Will AI Agents Create A Bank Run?

Source: Anthony Pompliano; Pomp Letter (Substack) · 2026-09-28

Torsten Slok, Apollo’s Chief Economist, sent shockwaves through the finance world yesterday when he wrote two simple paragraphs about a potential bank run driven by AI agents. Here is what he said:

“Muse and similar agentic AI assistants could soon sweep household cash automatically into accounts paying 3.3% to 5.0%, instead of the 0.1% national average on checking accounts.If every household used AI agents to optimize the return on their cash balances, banks could lose a large share of the cheap deposits they rely on to make loans, which would be a problem for the entire financial system.”

As you can imagine, the doomer crowd took these two paragraphs and had a field day. These people were so worried that you would have though we were back in the summer of 2008 and the entire financial system was about to collapse. Catastrophe everywhere!

But let’s stop the nonsense. AI agents are not going to take down the financial system. Let me explain why.

First, Torsten’s analysis assumes the world is static. He is arguing that every person will have an AI agent and those agents will collectively decide to move everyone’s money into higher yield products overnight. This vision is essentially Silicon Valley Bank on steroids. That is not how the world works though.

Consumers adopt new technology in waves. It has been 3 or 4 years since AI chat products hit the mainstream. I still know plenty of people who don’t use them on a daily basis. That doesn’t mean the products aren’t good or valuable, but it does mean that early adopters use products....early.

The other component of a static world is that banks with low yield offerings would have no response to this competitive threat. We know finance is one of the most competitive industries in the world, so it is much more likely that banks will either raise their interest rates or find another way to benefit from the rise of AI agents.

For example, where do you think a lot of the money flowing into these fintech products is going? Some of it ends up in traditional banks. Robinhood’s cash management feature is powered by JPMorgan Chase, Goldman Sachs, and Wells Fargo. Chime uses Bancorp Bank. Cash App from Block uses Wells Fargo and Sutton Bank.

The fear-mongering about a bank run will always sound interesting, but the odds are nearly 0%. Don’t take my word for it though. Cullen Roche had a great explanation that I agree with:

“In today's version of "Doomerism Debunked" we're looking at the idea of an agentic bank run and why it's overstated at best and wrong at worst. The argument goes like this. AI agents will sweep household cash out of 0.1% checking accounts and into 4% high yield savings accounts, banks will lose the cheap deposits they use to make loans, and the whole system gets into trouble.But think about where that money actually goes. When cash moves from a checking account at Bank A to a savings account at Bank B, Bank A's deposits fall and Bank B's deposits rise by the same amount. Total deposits in the banking system don't change. And most of the fintechs in the chart are either banks themselves or sweep your cash into partner banks. The money never leaves the system. It just changes addresses.It's also worth remembering that banks don't lend out deposits. Loans create deposits. Banks need deposits to settle payments and hold reserves, and if they lose cheap ones they can replace them with more expensive ones. That's a real cost, but it's a cost, not a collapse.Deposits really leave the banking system mainly through a few channels, such as physical cash withdrawals, payments to the Treasury (which comes back when spent), or money that ends up at the Fed's reverse repo facility. A bot moving your savings to a better rate isn't one of them.Now, the stronger version of this argument deserves a fair hearing. SVB was a run from one bank to other banks, and it still killed the bank. If AI agents can move uninsured deposits in hours instead of days, individual bank runs could happen faster than they used to. And sticky, low rate deposits have always been part of how banks manage interest rate risk. If that stickiness goes away, weaker banks have to manage their balance sheets more carefully.That's a legitimate risk for specific banks with bad balance sheets. It is not a system-wide run. The more likely outcome is that banks have to pay savers closer to market rates and their net interest margins get squeezed. That's bad for bank shareholders. It's pretty good for everyone holding cash.So yes, agentic banking will probably change the economics of deposits. But calling it a bank run confuses a repricing of bank funding with a panic driven collapse of the system. Those are very different animals.”

Said a different way, Haseeb Qureshi writes “business models built on friction and human laziness will (rightly!) get slaughtered in the coming years.” I could not agree with that more.

Ok, so what do I believe will happen as AI agents become more pervasive in the economy? There are 7 major takeaways that I think are worth sharing:

Expect banks to offer higher yields - banks won’t let depositors walk out the door, so they will respond to competition. Historically, the higher yields products haven’t taken away enough deposits to warrant a change. If that changes, the banks will change.

Expect more M&A - some banks will build products internally. Many banks will just acquire the fintechs that have already innovated on the product and captured significant users and deposits. Why risk building internally if you can simply spend money to ensure success? Banks aren’t dumb or in the game of taking high risk.

Expect more bank accounts, not less - just as people need bank accounts, agents will need bank accounts too. The crypto industry is right that agents will use crypto, but they will also use the traditional banking system. Not because traditional banking is superior, but because majority of the world is still on the old rails.