought 162.1 billion while official institutions were net sellers of 11.4 billion. In long-term securities, private foreign investors bought 111.4 billion while foreign official institutions sold 14.9 billion. The Treasury itself cautions that custodial data cannot attribute ultimate ownership with full accuracy.
The Federal Reserve’s custody accounts tell the same story. Treasuries held for foreign official and international institutions stood at roughly 2.69 trillion dollars in late May 2026, down about 225 billion on the year, within a total custody line near 2.97 trillion that itself fell close to 290 billion.
The marginal buyer is changing character.
From the price-insensitive official managing a peg, toward the price-sensitive private investor, the money-market vehicle, the stablecoin issuer, the hedge fund, and the bank balance sheet.
That does not mean no one buys America.
It means the buyer’s motive has changed. And motive governs behavior under stress.
One caveat cuts against over-reading even this, and honesty requires it. A 2025 Federal Reserve study finds the capital data badly understate foreign holdings routed through offshore financial centers, by something like 1.4 trillion dollars for the Cayman Islands alone, which means a chunk of recorded private demand may be leveraged or intermediary flow rather than the slow money of true reserve managers. That does not weaken the composition shift. It makes the front-end layer’s dependence on that demand more fragile, not less.
The Rail Is the Asset
The last and least watched layer of the old bundle is the one under all the others: settlement.
The assumption was not only that the asset would pay. It was that the rail through which the asset moved would remain open.
The reserve freeze shattered that assumption for one country. The response is now visible everywhere.
Europe’s central bankers increasingly speak of payments as sovereignty rather than plumbing. An ECB board member argued in April 2026 that Europe’s dependence on payment infrastructure owned outside the continent is a strategic vulnerability, noting that a large share of euro-area card transactions run on non-European schemes, and casting the digital euro and the interlinking of instant-payment systems as defenses against extraterritorial reach and disconnection. The same logic runs through the IMF’s January 2026 work on cyber risk, which treats concentration in a handful of cloud providers as a systemic concern. A payment claim is only as usable as the rail beneath it, and the rails, it turns out, have owners and jurisdictions and choke points. They have politics.
The same contest plays out, at far larger scale, in computation.
The decisive strategic asset of the decade is not only a financial instrument. It is compute, the chips, data centers, energy, water, grid interconnections, software stacks, and licenses that train and run artificial intelligence. Every major state now treats it as critical infrastructure whose control is security, not commerce.
S&P Global, in May 2026, framed compute sovereignty as a structural risk spanning hardware, software, jurisdiction, and operational control, noting that the licenses governing advanced chips can be suspended or revoked by vendor or government. Not theoretical. The United States authorized exports of up to the equivalent of 35,000 advanced Nvidia chips to state-backed entities in Saudi Arabia and the United Arab Emirates in late 2025, under strict security and reporting conditions.
The chip shipped with a security perimeter attached: licensing, reporting, end-use restrictions, and the continuing discretion of the exporting sovereign.
Compute is the purest case of an asset whose usable control depends on the continued permission of states, vendors, energy systems, and regulators, the same property the reserve managers found in their frozen accounts, now etched into silicon.
And compute is being financialized.
In March 2026 CoreWeave, an artificial-intelligence cloud provider, secured an 8.5 billion dollar delayed-draw term loan to expand its platform, with initial borrowing of about 7.5 billion expandable as its data-center assets reach stable operation. The facility matures in March 2032, was co-structured and book-run by Morgan Stanley and MUFG with Goldman Sachs and JPMorgan as additional lead arrangers, and was anchored by Blackstone Credit and Insurance. It was rated investment grade, A3 by Moody’s, the first investment-grade financing secured by high-performance computing infrastructure and its associated customer contracts.
Capital is moving into the physical layer of digital sovereignty.
But this is not risk-free. It is not a new Treasury.
The credit rests on customer contracts, utilization, power, depreciation, export permissions, and counterparty quality. Apex strength can mask periphery fragility. The investment-grade rating borrows the strength of a few dominant platform customers and lends it to hardware that loses value by the quarter.
The counter-case here is the one I find strongest against the lazy version of the thesis.
