ECB's Schnabel Laid Out Three Ways to Put Central Bank Money On-Chain. One Is Already Live.

At the Jackson Hole Economic Policy Symposium on August 28, 2026, ECB Executive Board member Isabel Schnabel gave a speech titled "Central Banks On-Chain." The framing was deliberate: not "central banks and blockchain," not "the ECB's digital currency strategy." Central banks. On-chain. The institution that issues euros is going to live on distributed ledger infrastructure. The question she addressed is how.

Schnabel's answer — three possible models — is the most rigorous public analysis of the architectural choices facing every major central bank in the tokenization era. And unlike most central bank speeches, this one came with an announcement: the ECB's Project Pontes launched on September 21, 2026, making Model 2 operational.

The Foundational Argument: Why Stablecoins Cannot Replace Central Bank Money

Schnabel's case for why central banks must go on-chain starts with a claim about what stablecoins cannot do. Central bank reserves are the ultimate settlement asset in wholesale financial markets because they carry no credit risk and no liquidity risk — the central bank can always make good on them, and it can expand the supply elastically during periods of financial stress. A stablecoin, regardless of its reserve composition, is a liability of a private issuer. If that issuer faces a run, the stablecoin cannot expand its supply to meet redemptions. The central bank can.

This distinction matters most in crises. During the 2008 financial crisis and the 2020 COVID shock, central banks provided liquidity to the system through reserve expansion — creating new central bank money to prevent settlement failures. A stablecoin-dominated settlement system would have no equivalent mechanism. Private stablecoins are, as Schnabel put it, "complements to central bank money, not substitutes for it." The system needs central bank money in the settlement layer to function as a lender of last resort — which means central bank money must be present on the same rails as the assets being settled.

Model 1: Direct Issuance on a Distributed Ledger

The most ambitious model: central bank reserves become native assets on a distributed ledger. The ECB issues tokenized reserves directly on a programmable platform — not a bridge to an existing system, not a representation of reserves held elsewhere, but actual central bank money living on-chain. Smart contracts can settle against it atomically. It participates in DeFi protocols as collateral. It is programmable, composable, and natively on-chain.

The advantage is completeness. Atomic delivery versus payment — simultaneous exchange of assets and central bank money in a single transaction with no settlement risk — becomes possible for any asset on the same ledger. Monetary policy implementation gains new tools: interest rate changes can be applied programmatically, reserve requirements can be enforced algorithmically.

The challenge is governance and infrastructure risk. A single unified ledger that carries both commercial bank money and central bank money becomes a systemically critical piece of infrastructure. Its failure mode is catastrophic rather than localized. Design choices about who can operate nodes, how forks are resolved, and how upgrades are implemented become monetary policy decisions — not just technology decisions.

Model 2: The Bridge (Pontes) — Now Live

The intermediate model keeps central bank reserves in existing systems — TARGET2 for the euro area — and connects those systems to distributed ledger platforms operated by market participants via a bridge layer. Tokenized assets settle on the DLT platform; the cash leg settles in central bank money via the bridge, which triggers a conventional TARGET2 transfer simultaneously.

This is Pontes. It launched September 21, 2026. The ECB operates a Eurosystem DLT platform that connects TARGET services to institutional DLT platforms. Settlement is available in two modes: conventional TARGET2 settlement or direct DLT settlement, depending on what the counterparties prefer. The cash leg is always in central bank money — the bridge ensures that.

Pontes's initial contact group of 61 institutions began work in September 2026. The ECB ran DLT settlement trials in 2024 with 64 participants across nine jurisdictions, settling approximately €1.6 billion in central bank money. Since March 2026, the ECB has accepted DLT-based assets as eligible collateral for its own credit operations — meaning institutions can already pledge tokenized assets to the ECB and receive central bank money against them.

Model 3: Private Settlement Tokens Backed by Central Bank Reserves

The third model does not put central bank money on-chain at all. Instead, regulated private institutions — commercial banks — issue settlement tokens that are fully backed, one-for-one, by central bank reserves held at the ECB. The token is a commercial bank liability; the reserve is a central bank asset. The token can settle on any blockchain; the reserve backing is off-chain at the ECB.

This model preserves the existing two-tier monetary structure — central bank money backstops commercial bank money — while enabling commercial bank money to participate in on-chain settlement. It is the approach closest to how tokenized deposits already work in the GBTD initiative: the tokenized deposit is a commercial bank liability, backed by the bank's reserves at the central bank.

The limitation is that Model 3 does not bring central bank money itself on-chain. Settlement finality — the moment at which a transaction cannot be reversed — occurs at the central bank level when the reserve backing is confirmed, not at the blockchain level when the token is transferred. For high-value institutional transactions where settlement finality is legally and commercially critical, this distinction matters.

The ECB's Current Position

Pontes is Model 2 in operation. Schnabel's speech did not foreclose Models 1 or 3 — it presented all three as live architectural options the ECB continues to evaluate. The progression from Model 2 to Model 1 — from bridge to native — is the medium-term direction. Pontes is described as offering "gradual integration of smart contract functionality and 24/7 operations" over time, suggesting the bridge model is a bridge to the native model, not a permanent endpoint.

The US parallel: the DTCC's October launch is closer to Model 2 — a bridge between existing book-entry systems and distributed ledger infrastructure — than to Model 1. The Fed's GENIUS Act implementing rules for stablecoins are closer to Model 3. No G7 central bank has yet committed to Model 1 for wholesale settlement. That commitment, when it comes, will be the defining moment of the programmable money transition.

→ GBTD — Model 3 in practice: tokenized deposits backed by bank reserves
→ DTCC October — the US equivalent of Model 2
→ GENIUS Act — the US stablecoin rules that approximate Model 3