The BIS Gambit: Tokenized Deposits and the Architecture of Trust in a Trustless System

The Bank for International Settlements has a stablecoin problem. Not the kind that surfaces in a code audit, but the kind that surfaces in a policy speech. At this year's Jackson Hole symposium, BIS General Manager Pablo Hernandez de Cos delivered what amounts to a technical indictment of the stablecoin industry, wrapped in the language of monetary policy. His thesis: tokenized deposits, not stablecoins, are the future of digital money. The reasoning is familiar to anyone who has examined payment infrastructure: stablecoins lack true interoperability, AML controls are inconsistent across jurisdictions, and the reserve-backed model introduces trust assumptions that central banks find unpalatable. But beneath the policy language lies a structural argument that deserves forensic attention. This is not market commentary. It is a claim about the architecture of settlement itself. And it is a claim that the market has not fully priced in. Tokenized deposits are exactly what they sound like: commercial bank deposit liabilities represented as tokens on a distributed ledger. They are not a new asset class. They are the existing two-tier monetary system - central bank money at the top, commercial bank money below - rendered in programmable form. The settlement layer, in the BIS vision, would be wholesale central bank digital currency, with commercial bank tokens settling against it on a unified platform. This is the Agora project, which BIS has been advancing with a consortium of central banks since 2024. Stablecoins, by contrast, are independent networks. USDT and USDC hold reserves - short-term Treasuries, cash, commercial paper - and issue tokens against those reserves. The tokens trade on permissionless or semi-permissioned networks, and settlement happens on the token's native ledger, not on a central bank's books. This is the fundamental structural difference. Tokenized deposits are bank liabilities. Stablecoins are issuer liabilities backed by a reserve pool. The trust anchor is different, and that difference has consequences. De Cos's argument, distilled: stablecoins create a parallel settlement system outside the banking framework, they fragment liquidity, they complicate AML enforcement, and they threaten monetary sovereignty. Tokenized deposits, he argues, keep settlement inside the regulated banking system while capturing the efficiency gains of distributed ledgers. The venue matters. Jackson Hole is not a casual conference. It is where the world's central bankers coordinate policy signals. When the BIS General Manager speaks there about tokenized deposits, it is not an academic exercise. It is a directional signal to 63 central banks, to the commercial banking sector, and to the stablecoin industry. The signal is clear: the official sector is building an alternative. The competitive landscape is stark. USDT holds roughly 62 percent of the stablecoin market with about $140 billion in circulation. USDC follows at around 25 percent with $80 billion. Tokenized deposits, by contrast, are still in pilot phase. The Agora project has not published production metrics. No major economy has deployed tokenized deposits at scale. The gap between the two systems is not measured in percentage points; it is measured in orders of magnitude. Let me examine the technical claims in order, because they deserve more rigor than the policy framing provides. The Interoperability Claim De Cos argues that stablecoin platforms lack true interoperability. This is partially correct, but the framing is misleading. Stablecoins have achieved interoperability through bridges, centralized exchanges, and payment processors. USDT moves across Ethereum, Tron, Solana, and a dozen other chains. The settlement is not elegant - it relies on wrapped assets, bridge contracts, and custodial intermediaries - but it functions at scale. USDT alone processes billions of dollars daily across multiple networks. Tokenized deposits, by contrast, are being designed as closed or semi-closed systems. The Agora project envisions commercial banks issuing tokens on a shared platform, with central bank money as the settlement asset. This is interoperable within the consortium, but it is not interoperable with anything outside it. A tokenized deposit from a German bank does not spontaneously become usable on a Japanese bank's platform. The interoperability is designed, not emergent. The BIS argument is that stablecoin interoperability is fragile because it depends on bridges and custodians, which are attack surfaces. This is a legitimate security concern. Bridge hacks have cost the industry billions. But the solution is not necessarily tokenized deposits. It is better bridge architecture, or native multi-chain settlement. The BIS position conflates interoperability that exists but is imperfect with no interoperability at all. There is also a deeper issue. The BIS vision of interoperability is centralized by design. All settlement flows through the central bank's ledger. This is efficient for oversight, but it reintroduces a single point of failure - not in the technical sense of a server crash, but in the governance sense of a single authority controlling the settlement layer. The architecture of trust in a trustless system