As BlackRock, JPMorgan, and the NYSE race to tokenise everything, the fund that guards global financial stability is issuing a quiet but urgent warning: build this wrong, and the next market crisis won’t just spread faster — it will be structurally impossible to stop.

There is a particular kind of danger in making financial systems too efficient. The history of modern finance is, in some essential way, the story of buffers — the friction, the delay, the human pause between a decision and its consequence. Margin calls that took hours to process. Settlement cycles that ran two business days. Clearing houses that smoothed out the chaos of simultaneous trades. These inefficiencies were not design flaws. They were, in the language of systems engineering, shock absorbers. And Wall Street has spent the last three years engineering a future in which they no longer exist.

That future now has a name — tokenised finance — and it is accelerating with the kind of momentum that tends to outpace regulatory architecture by a decade. BlackRock’s BUIDL fund, launched in 2024 on the Ethereum blockchain and now managing over $2.5 billion in assets, has become the flagship of a movement. JPMorgan’s Kinexys platform — which powers the bank’s new My OnChain Net Yield Fund, known as MONY — made JPMorgan the largest globally systemically important bank to launch a tokenised money-market fund on a public blockchain. Earlier this year, the New York Stock Exchange announced plans for a blockchain-based venue that would allow investors to trade tokenised stocks and ETFs around the clock — no settlement delay, no market hours, no pause. According to data from rwa.xyz, tokenised real-world assets have now crossed $27.5 billion on-chain, with U.S. Treasury products alone accounting for more than $12 billion of that total.

This is not speculative anymore. This is infrastructure being built in real-time, at systemic scale, by institutions that sit at the very centre of the global financial plumbing.

And now the IMF wants you to pay attention to what happens when the plumbing has no pressure-relief valve.

What Tobias Adrian Actually Said — and Why It Matters More Than Headlines Suggest

On April 1, 2026, the IMF published a note that deserves to be read in full rather than summarised in a press release. Authored by Tobias Adrian, the Fund’s Financial Counsellor and Director of the Monetary and Capital Markets Department, “Tokenized Finance” does not read like a bureaucratic warning. It reads like a structural diagnosis.

Adrian’s central argument is precisely the one that Wall Street’s boosters tend to dismiss: that tokenisation is not a marginal efficiency improvement. It is, as the note puts it, “a structural shift in financial architecture” — one that fundamentally alters how trust, settlement, and risk management function at the system level. The distinction matters enormously. You can regulate a product. Regulating an architecture requires thinking several orders of magnitude further ahead.

The paper identifies four specific risk categories that deserve unpacking, because each one is more counterintuitive than it first appears.

The Four Risks: Why Speed Is the Most Dangerous Efficiency of All

1. The Temporal Buffer Problem

Traditional finance has always embedded time into its risk management. When a bank faces a margin call in a T+2 settlement system, it has approximately 48 hours to locate liquidity, assess counterparty exposure, and if necessary, contact a regulator. Central banks are designed around this rhythm — their liquidity tools, their emergency facilities, their intervention protocols — all calibrated to business-day cycles.

As the IMF note warns, tokenised systems replace this architecture with automated margin calls, continuous settlement, and algorithmic feedback loops that compress the intervention window to near-zero. Think about what that means in practice. A stress event that currently gives a regulator 48 hours to respond might, in a fully tokenised system, leave a window of 48 minutes — or less. The irony is acute: the feature Wall Street is selling as a benefit (instant settlement, 24/7 operation) is precisely the feature that turns a liquidity problem into a systemic cascade.

This is not theoretical. The crypto markets have already run this experiment, repeatedly. In May 2022, the collapse of the Terra/Luna ecosystem wiped out approximately $40 billion in value in 72 hours — a speed of contagion that no traditional market mechanism could have produced. Tokenised finance, at institutional scale, with real-world assets as collateral, would be playing the same game with significantly higher stakes.

2. Concentration and Shared Infrastructure Risk

The second risk is one that financial stability analysts recognise from the pre-2008 era — but with a modern twist. When multiple institutions route through the same ledger infrastructure, errors in smart contracts or infrastructure failures can affect all participants simultaneously, rather than in the sequential fashion that allows for containment. The IMF calls this a concentration risk embedded in shared ledger architecture — a form of correlated exposure that has no real precedent in traditional finance.

One bug in a smart contract governing a trillion-dollar collateral ecosystem is not an operational nuisance. It is a systemic event.

3. Fragmentation and the Liquidity Silo Problem

Here is where the picture becomes genuinely paradoxical. Tokenisation promises to improve market liquidity — and in narrow, isolated conditions, it does. But if multiple platforms emerge with incompatible standards and siloed liquidity pools, the aggregate effect is the opposite. The IMF warns that fragmentation across platforms reduces netting efficiency and impairs the par convertibility between assets — the ability to treat different instruments as equivalent for settlement purposes — which is a foundational assumption of modern financial plumbing.

Imagine ten competing tokenisation platforms, each hosting its own version of tokenised Treasuries, each with slightly different legal frameworks and redemption mechanisms. In a calm market, they coexist. In a stress event — when institutions rush simultaneously to redeem and convert — the absence of common standards turns a manageable liquidity event into a multi-front crisis with no coordinating mechanism.

4. Cross-Border Chaos and the Jurisdictional Void

Tokenised transactions are, by their nature, borderless. A smart contract executing on Ethereum does not consult a legal jurisdiction before settling. But dispute resolution mechanisms remain stubbornly national — and the legal status of tokenised assets remains, in most jurisdictions, profoundly uncertain. Who owns a tokenised bond during an insolvency? Under which country’s law? In which court? The IMF flags that the principle of “code is law” — beloved by blockchain maximalists — is not a substitute for legal certainty, especially for instruments sitting inside systemically important institutions.

