AI, stablecoins and programmable trust are redefining fintech

Alex Lazarow is an author, speaker, global venture capitalist, and contributor to The FR.

As 2025 draws to a close, I’ve been reflecting on themes raised in dozens of conversations I’ve had with industry leaders, in boardrooms with founders and operators, and on the margins of big events. They paint a picture of a fintech landscape undergoing a more fundamental shift than the usual cycle of new products or distribution models.

It’s also been a year marked by meaningful liquidity events, from IPOs to M&A, and the mood feels decidedly optimistic. That momentum was reflected in the year’s notable IPOs, including Chime (a historical portfolio company) and eToro. Meanwhile, some of last year’s major themes, including embedded finance, BNPL and climate finance, were important but felt less dominant.

Across my conversations, a few themes consistently stood out, including:

• AI reinventing financial services from first principles; 

• stablecoins emerging as the Internet’s native settlement layer; and 

• the slow fusion of regulation and technology into programmable trust.

In short, the groundwork for fintech’s next era is already being laid out. Let’s dive into the key themes.

1. AI Is reinventing financial services

AI’s role in financial services is shifting from experimentation to real deployment, with major platforms now pushing capabilities directly into core workflows. A few of the largest AI platforms made recent announcements, including of note Anthropic with an excel plug-in with multiple financial skills. OpenAI announced integrations into multiple commercial and payments apps.

I sat down with Mike Krieger, co-founder of Instagram and now chief product officer of Anthropic, to discuss his views about the role of AI in financial services. 

“I think the companies that are going to grow the fastest are the ones building at the edge of model capabilities,” he said. “They're building in a way where every model release makes their product better — they're letting the model do as much as it can, and it's only going to get more capable and add more value.”

Anthropic’s approach to product development is built around that idea. Krieger pointed to customers like Parcha and Decagon, which stripped out custom code and rebuilt their workflows directly on Claude’s agent SDKs.

What struck me most was his pragmatism about interfaces. 

“At first I thought chat was too primitive,” he admitted. “Now I think it’s genius. You tell the system what you want, and everything underneath the chat … is what’s changing.” In Anthropic’s world, Claude isn’t a chatbot. It’s a computer, an agentic platform that learns and acts.

If Krieger is exploring AI as a product, Sean Neville, best known as co-founder of Circle, is exploring AI as a participant. His new venture, Catena Labs, is building what he calls a financial operating system for agents.

“In the future, the customers of our products will be AI agents,” Neville told me. “They’ll need to pay, get paid, borrow, lend — everything humans do.”

The concept of fintech for AI is an emerging theme that came up in discussions. Catena, part of a growing class of startups, is developing the plumbing to make that possible: agent identity, programmable trust, and what he half-jokingly calls Know Your Agent (KYA).

 “The immediate issue isn’t privacy or price,” he said. “It’s trust. Why would I trust an AI with my money?”

Investors see the same wave building.

Vivek Krishnamurthy of Commerce Ventures said there’s no stopping the train when it comes to human-directed agentic commerce. “Discovery and purchase through agentic platforms — ChatGPT, Perplexity — will be mainstream within twelve months,” he said. 

But he cautions we still have a long way to go on agent-to-agent payments. 

“The ecosystem needs common protocols, identity, and liability standards. When we get there, we’ll disrupt the legacy card and processor value chain,” he added.

All of this points to a broader reset in how financial products are being built.

Karim Gillani, founder of Luge Capital noted that while the last decade of fintech was about unbundling and rebundling banking products, the next phase will “reimagine how those products work fundamentally.”

AI as a building block, and as a customer, will likely be an important component. The next financial products won’t just empower humans with AI — they’ll empower machines to transact safely on our behalf.

2. Stablecoins: from curiosity to global rail

As I’ve written before, stablecoins have key advantages in speed, cost and reliability. “They’re fast, cheap, global, and always on,” said Gillani. “No other payment rail has all those characteristics.”

The industry is maturing. “People don’t want Tether or USDC,” Neville said. “They want dollars on their phone.” 

Circle’s founding thesis in 2013 was simple: put money on Internet rails. This vision has quietly materialized. Stablecoins now power payroll, cross-border remittances and corporate treasury operations from Lagos to São Paulo.  

The next battle is interoperability.

 “If I have $100 million, I don’t want 50 Amazon Coins and 10 USDC and 40 USDT, I need that complexity abstracted,” Neville said. He predicts a payments-optimized blockchain — multi-stable, gas-free, and near-instant — that becomes the TCP/IP of money.

Today, there are few barriers to entry. Jake Gibson of Better Tomorrow Ventures expects consolidation before that changes.

 “There are dozens of issuers and on-ramps being funded,” he said. “But payment rails always consolidate. In the end a few large networks will carry the volume.”

This consolidation also has a geopolitical dimension. Neville pointed out that the U.S. and China are effectively racing to define the next settlement standard — private stablecoins versus state-issued CBDCs. 

“The Internet was built through private-sector innovation and public-sector policy,” he said. “That’s how we’ll build the next generation of money.”

Many of these trends are unfolding at the same time. Jay Ganatra of Infinity Ventures described the convergence with AI this way: “AI and stablecoins are finally finding each other. Stablecoins were a solution in search of a problem — AI may be that problem,” he said.

3. Regulation × technology = programmable trust

The third current I’m seeing is subtler, but just as consequential over the long term: the merger of compliance and code.

Some of the most interesting conversations weren’t about delivering financial services, but about ensuring compliance.

Vivek predicts a shift in how fintechs position their products to banks. 

“AI lets fintechs sell banks Systems of Action instead of forcing them to replace their Systems of Record. That means they can finally capture SaaS and even payroll budgets — ten times the market we’ve been selling into,” he noted.

This may not stay the exclusive domain of fintech. Jeff Green, president of Hatch Bank, told me he expects to see more direct linkages between highly regulated institutions like banks and the capital providers and ecosystem players that fintech platforms once abstracted.

Where the next products will emerge

Looking forward, I expect some of the most exciting fintech innovations won’t be another credit product or an embeddable savings tool. AI will enable entirely new product categories — something I’m focused on professionally at Fluent Ventures — and open the door for a wider set of players. AI is also moving ideas into new markets far more quickly, radically accelerating the globalization of ideas.

Vivek sees this era as fintech’s second act. “The first 15 years were about digital distribution of existing solutions,” he told me. “The next 15 will fundamentally improve the core value proposition.”

Let’s hope so.

A version of this piece appeared in Forbes.