is, in this case, the architecture of a central bank. The AML Argument De Cos's second claim: AML controls are difficult to implement consistently on stablecoin networks. This is true, but it is a design choice, not a technical limitation. Permissionless networks do not have KYC at the protocol layer. Compliance happens at the exchange and custodian level. This creates gaps - a user can move funds between non-compliant venues, or use privacy tools to obscure the trail. Tokenized deposits solve this by construction. The issuing bank knows its customer. The settlement layer is permissioned. Every transaction is visible to the central bank. This is not a technical achievement; it is a governance choice. The same AML outcomes could be achieved with stablecoins if the regulatory framework required it - and indeed, MiCA in Europe is moving in that direction. The BIS argument is really about who controls the compliance layer, not whether compliance is possible. But there is a cost. The permissioned architecture that enables AML also enables surveillance. Every transaction, every balance, every counterparty relationship is visible to the central bank. For institutional users, this is acceptable - they are already subject to bank oversight. For retail users in jurisdictions with weak rule of law, the calculus is different. Stablecoins offer a degree of financial autonomy that tokenized deposits cannot. This is not a bug in the BIS vision; it is a feature. The BIS is explicitly choosing oversight over autonomy. The Trust Anchor The deepest difference is the trust anchor. A tokenized deposit is a claim on a commercial bank, backed by the central bank's settlement guarantee and, in most jurisdictions, deposit insurance. A stablecoin is a claim on a reserve pool - Treasuries, cash, commercial paper - held by an issuer. The reserve pool is audited, but the audit is not a guarantee. In a crisis, the reserve pool could be frozen, mismanaged, or subject to runs. This is where my own audit experience comes in. I have examined reserve attestations from major stablecoin issuers. The attestations are real, but they are point-in-time snapshots. They do not prove that the reserves are liquid in a stress scenario. They do not prove that the issuer can honor redemptions at par during a market-wide panic. The banking system has deposit insurance, lender-of-last-resort support, and a century of crisis management. Stablecoins have none of these. This is the strongest argument for tokenized deposits, and it is not a technical argument. It is a balance-sheet argument. However, the balance-sheet argument cuts both ways. The banking system's stability is a function of central bank support, which is a political choice, not a technical guarantee. In a hyperinflationary environment, or in a jurisdiction with an unstable banking sector, the central bank's balance sheet is not a safe harbor. Stablecoins denominated in dollars or euros offer an escape hatch. The BIS vision of tokenized deposits assumes that the domestic banking system is sound. That assumption does not hold everywhere. The Economics of the Two Systems There is a tokenomics dimension that the policy debate ignores. Stablecoins have a revenue model: the issuer earns yield on the reserve pool. Tether and Circle generate billions in interest income from their Treasury holdings. This is the seigniorage of the stablecoin era. It funds the infrastructure, the compliance teams, and the distribution networks. It is also the source of the conflict - the BIS sees this as private money creation, which is precisely what central banks are supposed to control. Tokenized deposits have no such revenue model. The bank earns the spread on the deposit, as it always has. There is no separate token, no reserve pool, no yield arbitrage. The economics are the economics of the banking system, not the economics of a new asset class. This means tokenized deposits will never attract speculative capital. They are not an investment. They are a utility. The market will not price them like a token, because they are not a token in the economic sense. The cost structure also differs. Stablecoin issuers bear the cost of reserve management, attestation audits, and compliance. These costs are passed on to users in the form of fees, or absorbed in the spread. Tokenized deposits, by contrast, are an incremental cost on top of existing banking infrastructure. The marginal cost of issuing a tokenized deposit is lower than the marginal cost of issuing a stablecoin, because the bank already has the compliance infrastructure. But the fixed cost - the IT overhaul, the ledger integration, the legal agreements - is substantially higher. This is why the BIS vision depends on central bank coordination. No single bank can build this alone. This has a practical consequence. The stablecoin market is driven by demand for dollar exposure in jurisdictions where dollars are hard to access. Tokenized deposits cannot serve that demand, because they are domestic liabilities. A tokenized deposit in euros is not a dollar substitute. The BIS vision does not address the fundamental use case that drove stablecoin adoption: access to hard currency in soft-currency environments. The Sovereignty Question The geopolitical dimension is the elephant in the room. US Treasury Secretary Scott Bessent