States are not passively ceding infrastructure and compute to firms. They are trying to recapture them. Brookings argued in February 2026 that full-stack artificial-intelligence sovereignty is structurally impossible for nearly every country, and that the realistic model is managed interdependence, governments building selective domestic capacity while dependent on a supply chain they cannot replicate. Canada, the European Union, and the Gulf states have all launched sovereign-compute strategies.
The story is not state to firm.
It is contestation. States wield export controls, industrial policy, licensing, public money, and security rules to reassert authority over infrastructure that briefly looked like it might slip away. The most aggressive version, where corporate balance sheets simply inherit sovereign power, ignores the chokepoints below the firm and above the chip: energy, water, land, permits, grid interconnection, and law.
Contestation, not conquest.
The Usable-Control Premium
Pull the layers together and a pattern surfaces.
State it with its limits in the same breath, because the limit is where most writers cheat.
Across every layer, the variable that increasingly governs a safe asset’s value under stress is not only credit. It is usable control, the degree to which the holder can actually deploy the claim when conditions turn hostile.
Usable control depends on where the asset sits in custody, under whose jurisdiction the issuer and registrar fall, who can freeze or seize it, whether it moves and settles when a clearing system is weaponized, whether sanctions render it inadmissible, and whether it can clear at the scale and speed required.
A long Treasury in a vulnerable custody chain, a digital dollar that can be frozen by lawful order, a compute contract exposed to export licensing, and a bar of gold that is seizure-resistant but hard to mobilize all sit at different points on that spectrum.
Two assets with identical ratings can carry different usable control.
That gap is the frontier.
Now the boundary, stated plainly.
That revocation risk is priced is established. The IMF has modeled geopolitical risk and sanctions. The ECB and the European Systemic Risk Board monitor fragmentation. The literature has built news-based geopolitical-risk factors and shown they earn a premium. If the message were only that politics is priced, this would be someone else’s work in borrowed clothes, and it should be rejected.
The open question, the one place something genuinely new might live, is narrower and more demanding.
Existing measures are mostly reactive. They are built from news, from realized comovement, from indices of attention. The question I cannot yet answer, and will not pretend to, is whether an ex-ante structural score of usable control, built from an asset’s fixed characteristics, its custody, its jurisdiction, its issuer freeze power, its vendor dependence, its settlement finality, and its admissibility under sanction, predicts how that asset behaves in an actual revocation event over and above what the existing geopolitical-risk beta already captures.
If a structural score adds nothing beyond the known factor, then usable control is a lens, not a source of return. It is still useful, but it is not alpha.
If it adds genuine incremental power across a panel of real events, the 2022 freeze, the successive sanctions designations, the chip controls, with the reactive factor properly stripped out, then there is a separable dimension of safety that allocators are not yet measuring.
I do not know which it is.
And that sentence is the most important one here.
The answer requires a cross-sectional event study on asset-level returns, spread changes, flows, haircuts, and settlement outcomes, with the known factors removed, on data that lives behind institutional terminals and a scoring frame that must be built and defended asset by asset. That is a research program, not a paragraph, and it cannot be closed by reasoning or by reading, both of which are now exhausted.
What can be said is that the published work visible from outside does not appear to have built that exact ex-ante structural score.
The ground appears open.
Whether it is open because it is valuable and undiscovered, or because it collapses into the existing factor on contact with data, is exactly the question.
Better to hand it over clean than to dress a guess as a finding.
What Would Prove This Wrong
A thesis that cannot be killed is not analysis. It is theology.
So here is what would kill this one, and what would feed it.
The frame is wrong, or overstated, if the dollar’s reserve share holds steady, if Treasury convenience stabilizes or recovers, if official demand proves resilient once offshore custodial plumbing is counted properly, if stablecoin demand stays too small or too unstable to matter even at the bill layer, and if artificial-intelligence capacity expands without serious bottlenecks in power, permits, interconnection, or sovereignty.
If those hold, what reads as unbundling is better called an ordinary mix of fiscal supply, regulation, and geopolitics. That result deserves to be taken seriously, because the strongest counter-argument rests on the same IMF and ECB work this argument leans on. Geopolitical risk may already be fully priced through known channels, and no new factor may be needed.
The frame strengthens if the official sector keeps backing away from the long end at the margin while private and stablecoin-linked demand supports the front; if gold accumulation persists beyond what valuation explains; if long-end convenience stays negative; if payment and compute sovereignty intensify; and if bills and long Treasuries keep behaving less like one asset and more like two layers with different buyers.
