On March 9, 2023, $42B left Silicon Valley Bank (SVB) in the space of a 10-hour business day.
Depositors didn’t wait in an orderly queue outside a branch like they might have done in the 1930s, limited to whatever cash was physically in the vault and whatever hours the doors were open. Instead, SVB depositors opened an app, anxiously jabbed their screens a few times, and moved their money before anyone else could. The bank opened on March 9 and was closed by regulators the next morning.
The point is how quickly it all unraveled, and how fast depositors pulled their money. For most of the 20th century, a bank account was really the only viable choice. It was the only place technology and regulation let ordinary people park spare cash, and FDIC insurance made putting it there feel safe enough to stop the runs that had become common.
None of those three conditions holds the same way anymore. Most importantly, the technology constraint is gone. Once money can move as fast as it did out of SVB, the question stops being where deposits are safest and starts being where they can best be put to use.
It’s a significant market, too. The American Bankers Association itself, citing an April 2025 Treasury estimate, puts $6.6 trillion in transactional deposits as the pool theoretically exposed to this checking account exodus.
If not checking accounts, then where?
I think the answer, for most people, moves away from the bank account, and mostly away from checking accounts specifically. Most people leave spare cash sitting in a checking account paying a fraction of a percent, not because it's the best place for it, but because of a lack of convenient alternatives. That’s the shift I believe we’re going to see in the coming years, and it’ll trigger a retail banking exodus to tokenized credit products.
Bank accounts will still hold a small operating balance, but most will be put to work somewhere with an actual mandate attached: A defined term, a defined yield, a defined level of risk. It’s not a new concept, it’s what corporate treasurers have done for decades, holding idle cash in T-bills instead of a checking account. Someone chooses a 6-month position because they don't need that money for 6 months, and it earns more there than it does sitting idle. But retail customers don’t have a private, savvy treasurer managing their balances.
Imagine if they did.
Advent of the agents
That's what I believe is coming next, except the treasurer isn't a person on a salary, it's an agent that never sleeps and charges a fraction of a management fee.
Instead of a person logging into a banking app once a week to check a balance, an agent holds the keys and moves funds continuously between products with explicit mandates: A 6-month position here because that portion of savings won't be touched for 6 months, a shorter-duration slice there for money that might be needed sooner. This will all be wrapped up in a nice fintech UX while a far more exciting back end, hidden from the customer, takes off.
Robo-advisors? I hear you ask. Robo-advisors have been doing mandate-driven, continuous rebalancing for years. It’s the rails underneath that have changed. A bank account requires a bespoke integration and a human in the loop for anything above a small threshold, which means an agent moving between five providers needs five separate permissioned relationships, each running on its own business hours.
Onchain infrastructure is one standard rail that any agent can transact on the same way, regardless of which product it's moving into, and it doesn't stop for a weekend. A spending limit or an approved counterparty can be a rule the wallet enforces itself. Anyone can check every position, every move, every balance in real time.
Fintechs are already shifting onchain
A retail customer doesn’t need to understand the onchain plumbing underneath, and the products already live are starting to prove that. Robinhood Earn puts Maple credit underwriting behind an app 30 million people already use. Bitso, with more than 9M users across Mexico, Argentina, Brazil, and Colombia, added stablecoin yield next to its core banking features. Littio built a Colombian neobank around the same idea. Plasma One launched in the UK doing the same thing from the ground up rather than bolting it onto an existing product. None are selling "onchain." They're selling a rate and a term instead of a flat balance that never moves.
The strongest signal we’re seeing a shift is that all of these companies spotted the same opportunity and arrived at the same structure, in different markets, serving different customers, independently of each other. It means the shape of the product isn't one company's bet, it's what the category converges on once the constraint that used to force everyone into a checking account stops applying.
On one side, onchain credit providers like Maple have spent years building the tokenized credit products that generate this kind of yield. On the other side, fintechs have built the distribution to put it in front of retail customers. Between them, supply has been sitting ready, waiting for retail demand to catch up.
Banks already see this coming, and they're not being quiet about it
Banks aren't waiting to see whether deposits actually move out of checking accounts and into products like these before reacting. A New York Fed staff paper published in February found that banks holding stablecoin-related deposits end up operating more narrowly, sitting on larger reserves and lending a smaller share of what they hold.
Fed researchers have separately modeled a scenario where a stablecoin-driven deposit shift could pull $1T out of the banking system if that liquidity doesn't recycle back through it. The Fed hasn't seen that scale of outflow yet, but in my opinion, this is the regulatory establishment independently confirming the direction this is heading, before the scale has caught up with it.
The economics explain why banks should be worried. Most currently pay close to nothing on deposits, somewhere around 10 basis points, and keep the rest of the spread for themselves. No one competed on that rate, because a checking account wasn’t really a product with terms, it was just the default.
That only works as long as customers have nowhere else to put idle cash. Once a real alternative exists, banks have two options, and both cost them. Raise what they pay to hold onto deposits, and the spread compresses. Hold the rate where it is, and the deposits leave anyway. No version of this leaves bank earnings intact.
Not risk-free, just held differently
The honest objection is that none of this makes risk disappear, it just moves it somewhere else. Community banks are the most exposed piece of this story. The reserves behind stablecoins concentrate in a handful of large banks rather than spreading out the way retail deposits do today, so a smaller bank could lose deposits to this shift without gaining anything back.
The people building agent wallets aren't pretending otherwise. Current guidance across that ecosystem is that agent wallets are built for operating funds rather than core savings. Self-custody or qualified custody remains the recommended route for the bulk of what someone holds.
What an agent actually needs from a product like Maple is a set of clear rules it can work within:
- Defined strategies an agent can choose between
- A structure simple enough for an agent to read and act on
- Positions and balances checkable in real time
- A real choice on the mix itself, trading duration and risk for a better rate, rather than accepting whatever a bank happened to offer that week
Over the next five years: Banks will narrow
None of this ends with banks gone. It ends with them performing a smaller, more specific job. Banking needs to evolve toward tokenized credit products: A defined term, a defined yield, a defined level of risk. Instead of one flat rate on a checking account, your agent picks a position that matches what you actually need, 6 months here because that money isn't needed sooner, 12 months there because it earns more and you can afford to wait.
That's going to be the natural evolution of banking. And it just happens to be getting built onchain.
Sidney Powell
CEO and Co-Founder
Sidney Powell is the co-founder and CEO at Maple. Sid serves as the company's voice in onchain asset management, yield, and institutional lending.
