Tokenized bank deposits reached $4.8 billion at end of 2025. Twenty-eight financial institutions and central banks have completed transaction scenarios. The market remains small relative to the $303 billion stablecoin sector, but institutional momentum is accelerating — SWIFT, JPMorgan, DBS, and ...
"Tokenized deposits plus AI agents could strip $700 billion in duration risk appetite from U.S. banks by killing deposit stickiness." — Rosie Levy and Srini Ramaswamy, Dallas Fed researchers, August 25, 2026
Tokenized bank deposits reached $4.8 billion at end of 2025. Twenty-eight financial institutions and central banks have completed transaction scenarios. The market remains small relative to the $303 billion stablecoin sector, but institutional momentum is accelerating — SWIFT, JPMorgan, DBS, and the Bank for International Settlements are all running live or pilot programs.
On September 10, 2026, Canada's Office of the Superintendent of Financial Institutions ruled that tokenized deposits are legally identical to traditional bank deposits, adopting a technology-neutral approach that requires no new legal category. Meanwhile, the FDIC in April 2026 proposed separate rules under the GENIUS Act, drawing distinctions between "deposit tokens" and "tokenized deposits" and creating regulatory categories that do not exist in Canada. The Dallas Fed warned in August 2026 that programmable deposits interfacing with AI agents could strip $700 billion in duration risk capacity from U.S. banks.
Three regulatory approaches are now competing globally: deposit tokens as existing bank liabilities (Canada, OSFI), deposit tokens as distinct instruments requiring new rules (U.S., FDIC), and deposit tokens settled via wholesale CBDC (BIS, Singapore MAS). The outcome determines whether banks keep their deposit franchise or lose it to stablecoin issuers.
Table of Contents
On September 10, 2026, OSFI published a statement confirming that tokenized deposits are not a separate class of liability from ordinary bank deposits. The ruling adopts a "technology-neutral" approach: the underlying technology used to build or deliver a product does not change its legal classification.
The practical implications are straightforward. Banks issuing tokenized deposits must still comply with Guideline B-13 (technology and cyber risk) and Guideline B-10 (third-party risk management). Institutions must speak with their OSFI lead supervisor before launching products. Under the Basel framework, qualifying tokenized deposits fall into Group 1a for capital treatment purposes.
The ruling arrives in the context of active Canadian experimentation. In March 2026, the Bank of Canada, RBC, TD Bank, and Export Development Canada completed Project Samara — a C$100 million tokenized bond issuance pilot on distributed ledger. BMO, with CME Group and Google Cloud, announced a roadmap for tokenized cash and deposit capabilities.
No Canadian bank has publicly announced a tokenized deposit product launch yet. But the regulatory clarity now exists for one to proceed. Canada's approach stands in direct contrast to the U.S., where regulators are constructing new categories rather than fitting tokenized deposits into existing ones.
On April 7, 2026, the FDIC issued a proposed rulemaking to implement GENIUS Act provisions for FDIC-supervised institutions. The rule draws a distinction between "deposit tokens" and "tokenized deposits," creating separate regulatory treatment for each category.
The insurance treatment of stablecoin reserves is where the distinction bites hardest. Under the proposed rule, deposits held as payment stablecoin reserves would be insured to the issuer — classified as a Permitted Payment Stablecoin Issuer, or PPSI — and not on a pass-through basis to individual stablecoin holders. Insurance coverage applies up to $250,000 with respect to the PPSI, not to each individual stablecoin holder. Tokenized deposits that meet the statutory definition of "deposit" under the Federal Deposit Insurance Act would be treated no differently from conventional deposits.
The comment period closed June 9, 2026. The Bank Policy Institute, The Clearing House, and the Consumer Bankers Association submitted joint comments on the FDIC's GENIUS Act rule, signaling coordinated industry engagement with the rulemaking process.
The contrast with Canada is stark. The U.S. is creating new regulatory categories for instruments that Canada treats as existing liabilities. Whether this additional complexity produces better outcomes or simply slows adoption remains an open question. What is clear is that U.S. banks face a longer path to product launch than their Canadian counterparts.
On August 25, 2026, the Dallas Fed published a paper by researchers Rosie Levy and Srini Ramaswamy that quantified the systemic risk embedded in deposit tokenization.
The core finding: if tokenized deposits make depositors 10% more rate-sensitive, banks lose approximately $700 billion in capacity to hold long-term interest-rate risk. If tokenization causes deposits to leave 10% sooner, banks lose approximately $580 billion in capacity to absorb the interest-rate risk of long-term loans and securities.
The mechanism the paper describes is specific. Programmable deposit tokens with smart contracts that interface with AI agents enable automated, near-instantaneous reallocation to whichever bank pays the highest yield. This kills "deposit stickiness" — the behavioral inertia that allows banks to borrow short and lend long, the foundational business model of commercial banking.
An earlier Dallas Fed paper, published July 14, 2026, explained how tokenized deposits use blockchain architecture inside a traditional banking framework. The August paper builds on that foundation to model the consequences of what happens when that architecture makes depositor behavior faster and more rational.
