A research paper from the Federal Reserve Bank of Dallas puts a specific figure on a risk regulators have mostly discussed in the abstract: as much as $700 billion could leave US banks' lending capacity if deposits shift into tokenized deposits and stablecoins. CoinDesk reported the estimate on August 26, 2026.
The warning is not that money disappears. It is about where the money sits. Traditional bank deposits are the raw material banks lend against. Move those balances into on-chain formats that sit outside the fractional-reserve system, and the base that funds mortgages, business loans, and credit lines shrinks.
The mechanics behind the number
Banks fund loans with deposits. A dollar parked in a checking account does not stay idle: the bank lends most of it out, keeping a fraction in reserve. That multiplier is how a banking system turns a given pool of deposits into a much larger stack of credit.
Tokenized deposits and fully-reserved stablecoins break that loop. A stablecoin backed one-for-one by cash and short-term Treasuries does not get re-lent by its issuer. The dollar backing it sits in reserve, not in a small-business loan. When a depositor moves $1,000 from a bank account into a stablecoin, the bank loses a funding source and the credit multiplier attached to it. Scale that across the system and the Dallas Fed's $700 billion figure is the shortfall in lending capacity that follows.
The paper frames this as a structural consequence of design, not a run. Even an orderly, gradual migration produces the drain, because the destination products are built to hold reserves rather than recycle them into loans.
Timing against a busier stablecoin backdrop
The estimate arrives while the US stablecoin framework is being built out in real time. The GENIUS Act set reserve and identity rules for stablecoin issuers, and industry groups have been shaping how those rules apply. We covered the Blockchain Association's position on the GENIUS Act identity requirements earlier this month.
Banks are not sitting still either. Thirty-nine US state banking associations formed the BankChain Alliance to build shared blockchain rails, targeting a 2027 launch for tokenized deposits and automated settlement. That project is a direct answer to the disintermediation the Dallas Fed describes: if deposits are going to be tokenized, banks would rather issue the tokens themselves and keep the balances on rails they control.
The distinction the paper leans on matters here. A tokenized deposit issued by a bank can, in principle, still support lending because the balance remains a claim on that bank. A stablecoin issued by a non-bank, fully reserved against Treasuries, does not. The $700 billion risk is heaviest in the second case, where deposits leave the banking system entirely rather than changing form inside it.
The stablecoin funding layer and everyday spending
For anyone using stablecoins for day-to-day spending, the paper is a reminder that these balances are not just a payments convenience. They are a funding-layer decision. Money held as USDC or USDT is money that a bank is not lending against.
Crypto cards sit on top of this layer. Most stablecoin-linked cards top up from balances that already live off-bank, so the shift the Dallas Fed models is, in part, already underway among crypto users. The macro question is what happens when that behavior moves from a niche to the mainstream. If a meaningful share of ordinary checking-account balances follows, the credit contraction the paper describes stops being theoretical.
Crypto prices offered little drama alongside the report. As of August 26, 2026, Bitcoin traded near $78,515, down 0.7% on the day, with Ether around $2,465, down 0.6%, per CoinMarketCap. The Fear and Greed Index sat at 80, in "extreme greed" territory, even as this longer-term structural warning landed.
The policy tension the paper exposes
The report captures a genuine conflict at the center of US crypto policy. Regulators want stablecoins fully reserved so they cannot blow up the way undercollateralized tokens have. Full reserving is also precisely what removes those dollars from the lending economy. The safety feature and the disintermediation risk are the same design choice.
A Federal Reserve regional bank attaching a $700 billion estimate to that tradeoff raises the stakes of the debate. It does not settle it. The figure is a projection of what could happen if deposit migration reaches scale, not a measurement of what has already occurred. But it hands lawmakers and bank supervisors a concrete number to argue over as tokenized-deposit projects like BankChain move toward launch.
Overview
The Dallas Fed estimates up to $700 billion in lost US bank lending capacity if deposits migrate into tokenized deposits and stablecoins, because fully-reserved on-chain products hold their backing in reserve instead of re-lending it. The warning lands as the GENIUS Act framework and the bank-owned BankChain Alliance take shape, sharpening a policy tension: the full reserving that makes stablecoins safe is the same feature that pulls dollars out of the credit system. The number is a projection, not a measured outcome, but it gives the disintermediation debate a hard figure to work with.
