The €100bn DLT Bet in Capital Markets:
Moving from Lighthouse to Execution

ArgumentThe Through-Line

The tokenization of markets was never a technological problem. Native digital issuance has worked since 2019, and atomic DvP in central bank money since Project Helvetia. What was missing was something else: a Settlement Asset, interoperability, and a use case with enough pull to cross the threshold into a network effect. For most of the decade, the industry substituted pilots for all three — and a pilot, by its very nature, generates neither recurring volume, nor secondary liquidity, nor the counterparty network that would make the build-out economically defensible in the first place, beyond the crypto headline.

The shift in regulation is now matched by capital: the large infrastructure providers are beginning to invest heavily in crypto capital-markets infrastructure. The regulatory side is setting the tempo. The EU’s Market Integration Package of December 2025 proposes to raise the aggregate ceiling from €6bn to €100bn, to let MiCA e-money tokens satisfy the Cash Leg, and to mandate interoperability standards. The United States has enacted a federal stablecoin regime, and the SEC has cleared the DTCC’s path onto the Canton Network. The United Kingdom has committed to a live synchronization service in central bank money. Each of these steps clears one of the three blockers, and together they are what is drawing the investment. Yet none of them, on its own, produces what turns a pilot into a P&L: a workflow that multiple houses must use jointly and repeatedly because the old method is demonstrably worse.

This briefing follows exactly that logic. It first separates the instruments — because half the market’s confusion is taxonomic in nature — then the regimes, because the regulatory perimeter now determines who is allowed to do what, and finally the players, sorted by the business they actually run rather than by the press release. It closes on the pilot question itself: why the isolated proofs of concept delivered so little — and what concretely differs today.

IThe Instrument Taxonomy

Three fundamentally different things travel under the word “tokenization,” and conflating them is the most common error in practice. Once you cleanly separate three axes — the Asset (what is tokenized), the Settlement Asset (against which money it settles), and the Issuance Regime (under which legal framework it is issued) — most of the strategic disputes resolve themselves.

Native Digital Securities vs. Tokenized Wrappers

The decisive line runs between an asset that originates natively on a ledger and a token that merely references an asset held in custody elsewhere.

Asset tokens, or native digital securities, are the security itself: issued on the ledger and recorded there as legal proof of ownership. Switzerland’s register securities (Registerwertrechte), Luxembourg’s DLT-issued dematerialized securities, and Clearstream’s D7-DLT issuances belong here. No “real” off-chain version exists — the token is the instrument. This is the clean architecture — and it is the one regulators have worked hardest to enable.

Tokenized assets, or wrappers, securitize a claim on an externally held asset — economically, a digital depositary receipt. xStocks, BlackRock’s BUIDL, Ondo’s OUSG, and tokenized equities fall into this category: a licensed custodian holds the underlying share, bond, or fund unit and issues a transferable 1:1 token against it.

The Settlement Asset: A Hierarchy of Cash Legs

No security or token settles itself. The quality of a tokenization is determined in large part by its settlement — and a clear counterparty-risk hierarchy must be observed.

The entire efficiency promise of tokenization — atomic, instantaneous DvP — depends on a tokenized Cash Leg of sufficient quality. Throughout the pilot decade this Cash Leg simply did not exist, so the benefit could not be demonstrated at all. The changes of 2025/26 — EMTs as a permitted Cash Leg, live deposit tokens, wholesale CBDC in production — are what finally make building out the rest of the stack worthwhile.

IIThe Regulatory Architecture

Five jurisdictions matter, and they have divided into two philosophies: enable native issuance through securities law (EU member states, Switzerland) or regulate the money and the service providers (US, MiCA crypto track). The EU is now attempting to do both at once.

EU — The DLT Pilot Regime and Its Rescue

The DLT Pilot Regime launched in March 2023 with three infrastructure categories: a DLT MTF, a DLT settlement system, and a combined DLT trading and settlement system. It is widely regarded as a near-failure — by mid-2025 only a handful of infrastructures had been authorized, and live transactions remained negligible. ESMA’s review of June 2025 attributed this not to a lack of interest but to the design: thresholds set too low (an aggregate ceiling of €6bn), a perceived three-year sunset date that deterred investment, and unresolved questions around the Cash Leg and interoperability. ESMA recommended making the regime permanent, raising the limits substantially, and broadening the range of eligible assets.

