That creates a second-order portfolio question for holders of financial-sector ETFs. Broad financial funds can look diversified, but their underlying banks do not all face this risk equally. Institutions with brokerage, advisory and cash-management ecosystems may be better able to keep assets inside their platforms, while deposit-heavy banks without those links could face more pressure on funding costs and net interest margins. Investors in sector ETFs should therefore look past the headline “financials” label and consider what business models dominate the index.
There is also an important behavioural caveat. Moving emergency cash into higher-yield vehicles is not the same as increasing risk capacity. Automated cash optimization may improve returns on idle balances, but it can also blur the line between liquidity management and investing. Long-term investors should separate true near-term cash needs from portfolio capital, because convenience tools can encourage unnecessary movement between accounts or products. The article points to a real structural risk for banks, but for ETF investors the more durable issue is whether AI changes where cash sits, how sector profits are earned and how much hidden concentration exists inside financial-sector exposure.
Banks make a lot of their money from depositors leaving their cash sitting idle, but those days could be coming to an end, Bank of America says.
In a client note on Wednesday, BofA said agentic AI assistants like Meta’s Muse — which can perform tasks autonomously, including making purchases and dealing with account admin— pose a “multiyear evolutionary risk” to banks’ business models.
The thinking goes that these AI assistants could eventually be able to manage and move money for people, automatically allocating funds to certain investments instead of leaving them parked in low-yield savings or checking accounts.
“The takeaway for banks is the precedent, not the e-commerce use case: scaled platforms are plugging payment credentials into a third-party agent,” Ebrahim H. Poonawala, a research analyst at Bank of America, wrote. “The gap between ‘find me a better product and pay for it’ and ‘find me a better yield and move my excess cash’ is narrowing.”
He added: “Whereas a chatbot can tell customers they are earning too little, an agent can identify excess liquidity, compare yields and act.”
In savings and checking accounts, banks offer interest rates typically well below even 0.1%. Meanwhile, many money market funds yield around 4% annually.
Federal Reserve data shows that around $5.4 trillion dollars is sitting in checking accounts, money that is missing out on higher returns that can compound over the years. According to a study from Bankrate, $10,000 earning 4% per year would earn $2,167 over a five-year period, while one would earn just $5 over that time at a 0.01% interest rate.
Banks losing out on those returns would be a big hit to their margins. Investors have recognized the risks for financials sector stocks. The State Street Financial Sel Sec SPDR ETF (XLF) is down 2.4% since the start of trading on Tuesday, while the Invesco KBW Bank ETF (KBWB) has fallen 3.2%.
Still, agentic AI won’t take hold overnight, giving banks time to react, Bank of America said.
One route they can take is to “self-cannibalize” and institute their own AI assistant to help customers to manage their money. This way, BofA said, the banks can at least keep the money in-house with their own investment products.
“Banks with integrated consumer banking, brokerage and wealth platforms should be better positioned to keep balances within their ecosystems,” Poonawala wrote. “Those without may need to improve their own offerings, partner with third parties, or accept greater funding and margin pressure.”
Enjoyed this article? Sign up for our newsletter to receive regular insights and stay connected.