The signals are specific, public, and dated, which is the point, because a frame you cannot monitor is a frame you cannot trade.
Watch the monthly Treasury International Capital release for the official-versus-private split.
Watch the weekly Federal Reserve custody figures for official Treasury balances.
Watch the quarterly reserve-composition data from the IMF and the ECB for the dollar, euro, Treasury, and gold shares, valuation-adjusted.
Watch stablecoin assets and their effect on Treasury bill yields.
Watch the power, permits, interconnection, and export policy that decide whether the compute buildout proceeds.
And watch one thing above all the rest.
Under the new Warsh Fed, the question is no longer theoretical. Warsh, sworn in this May, has signaled both continuity and change, and has opened the debate over a leaner balance sheet, even as that question runs on over how far the Fed can shrink it. The ultimate backstop of Treasury usability is not the bond. It is the central bank’s willingness to lend dollars against the bond when the system needs them most, through the swap lines, the FIMA repo facility, and the broader liquidity architecture. A Fed reconsidering the size of its own footprint is a Fed whose crisis-liquidity posture the world will now watch with care.
The day a Fed signals reluctance to lend dollars freely in a crisis, the usability of the entire dollar reserve layer changes.
That is the catalyst calendar.
The Line That Matters
The dramatic claim that some corporation has become safer than the United States deserves a proper burial, because refusing it is the discipline of the whole argument.
The claim usually rests on a comparison of credit-default-swap spreads, and the comparison dies on contact with microstructure.
The market for single-name protection on the United States is tiny, and Federal Reserve research argues that quoted spreads in recent years may rest on few, if any, genuine transactions, which makes them unreliable as a measure of expectations. The deeper mechanical problem is the cheapest-to-deliver option. In the 2023 debt-ceiling episode the cheapest bond deliverable into a hypothetical default auction was a deeply discounted thirty-year Treasury trading near the mid-fifties, so the protection payout, and therefore the spread, reflected that discounted bond and the settlement mechanics rather than any judgment about American solvency. The net amount of protection outstanding was tiny against the deliverable Treasury issue.
Comparing that artifact to a corporate spread is a category error.
Building a grand thesis on it is how a writer gets a single fixed-income specialist to throw out everything else.
The sovereign has not been supplanted by the firm.
What is happening is quieter and larger.
The market is not replacing the sovereign with the firm. It is repricing which instrument, which institution, and which rail controls each separate layer of a safety that used to arrive in one place.
The dollar still rules the funding layer.
Treasury bills still rule the front end, now reinforced by digital dollars engineered inside the regulatory perimeter.
Long Treasuries are sliding out of the old safe-haven role under the weight of supply and a weakened convenience yield.
Gold is the insurance layer, rising as a hedge against seizure while remaining hard to mobilize at scale.
Payments and compute are the contested layers where states fight to recapture control.
The risk-free asset is not dying.
It is unbundling.
And the work of an allocator this decade is to source each layer of safety on purpose, rather than trusting one instrument to deliver them all.
The premium the world paid for the old bundle is gone, because the bundle itself is coming apart.
The task now is to price the pieces.
Whether usable control proves to be merely the right way to see that task, or also a way to earn from it, is the question I take up next.
And I intend to answer it with data, not conviction.
About the author
Shanaka Anslem Perera is an independent analyst writing on macroeconomics, monetary systems, geopolitics, and the economics of computation. His research is produced privately for, and read by, a small number of hedge funds, investment managers, and sovereign and institutional allocators.