MiCA — The Crypto Perimeter, Not the Securities Perimeter

MiCA became fully effective in 2024/25 and is the world’s most comprehensive crypto framework; more than 90 CASPs were authorized in its first year. The stablecoin regime distinguishes e-money tokens (single-currency, issued by an EMI or bank) from asset-referenced tokens (multi-asset), each with obligations on reserves, redemption at par, and disclosure, plus an escalation into joint EBA supervision above certain size thresholds. For capital-markets work, one boundary matters above all: MiCA expressly does not capture instruments that qualify as financial instruments — those remain under MiFID and CSDR. Tokenized securities are therefore a question of securities law; only the stablecoin money side is a MiCA question. Grandfathering for established providers runs, in many places, through mid-2026.

Luxembourg — The Workbench of Native Issuance

Luxembourg has worked in four deliberate steps. Blockchain I (2019) recognized DLT for the circulation of securities. II (2021) allowed authorized issuers to issue via DLT and created the central account-keeper for uncertificated debt securities — the basis of the EIB’s digital bond program. III (2023) extended financial-collateral law to DLT-held instruments and placed creditors on an equal footing. IV (Dec. 2024) introduced the optional Control Agent as an alternative to the central account-keeper, opened the framework to equity securities and fund units, and enabled direct holding. The Control Agent model is the real lever: it permits a flatter custody structure and opens DLT issuance to the fund industry — directly relevant to the retailization of private funds.

Switzerland — The Most Complete Regulated Stack

The Swiss DLT Act (2021) made register securities (Registerwertrechte) a fully valid legal form and created a FINMA license for DLT trading systems. Only that allowed SDX to operate as the first FINMA-licensed DLT central securities depository and exchange at once — and made the settlement of genuine digital bonds against wholesale CBDC in Project Helvetia legally clean rather than experimental. Switzerland is the only jurisdiction in which native issuance, regulated trading, on-DLT custody, and central bank money settlement are all live in production today.

United Kingdom — Infrastructure-Driven, Anchored in Central Bank Money

The United Kingdom is betting on modernizing the settlement rail rather than licensing a parallel ecosystem. The Digital Securities Sandbox (Bank of England + FCA) is shepherding around 16 firms through live issuance and settlement. The renewed RT2 core (April 2025) is designed for near-24/7 operation. A Synchronisation Lab (spring 2026, 18 participants) is testing the connection of RT2 to external and DLT ledgers for conditional settlement in central bank money; a live synchronization service is targeted for 2028. HM Treasury is piloting a digital gilt (DIGIT), and the Bank is working to make tokenized counterparts of already-eligible assets usable as collateral. On stablecoins, the Bank is deliberately restrained: in core settlement, liabilities should not migrate out of central bank money into private stablecoins — it would sooner extend RTGS than relinquish the Cash Leg.

United States — Money First, After a Long Standstill

The United States changed course decisively in 2025. The GENIUS Act (July 2025) created a federal framework for payment stablecoins: licensed issuers, high-quality 1:1 reserves, redemption at par, a bankruptcy-remote structure, BSA compliance — and a USD 10bn threshold above which issuers fall under direct federal supervision. Notable is the ban on paying interest to holders: it draws a sharp line between stablecoins (payment) and tokenized money market funds (yield). On the securities side, the SEC’s no-action letter of December 2025 allowed the DTCC to pilot the tokenization of highly liquid assets; the broader “modular taxonomy” (collectibles / commodities / stablecoins / securities) superseded the earlier stance of treating tokens as securities when in doubt. Asia moved in parallel — Hong Kong’s Stablecoin Ordinance (Aug. 2025) and Singapore’s MAS framework license fiat-referenced issuers on reserve-backed terms.

IIIThe Players, by Business Model

Sorted not by industry label but by the business each actually runs. Four archetypes are vying for different layers of the same emerging stack, and the striking thing is how little, so far, they get in one another’s way.

Synthesis Matrix

IVThe Pilot Trap

Here is the question the industry prefers not to dwell on: why did so many institutions run tokenization transactions, only to shut the experiment down for lack of practical utility? The answer is structural, and it does no one a favour to soften it.

A pilot is optimized for a press release, not for a P&L. A single bilateral digital bond issuance proves technical feasibility — but that was never seriously in doubt. What an isolated transaction cannot generate is precisely what matters commercially: recurring volume, a secondary market, and a network of counterparties who must all be present at once for even a single trade to come together.

