A recent report from the Federal Reserve Bank of Dallas indicates that AI-driven bank accounts are accelerating the movement of funds between financial institutions, potentially undermining the traditional banking model that relies on low-cost, stable deposits to finance long-term credit. The core of this structural disruption lies in the erosion of the long-term financing value embedded in stable deposits.

Based on commercial bank balance sheet data from July 15, analysis shows that the interest rate risk exposure for related assets is estimated at up to $7 trillion, with approximately $5.84 trillion directly stemming from the duration characteristics of non-large time deposits. Sensitivity analysis further quantifies the risk: if deposit price sensitivity increases by 10%, under the assumption of a 4-year weighted average maturity, overall duration risk tolerance over a 10-year period would shrink by roughly $700 billion. Meanwhile, if the weighted average maturity shortens by 10%, the model estimates that maturity conversion capacity would decline by approximately $580 billion. The 10-year equivalent concept aims to translate risk exposure into interest rate risk comparable to that of 10-year Treasury bonds, though the actual credit impact depends on how banks restructure their assets and funding sources.

Facing declining funding stability, banks are forced to adjust their balance sheets to maintain operations. One path involves issuing more term debt to lock in funding sources, but this constitutes wholesale funding activity, which the Dallas Fed explicitly notes would directly raise borrowing costs for businesses and individuals. Another strategy is to increase holdings of reserves and Treasury securities as a buffer against more rapid and unpredictable outflows, but this inevitably crowds out resources available for originating illiquid credit, leading to a contraction in credit supply.

Although digital deposits remain in their early stages with uncertain ultimate scale, potential systemic implications have already drawn regulatory attention. The report's authors emphasize that these views represent only their personal research conclusions and should not be regarded as the official stance of the Dallas Fed or the Federal Reserve System. However, as AI technology deepens its penetration into the financial sector, the exposure of vulnerabilities in bank funding structures may become a recurring theme.