has explicitly endorsed stablecoins as a tool to strengthen dollar hegemony and create demand for US Treasuries. The BIS, representing 63 central banks, sees dollar-denominated stablecoins as a threat to monetary sovereignty. If USDT and USDC become the default settlement rails for cross-border payments, non-US central banks lose control over their domestic payment systems. Capital flows become harder to monitor. Monetary policy transmission becomes weaker. This is why the BIS is pushing tokenized deposits. It is not because tokenized deposits are technically superior in every dimension. It is because they keep settlement inside the banking system, where central banks have authority. The technical arguments are scaffolding for a political conclusion. The regulatory divergence between Washington and Basel is now explicit. The US Treasury sees stablecoins as a strategic asset. The BIS sees them as a strategic threat. This is not a technical disagreement; it is a geopolitical one. The market will have to navigate two competing regulatory frameworks. In the US, stablecoin legislation is advancing - the GENIUS Act and related bills are moving through Congress. In Europe, MiCA has already imposed comprehensive regulation on stablecoin issuers. In Asia, jurisdictions like Hong Kong and Singapore are running sandboxes for both stablecoins and tokenized deposits. The result is a patchwork of rules that will define the competitive landscape for the next decade. The market impact is indirect but real. Stablecoin issuers face a future of regulatory fragmentation. The BIS signal will accelerate central bank experiments with tokenized deposits, which will in turn create a parallel settlement infrastructure. The stablecoin market will not collapse - the network effects are too strong - but its growth ceiling will be defined by regulation, not by technology. The Blind Spots Here is the blind spot. The BIS argument assumes that the banking system is the natural home for digital settlement. But the banking system has spent the last decade demonstrating that it cannot innovate at the speed of the market. Tokenized deposits require IT overhauls, cross-border legal agreements, and central bank coordination. The Agora project is promising, but it is a pilot. The timeline for production deployment is measured in years, not quarters. Meanwhile, stablecoins have already achieved what tokenized deposits are still designing: global settlement, deep liquidity, and network effects. USDT has roughly $140 billion in circulation. USDC has about $80 billion. The infrastructure exists. The users exist. The liquidity exists. The BIS can argue that this infrastructure is fragile, but fragility is not the same as absence. The more likely outcome is bifurcation. Tokenized deposits will dominate institutional settlement - cross-border wholesale payments, interbank clearing, securities settlement. Stablecoins will continue to dominate retail, Web3, and emerging markets. De Cos himself acknowledged this division of labor. The two systems will coexist, but they will not be interoperable. That is the real risk: a fragmented global payment landscape with two parallel settlement rails, each with its own governance, its own compliance regime, and its own geopolitical alignment. There is also a second blind spot: the assumption that central banks will actually deploy tokenized deposits at scale. The history of CBDC projects is a graveyard of pilots that never reached production. The technical challenges are real, but the political challenges are larger. Central banks are conservative institutions. They do not move quickly. The BIS can signal direction, but it cannot force deployment. The stablecoin industry should not panic; it should watch the pilot programs and prepare for a long transition. And there is a third blind spot, perhaps the most important. The BIS is not neutral. It is an institution with a mandate to preserve the existing monetary order. Its preference for tokenized deposits is not a technical judgment; it is an institutional survival instinct. The stablecoin industry should understand this. The debate is not about which technology is better. It is about who controls the settlement layer of the global economy. The BIS has made its move. The question is whether the market follows. Tokenized deposits are not a competitor to stablecoins in the retail sense; they are a competitor in the institutional settlement sense. The next three to five years will determine whether the banking system can actually deliver on the promise of programmable money, or whether the stablecoin networks - for all their flaws - remain the de facto settlement layer for the digital economy. Where logic meets chaos in immutable code, the answer will be written not in policy speeches, but in settlement data. The architecture of trust in a trustless system is being redesigned, and the architects are not the ones who built the first version. The stablecoin industry built the rails. The central banks are building the alternative. Both cannot be the default settlement layer. The market will choose, and the choice will be made in the data - in settlement volumes, in liquidity depth, in the cost of moving money across borders. Watch the pilots. Watch the deployment timelines. The policy speech is the beginning, not the end.