Why the Isolated Use Cases Delivered So Little

• No settlement asset. Without a tokenized Cash Leg of sufficient quality, the actual benefit — atomic DvP — could not be demonstrated. Most pilots settled the Cash Leg off-chain via a manual trigger, thereby preserving the very settlement-risk window that tokenization was meant to eliminate. Efficiency was asserted, not proven.
• No interoperability — fragmentation as an “economic tax.” Each pilot ran on its own chain, with its own standards, its own smart-contract logic, its own settlement design. An asset stranded on an isolated network is — as the post-trade utilities’ own joint paper puts it — a source of high operating costs and divided liquidity.
• No counterparty on the other side. A token is transferable at any time, but transferability is not yet a market. Absent the buyer, “instant settlement” settles nothing. Market makers will not commit balance sheet to a venue with no flow, and there is no flow until the market makers are already there: the cold-start problem, undisguised.
• Assets, not workflows, were the starting point. The same recurring design error: the instrument was tokenized, only to then go looking for a platform and a problem it could solve. The right way around would be to find an expensive workflow — margining, intraday repo, collateral substitution — and make it cheaper through efficient tokenization.
• Sandboxes for experiments, not for business. The DLT Pilot Regime’s €6bn ceiling and its perceived sunset date kept everyone from investing at production scale. The regime got the one-off experiments it was built for — and little else.
• Was the headline more important than the deal? Innovation budgets, league-table placings, and the prestige of the “first digital bond” were measured by the announcement, not by the volume that followed. In many houses, the pilot thus became the goal itself — rather than the first step of a viable business.

What Now Actually Differs

Whether the step beyond the pilots is worthwhile hinges on whether the three structural blockers dissolve simultaneously — because solving one without the others changes nothing.

The delay deserves an honest explanation, because the usual one — the technology wasn’t ready — is false. Take collateral mobility, the one use case with a credible path across the network threshold. It did not need to be invented. HQLAx has run it on a distributed ledger since December 2019: built on Corda, backed by Deutsche Börse, swapping ownership of collateral baskets between the largest dealers without the underlying securities ever leaving their custodians. It does exactly what the efficiency case promises, and it has done so, in production, for the better part of a decade. That it has not swept the industry is the tell. The obstacle was never the rails; it is that capital markets is structurally misincentivised to fix a cost it has organised itself around. Every desk funds itself, every silo optimises locally, and the savings — a recurring sum that runs comfortably into the hundreds of millions — fall in a gap no single P&L owner is paid to close. An industry that does not fully grasp the price of its own fragmentation does not adopt the cure simply because the cure exists.

So the pilot era can finally end — not because the technology improved, but because the settlement asset, the interoperability mandate, and a real use case are arriving together for the first time. Those who treated their pilots as option value, reusable infrastructure and standards positioning rather than one-off demonstrations, are ready to make the switch. Those who booked each pilot as a finished deliverable are left holding a portfolio of impressive, unconnected, non-recurring transactions, and are only now beginning the real work.

VTaxation Across the Regimes

Tax is no footnote to legal form — it is the same question, only from the tax authority’s vantage point. Because the tax characterization follows the legal form, a single structuring decision determines the entire downstream treatment: the applicable statutory provision, the automatic-information-exchange regime, and whether the holder can reach withholding-tax relief and treaty benefits at all.

The CARF/CRS Switch — The Consequential One

From 2026 (DAC8 in the EU, first exchanges in 2027), the OECD’s Crypto-Asset Reporting Framework runs alongside the Common Reporting Standard (CRS). Which regime an instrument falls under is decided by a precise test — and that test maps almost exactly onto the CSDR-versus-MiCA dividing line.

The consequence is significant. A native digital security held in custody via a central securities depository remains a CRS financial asset: reported by financial institutions, on the familiar account basis, alongside any other security in the portfolio. SDX, for instance, obtained confirmation from the Swiss Federal Tax Administration that its digital bonds are CRS financial assets and not crypto-assets — precisely on the disintermediation argument. A public-chain wrapper, by contrast, which can migrate to a self-hosted wallet, is a crypto-asset: it pulls holders and service providers into CARF’s transaction-level reporting — every acquisition, every disposal, every transfer to a self-hosted wallet — operationally heavier and reputationally crypto-tainted in a way that many institutional committees will not endorse.

Withholding Tax and Treaty Access

Here the depositary-receipt analogy bites hardest. A security held in custody via a central securities depository plugs directly into the well-oiled machinery for withholding-tax relief at source, refunds, tax vouchers, and double-taxation treaties — the documentation chain on which treaty relief depends runs through recognized custodians. A public-chain crypto-asset or a self-custodied token, as a rule, does not reach relief at source: recognized custody and documentation are simply absent. As with the ADR, all treaty access runs through the custodian’s compliance apparatus, not through the token — and where no custodian in the recognized sense stands behind it, the investor bears the gross withholding tax and forfeits the refund. For a cross-border institutional book, that is no rounding error.