Primary sources
Congressional Budget Office, The Budget and Economic Outlook: 2026 to 2036; Wenxin Du, Ritt Keerati, and Jesse Schreger, Decoupling Dollar and Treasury Privilege, Federal Reserve International Finance Discussion Paper No. 1427 and NBER Working Paper No. 35000 (December 2025); Bank for International Settlements, Rashad Ahmed and Iñaki Aldasoro, Stablecoins and Safe Asset Prices, Working Paper No. 1270 (May 2025, revised February 2026); United States Congress, S.1582, the GENIUS Act (enacted July 18, 2025); United States Department of the Treasury, Treasury International Capital data (March 2026 release); Board of Governors of the Federal Reserve System, H.4.1 release (late May 2026); World Gold Council, Gold Demand Trends Q1 2026 and the Central Bank Gold Reserves Survey 2025; International Monetary Fund, Global Financial Stability Report (April 2025, Chapter 2) and the Currency Composition of Official Foreign Exchange Reserves; European Central Bank, The International Role of the Euro (June 2026) and remarks on payments sovereignty (April 2026); European Central Bank and European Systemic Risk Board, joint work on geoeconomic fragmentation; Federal Reserve Bank of Chicago, What Does the CDS Market Imply for a U.S. Default? (2023); Federal Reserve Bank of New York, Liberty Street Economics, The Evolving Market for U.S. Sovereign Credit Risk; Federal Reserve Board research on reserve holdings and geopolitical alignment; CoreWeave, Inc., financing announcement and United States Securities and Exchange Commission filings (March 2026); S&P Global, Compute Sovereignty (May 2026); Brookings Institution, analysis of artificial-intelligence sovereignty (February 2026); Council on Foreign Relations, analysis of the immobilized Russian reserves; and reporting by Reuters on the Moody’s downgrade, the CoreWeave financing, the Gulf semiconductor export approvals, and the confirmation and swearing-in of Federal Reserve Chair Kevin Warsh.
Disclaimer and disclosures
This article is published for informational and educational purposes only and represents the analysis and opinion of the author as of June 3, 2026. It does not constitute, and must not be relied upon as, investment, financial, legal, tax, accounting, or other professional advice, nor a recommendation, solicitation, offer, or inducement to buy, sell, or hold any security, currency, commodity, digital asset, derivative, or other financial instrument, or to adopt any investment or trading strategy, in any jurisdiction. Nothing herein is tailored to the circumstances, objectives, financial situation, or risk tolerance of any individual or institution, and no fiduciary, advisory, or client relationship is created by its publication or by your reading of it.
The analysis contains forward-looking statements, scenarios, probabilities, falsification conditions, and research hypotheses that are inherently uncertain and subject to change without notice. Markets, policies, credit ratings, reserve compositions, legislation, regulations, and the data described here can and do change. Projections, including those of the institutions cited, may not be realized. Named institutions, instruments, issuers, and securities are referenced solely as the objects of factual and analytical commentary, not as endorsements or criticisms of any entity.
Figures and statements attributed to third parties, including the Congressional Budget Office, the Bank for International Settlements, the International Monetary Fund, the European Central Bank, the Federal Reserve and the Federal Reserve Banks, the World Gold Council, the rating agencies, corporate and regulatory filings, and news organizations, are reported in good faith from sources believed to be reliable as of the date of writing. Their accuracy, completeness, and continued currency are not guaranteed. Certain figures, including estimate-quality, survey-based, issuer-reported, or model-derived data, are identified as such in the text. Independent verification against primary sources is encouraged before any reliance.
The author and any affiliated persons or entities may hold, or may in the future hold, positions in the assets, sectors, currencies, or instruments discussed, and may transact at any time without notice. To the maximum extent permitted by applicable law, the author and the publisher disclaim all liability for any direct, indirect, incidental, consequential, special, or other loss or damage arising from any use of, or reliance on, this material or its contents. Readers are solely responsible for their own decisions and should consult their own licensed and qualified professional advisers, and conduct their own independent due diligence, before acting on any information or view expressed here. This material is the intellectual property of the author, may not be reproduced or redistributed in whole or in part without attribution, and is not directed at, or intended for use by, any person in any jurisdiction where such publication, distribution, or use would be contrary to law or regulation. Any analytical framework introduced here, including the usable-control score, is a research hypothesis rather than an established model, unless and until it is independently tested, replicated, and validated on appropriate data.
If settlement infrastructure itself becomes part of the risk-free stack, then whoever controls settlement infrastructure gains enormous structural importance.
Basically saying modern finance increasingly works like Jenga built from objects everyone agrees are unmovable—and then we keep adding more floors because the tower hasn’t fallen yet.
The usable-control frame maps directly onto physical supply chains. The DFARS January 1 deadline is a usable-control failure - the magnet exists, but the supply chain that makes it deployable under the new procurement architecture doesn't. Different layer, same structure.