Fintech Investors Left Money20/20 Debating The Future Of Fintech

This piece was first published in my Forbes piece here.

I am just back from Money20/20. As I do every year, I like reflecting on the key themes coming out of the event.

Perhaps the question of this event was: what’s fintech anymore anyways?

Ryan Falvey, founder of Restive, saw two camps at the show: “those at the conference that I think are living in the fintech world of 2015 through 2021, and those that are really embracing the impact AI is going to have on financial services.”

He went on: “I was struck by the combination of melancholy, delusion and lack of energy by some at the conference who seemed to think the AI bubble ‘was going to pop’ and we were going to go back to how it was.” He added, “AiFi (AI-native Financial Services) has really changed everything.”

This post will look to answer a few fundamental questions. Let’s start with the biggest.

Is Fintech Back? And If So, What Is It Anymore Anyways?

Jay Ganatra of Infinity Ventures, reflecting on whether fintech is back, answered “The sector was left for dead for a few years and it feels like it’s coming back with a vengeance.” A lot of what is driving that is AI. Jay pointed out that “we’re now starting to see companies like Stripe and Ramp move squarely into the AI infra layer.”

Vikas Raj, co-founder and managing partner of ResilienceVC, was upbeat about what he saw. “The most interesting companies we saw are building around AI and run at a fraction of the operating cost, which lets them profitably serve customers that legacy providers couldn’t reach, with products that adapt rapidly to each user.”

Jake Gibson of Better Tomorrow Ventures asked cautiously, reflecting on the impact of AI in venture: “Do I think that the technology landscape has permanently changed, and huge durable businesses are going to be built? Yes. Do I think the market is going to over-extrapolate and do the same thing it always does? Also yes.” He is worried about incentives. “There’s more money than ever.” But it’s harder than ever. “It’s just going to fewer places…the safest thing you can do for your career is increasingly the riskiest thing you can do with your LPs’ money.” Reflecting that it is harder for great fintech businesses that may not be growing at AI levels, to raise capital.

My take: fintech is turning into a part of the AI economy (Ryan’s AiFi), which is good for the sector. This will reward companies that can build capital-efficiently if they are not growing at hyper-AI growth levels.

Is The Exit Window Open?

Aaron Fu, head of venture investments at DCG, told me: “The exit window is certainly open and we expect the momentum of M&A to continue through in the coming months.” He also argued acquihires have accelerated, with acquirers looking to bolt on key capabilities.

Dan Rosen of Commerce Ventures was less sanguine, when asked if the exit window had cracked open: “Cracked. The sizing of M&A has been varied and the frequency intermittent.”

My take: both are actually describing the same market from different vantage points. Infrastructure that fills gaps in others’ stacks is still being acquired. But at the same time, IPOs remain lumpy. In the last two years, I saw the exit of one of my portfolio companies, Chime, and recently wrote about the wave of M&A in stablecoins, one of our global themes.

Ok, AI Pays, But Who Is On The Hook?

Just before the conference, JPMorgan said it would block outside AI agents from client accounts. Citi and Bank of America followed shortly after. At the same time, Robinhood shipped an agentic card (which I gave to my AI agent, Instinct).

So who is responsible in an agentic banking era? Dan said (questioning the definition of agentic banking) “Assuming it is a card-based payment and the merchants follow network rules and protocols, the liable party is the party that provides the mandate to the agent either when creating it or authorizing it to transact (typically, its developer or operator).”

But there is clearly much infrastructure to build. Karim Gillani of Luge Capital called this a liability gap: “The merchant, the issuer, the agent developer and the consumer all assume someone else is on the hook. Whoever closes that gap with a clear, insurable framework will own a very valuable piece of infrastructure.”

Monica Brand Engel, co-founder and managing partner of Quona Capital, stressed the value that incumbent infrastructure has built over time, and what needs to be replaced in its absence: “Cards let you fix mistakes after the fact through chargebacks. On real-time rails like Pix, UPI and FedNow, the money is gone in seconds. Trust has to be verified before the money moves, not after.” Monica thinks the card networks “will build their own version, but the bigger opportunity is a neutral trust layer that works across every rail, from cards to instant payments to stablecoins.”

New fronts will also be open. Aaron argues agents need to “settle instantly, often in amounts too small for card economics,” which is why “programmable money, stablecoins especially, becomes the native rail for machine-to-machine commerce.”

My take: As I wrote earlier this year, every technology wave builds a trust layer. A lot of trust infrastructure was built for the digital bank and card world. New infrastructure will need to be built for the AI world. This seems to be a rising area of focus and agreement. As Jay put it, “I think liability and identity are actually the same problem. When an agent transacts autonomously, the question of ‘who is responsible’ can’t be answered without first answering ‘who is this agent, really?’” Karim was more concrete. “Today, most agents borrow a human’s credentials, so merchants can’t tell a legitimate agent from a fraudulent bot.”

Is A Stablecoin Turning Into A Narrow Bank?

One of my questions on the future of stablecoins is whether they are morphing into new narrow banks. I was surprised most folks said yes.

“A fully reserved stablecoin holds safe assets one-for-one and moves money, but doesn’t lend,” Karim said. “That’s a narrow bank in everything but name. The open question is whether regulators treat it like one.” Aaron agreed: “In substance, a fully reserved stablecoin already is a narrow bank: every dollar backed by cash and Treasuries, no lending.”

The debate was more on what stablecoins’ are actually best used for. Dan sees an enormous challenger opportunity, “globally accessible ‘bank’ accounts for the masses” that store value in a stable currency of choice while also enabling spending flexibility. “Now that we’re seeing debit cards attached to these accounts, the picture is becoming even clearer,” he said.

Monica, conversely, sees it more as plumbing. Issuance and acceptance “are still relevant topics,” she argued, but “the conversation is shifting more towards distribution, integration, and interoperability.”

My take: Dan and Monica are both right, it just depends on your vantage point. The same token can be plumbing in one market and a bank account in another. Aaron put it best: “In the U.S., a stablecoin is a faster settlement layer; in Lagos or Buenos Aires, it’s the most reliable dollar account someone has ever had.”

Is A One-Person Fintech Possible?

I thought Dan hit it on the head here “Possible? Maybe. Likely? Not a U.S. chartered financial institution…at least not yet.” He could see one person building “a personal financial management app or financial services lead gen,” but not a bank. Aaron thinks AI “has collapsed the cost of shipping software” but “hasn’t collapsed licensing, compliance, banking partners or distribution.”

My take: I’m actually skeptical on the notion of single-founder unicorns in general (though there will be plenty of single-founder startups). But in fintech, I believe it is harder for complex, regulated business models.

Parting Thoughts:

Perhaps symbolically, Money20/20 announced it was moving its venue.

Similarly, the fintech conversation, and particularly what’s happening at the frontier with stablecoins and agentic finance, is moving.

Can’t wait to see what 2027 has in store for us.