The Emerging Market Dimension: Who Bears the Tail Risk?

If you are reading this from London, New York, or Frankfurt, the systemic risks of tokenised finance feel abstract — contained, perhaps, to the corridors of Canary Wharf or Wall Street. That is a dangerous perspective.

For economies in Latin America, sub-Saharan Africa, and South and Southeast Asia, the tokenisation wave carries a specific and asymmetric danger: monetary sovereignty. A previous IMF analysis documented how dollar-backed stablecoins are already accelerating currency substitution in high-inflation economies — Argentina, Turkey, and Venezuela are the textbook cases. Privately issued global stablecoins, if they become dominant settlement assets in tokenised markets, could displace local currencies in exactly the jurisdictions where central banks have least capacity to respond.

This is the emerging market dimension of tokenisation risk that most Western commentary ignores entirely: the possibility that the infrastructure of the next financial crisis will be built in New York but the worst of its consequences will be felt in Nairobi, Jakarta, or Buenos Aires. The capital-flow volatility implications alone — tokenised assets enabling frictionless cross-border movement — could destabilise exchange rates in ways that make the 1997 Asian financial crisis look well-cushioned by comparison.

Why This Is Not Anti-Innovation Scaremongering

Let us be precise about what this argument is not.

The efficiency gains from tokenisation are real and measurable. Atomic settlement — the simultaneous exchange of asset and payment — eliminates counterparty risk in a way that the T+2 system never could. Embedded compliance through programmable assets reduces the cost of regulatory reporting. Continuous liquidity management allows institutions to optimise collateral use in ways that genuinely benefit end investors. BlackRock’s Larry Fink has called for the entire financial system to run on a common blockchain — and while the maximalism of that vision invites scepticism, the underlying logic about settlement efficiency is sound.

The IMF note, to its credit, does not dispute any of this. Adrian explicitly acknowledges the benefits. His argument — and mine — is architectural, not ideological. The question is not whether to tokenise finance. That ship has already sailed, and the $27.5 billion on-chain today is merely the early tide. The question is how the foundational infrastructure is built, and who bears the systemic risk when it fails.

A high-speed train is not inherently dangerous. A high-speed train without adequate braking systems, operating on tracks without standardised gauges, running through tunnels with no emergency protocols — that is a different proposition entirely.

The IMF’s Five-Pillar Prescription: A Policy Architecture Worth Taking Seriously

Adrian proposes a five-part framework that is more rigorous than most regulatory roadmaps currently circulating in parliamentary committees and central banking briefing rooms.

First, anchor settlement in safe money — specifically, wholesale central bank digital currencies (wCBDCs) or equivalent public-sector settlement assets. The principle is straightforward: the trust that makes financial systems function cannot be fully outsourced to private infrastructure. Settlement in private stablecoins, however well-designed, creates a dependency on entities that lack the public accountability and unconditional backing of a central bank.

Second, apply consistent regulation to economically equivalent activities. A tokenised money-market fund that functions identically to a traditional money-market fund should face the same regulatory requirements — regardless of the technology underlying it. This sounds obvious. In practice, regulatory arbitrage between tokenised and traditional instruments is already emerging, and the IMF is right to call it out early.

Third, establish legal certainty for tokenised assets — clarifying ownership rights, creditor protections, and jurisdictional applicability before the scale of these markets makes retroactive legal reform prohibitively complex.

Fourth, promote interoperability standards that prevent the fragmentation of liquidity across incompatible platforms. This is a role for standard-setting bodies — the BIS Innovation Hub, IOSCO, the Financial Stability Board — rather than individual institutions, and it requires coordination across jurisdictions that rarely cooperate at speed.

Fifth, and most ambitiously, adapt central bank liquidity tools to a 24/7 automated environment. This is the deepest structural challenge. The Federal Reserve, the European Central Bank, the Bank of England — all of them are currently calibrated to intervene at business-day frequency. A financial system that settles continuously requires lender-of-last-resort tools that operate continuously. That is not a software update. It is a fundamental reimagining of what central banking looks like in the digital age.

The Window Is Open — For Now

There is a line near the end of Adrian’s note that should be quoted wherever serious people discuss financial architecture: “The window for shaping the architecture of the tokenised financial system is open, but it will not remain so indefinitely.”

That is not bureaucratic boilerplate. It is a precise statement about the path-dependency of infrastructure decisions. Once JPMorgan’s MONY fund has $100 billion under management. Once the NYSE’s tokenised equities platform is processing ten million trades a day. Once the legal precedents that emerge from the first tokenisation-related insolvency have hardened into case law in six different jurisdictions simultaneously — at that point, the architecture is largely fixed. You can regulate at the margins. You cannot redesign the plumbing while the city runs on it.

Policymakers — at the Federal Reserve, the Bank for International Settlements, the Financial Stability Board, and in finance ministries from Washington to Brussels to Singapore — have a narrow window to establish the public infrastructure and coordination frameworks that could make tokenised finance genuinely safe. What that requires, practically, is not regulatory hostility to innovation but something considerably harder: regulatory ambition. The willingness to build public infrastructure ahead of private demand. The discipline to enforce interoperability before fragmentation becomes entrenched. The foresight to extend central bank tools into a 24/7 world before the first crisis demonstrates why it was necessary.

The private sector is moving fast. BlackRock, JPMorgan, the NYSE — they are not waiting. The public sector, historically, moves slower. In this case, the asymmetry between those two speeds is itself the systemic risk.

The IMF has named it. The architecture remains, for now, unbuilt. That is the most important financial stability question of 2026 — and the one that will define the next crisis when it comes.