The Dallas Fed's findings align with warnings from the International Monetary Fund. IMF Note 2026/001, authored by Tobias Adrian and published in April 2026, warned that "market stress episodes expected to develop more rapidly than in conventional systems, narrowing the window for authorities to intervene." The combination of programmable deposits and AI-driven allocation compresses the timeline for deposit flight from days to seconds.
Twenty-eight financial institutions and central banks have completed transaction scenarios for tokenized deposits, and the infrastructure buildout is accelerating across three distinct models.
Model 1: Tokenized deposits on private/permissioned ledgers. SWIFT launched a blockchain-based shared ledger with 17 banks from six continents piloting live cross-border payment transactions using tokenized bank deposits. JPMorgan and DBS are exploring an interoperability framework for "exchangeability and settlement of tokenized deposits" between JPMorgan's Kinexys Digital Payments and DBS Token Services.
Model 2: Bank-issued stablecoins backed 1:1. Twenty-one banks — including Citi, Goldman Sachs, and UBS — formed a global bank stablecoin consortium targeting a USD token launch in H1 2027. This approach creates a stablecoin instrument issued by banks rather than by non-bank entities such as Circle or Tether.
Model 3: Tokenized deposits settled via wholesale CBDC. Singapore's MAS announced a 2026 pilot for tokenized government bills settled using wholesale CBDC, following a 2025 trial with DBS, JPMorgan, and Standard Chartered. The BIS has concluded that tokenized deposits settled in wholesale CBDC are "far better suited to serve as the private money component" of future financial infrastructure.
These three models are not mutually exclusive in theory, but they imply different regulatory frameworks, different capital treatment, and different competitive dynamics between banks and non-bank issuers.
The distinction between tokenized deposits and stablecoins is not semantic. It is structural.
GENIUS-compliant stablecoins are backed 1:1 by low-risk liquid assets, primarily Treasuries. They are not bank deposits. Deposit insurance does not apply to holders. No fractional reserve lending occurs against them.
Tokenized deposits are a claim on the issuing bank, backed by the bank's full balance sheet. Deposit insurance applies. Banks can lend against them through fractional reserve banking, the same mechanism that applies to conventional deposits.
This is the core tension. Stablecoins disintermediate banks from the deposit-taking function. Tokenized deposits keep banks in the intermediation chain. As Fnality International noted, "Global regulation of stablecoins and tokenized deposits remains fragmented in 2026."
The $303 billion stablecoin market represents capital that sits outside the banking system's fractional reserve mechanism. Every dollar held in a stablecoin backed by Treasuries is a dollar that a bank cannot lend against at a leverage ratio. The $4.8 billion tokenized deposit market represents capital that remains inside the banking system.
If tokenized deposits win the adoption race, banks preserve their lending franchise. If stablecoins win, a parallel narrow-banking system emerges — one where issuers hold full reserves and no lending occurs against the liabilities. The regulatory choices being made in 2026 by OSFI, the FDIC, the BIS, and MAS are shaping which outcome prevails.
Canada adopted a technology-neutral approach. Tokenized deposits equal bank deposits, full stop. No new legal category required. OSFI's September 10 ruling gives Canadian banks a clear path to launch.
The U.S. is creating separate regulatory categories for deposit tokens and tokenized deposits, adding complexity and uncertainty. The FDIC's proposed rule treats stablecoin reserve deposits differently from pass-through deposit insurance, a distinction with material consequences for stablecoin holders.
The Dallas Fed quantified systemic risk. $700 billion in bank duration capacity is at stake if deposits become programmable and rate-sensitive. $580 billion at stake if deposits simply leave faster. AI agents compound the effect.
28 institutions across six continents are in live or pilot stages. SWIFT (17 banks, cross-border payments), JPMorgan and DBS (interoperability framework), Singapore MAS (wholesale CBDC settlement), and a 21-bank consortium (USD stablecoin, H1 2027 target) are all active.
The deposit-vs-stablecoin question is structural. It determines whether banks keep their fractional reserve franchise or cede it to non-bank issuers operating under narrow-banking constraints.
The regulatory divergence between Canada's "same rules" approach and the U.S. "new rules" approach will determine how fast tokenized deposits scale. Canada's clarity gives its banks a head start. The U.S. regulatory split between the FDIC, OCC, and the Fed creates uncertainty that the comment-and-rulemaking process may not resolve quickly. Banks operating across both jurisdictions face the additional burden of reconciling two different regulatory philosophies applied to functionally identical instruments.
The $4.8 billion tokenized deposit market is small relative to the $303 billion stablecoin market. But the institutional weight behind deposit tokens — 28 institutions, SWIFT, the BIS, and now explicit Canadian regulatory backing — suggests the two instruments are on a collision course. The Dallas Fed's $700 billion warning is a reminder that the stakes extend beyond fintech competition into the structural stability of the banking system itself.