Income Characterization — Germany as the Model Case

Legal form determines the statutory provision, and the German treatment stands as exemplary for the EU pattern. Under the BMF circular of 6 March 2025, a security token is treated like a security — investment income under §20 EStG, a flat withholding tax (Abgeltungsteuer) of 26.375%, tax withheld at source, reported in Annex KAP. A crypto-asset held as private assets falls under §23 EStG as a private disposal transaction — taxed at the personal rate of up to roughly 45% plus surcharge, but tax-free after a one-year holding period.

Value-added tax is largely neutral and not a differentiator — both the crypto exchange (in the logic of the CJEU’s Hedqvist judgment) and securities transactions fall within the exempt sphere. Stablecoins and wholesale CBDC are excluded from CARF as specified e-money products or as CBDC, respectively, but folded into the amended CRS — which again shows: the money side is being pulled toward the regulated reporting perimeter, not away from it.

VIWhy Institutions Should, for Now, Hold CSDR Rather Than MiCA Instruments

MiCA and CSDR are mutually exclusive by construction. An instrument is either a financial instrument under MiFID — and thus within CSDR’s settlement and securities-law perimeter — or a crypto-asset under MiCA. It cannot be both at once; MiCA expressly carves out financial instruments. The decision is made at structuring, and for any regulated balance sheet it should today fall, without hesitation, on the CSDR side.

The reason is not ideological. It is that a MiCA instrument carries, for an institutional holder, a whole series of structural deficits — and, crucially, that no benefit stands against them which an institution could actually realize.

The Structural Deficits of MiCA Instruments

Each row, on its own, is a friction. Taken together, they are disqualifying for a regulated investor: an instrument that cannot be posted as collateral, that ties up 1,250% capital, that sits outside the protection of settlement finality, that may breach mandate limits, and that forfeits treaty relief is no marginal decision — for most institutional purposes it is simply untenable.

A native digital security issued and settled via a central securities depository — including a DLT central securities depository under the Pilot Regime — captures the entire DLT efficiency set: programmability, near-atomic DvP, 24/7 capability, automated servicing. And it does so without leaving the financial-instrument perimeter. The institution therefore never has to surrender its protections to reach the technology. The MiCA path demands the opposite: to give up settlement finality, collateral eligibility, normal capital treatment, mandate compliance, and treaty access — in exchange for permissionless transferability, DeFi composability, and retail reach, the very properties an institution does not need and often may not use. In every direction that matters for a regulated holder, it is a bad bargain.

Why “For Now” — and Why the Caveat Supports the Thesis

The judgment is expressly point-in-time. It would soften if the regulatory-capital treatment were recalibrated, if custody and settlement finality on the MiCA side reached securities-grade level, and if collateral and mandate frameworks adapted. But note the direction of the reforms underway: the Market Integration Package does not narrow the gap by making MiCA crypto-assets institution-eligible, but by pulling tokenization into the CSDR/financial-instrument perimeter — EMTs as a regulated Cash Leg, settlement finality extended to DLT via the SFR, mandatory interoperability of DLT post-trade infrastructures, a higher Pilot Regime ceiling. The regulation is making the CSDR side DLT-capable, not the MiCA side institution-capable. That does not weaken the CSDR-default thesis — it is its strongest argument.

VIIStrategic Implications

Six propositions that follow directly from the foregoing — tailored to an institution that must decide where to commit.

• Default to the CSDR side. For any regulated balance sheet, the legal-form decision should fall categorically in favor of the financial instrument and against the crypto-asset. The DLT efficiency is fully available within the CSDR perimeter; the MiCA wrapper forfeits settlement finality, collateral eligibility, capital treatment, mandate compliance, and treaty access — for benefits institutions cannot use. MiCA exposure is the exception requiring justification, not the default.
• Choose the layer, not the technology. The four archetypes occupy different layers and, so far, barely get in one another’s way. The first decision is which layer a house defends or attacks — securities issuance, the wrapper, distribution/data, or the Cash Leg — because each has its own incumbent, its own moat, and its own timeline.
• The Cash Leg is the contested terrain. Whoever controls the settlement asset controls the economics of the network. The contest between bank deposit token and stablecoin is the one to watch, and it is being fought right now.
• Critical mass before openness. The infrastructure should be designed as an open connectivity layer with proprietary differentiation at the service layer — but the opening follows after critical mass, not the other way around. Whoever opens before liquidity is there arms the competitors without building a network of their own.
• Think from the workflow, lead with collateral. The only use case with a credible path across the network threshold today is collateral mobility — the asset pools already exist at scale, and the cost of stranded collateral is measurable. A tokenization initiative that addresses no concrete, expensive workflow remains a pilot without a business model.
• Bet on interoperability, not on a winning ledger. No single chain will dominate; the incumbents have conceded as much. Value accrues to standards, gateways, and those who knit the Network of Networks together — the connectivity position is what counts, not the protocol bet.