The Intelligence Stack: AI, Stablecoins, and the Next Financial System

The convergence of intelligence and value is rewriting the rules of banking.

We’re standing at the intersection of two profound transformations in financial services: the commoditization of intelligence through AI and the tokenization of value through blockchain and stablecoins. Each of these forces alone would be disruptive. Together, they’re setting the stage for a fundamental rewrite of financial infrastructure.

Across the globe, AI is moving from pilot programs to production-grade platforms. Whether it’s JPMorgan’s massive in-house LLM initiative or Santander’s full-stack partnership with OpenAI, banks are no longer just testing, they’re transforming.

Here are a few of the story arcs I've been following (find links in the comments).

The J is for Juggernaut: AI at Enterprise Scale

JPMorgan Chase is arguably the most aggressive legacy bank deploying AI at scale. Not tests, but production across all core functions. Recent reports show the bank now runs 400+ AI use cases in production (that's insane, by the way), involving over 200,000 employees using its in‑house “LLM Suite” to drive operational efficiency, fraud mitigation, credit risk, and wealth management workflows. This broad deployment has delivered:

• 83% reduction in research time
• Significant gains in fraud detection, trading, lending, and CX
• 10–20% productivity gains for engineers via code assistants

Everyday AI is really working: JPMorgan’s LLM platform is used daily across its global workforce, enabling code assistants that boost engineer productivity by 10–20% (compare this to startups), and back‑office tools like EVEE Q&A and coding copilots that dramatically speed response. Their AI integration now spans 175–450 active use cases, scaling rapidly with structured ROI tracking and strategy for enterprise integration.

Leadership and organizational transformation are transforming everyday work. Teresa Heitsenrether, their Chief Data & Analytics Officer, leads the deployment of LLM Suite, which is now in the hands of 220,000 employees. Having led similar efforts integrating Microsoft Copilot and use case specific AI tooling at SMBC, I can say firsthand: setting up this kind of deployment in a regulatory environment is hard. JPMorgan is owning the AI infrastructure narrative in American banking, and setting a bar that’s hard to match.

JPMorgan is also doubling down on AI infrastructure with 2,000 employees in AI, including a 200‑person AI research unit, embedded in its broader strategy to retrain and evolve its workforce. How many people does your bank have working in AI? JPMorgan's investments in both AI tech, infrastructure, and teams are demonstrating why they own this narrative across banking's business model, at least in the Americas.

AI-Native Banking: Santander Sets the Benchmark

Jumping to EMEA, where my former colleagues are starting to make some real noise when it comes to AI infrastructure. Santander’s partnership with OpenAI appears to be more than a headline, it’s a signal. With 70+ AI projects already in flight and over 1,000 developers involved, Santander is embedding generative AI into its operational DNA, aiming to become the first truly “AI-native” global bank. I got to see these teams up close for many years, if Executive Chair Ana Botín says this is going to happen, count on that thing happening.

Santander frames AI not as a tool, but as a foundational pillar of its strategy across customer service, compliance, lending, and fraud detection. And unlike many banks still testing AI in sandbox environments, Santander’s move is production-grade and enterprise-wide. Watch carefully who they partner with, as it will be a slightly broader stack than who most banks work with (they've been very partner forward since the early development of Openbank and the launch of Santander InnoVentures, now Mouro Capital).

Related stories in Finextra and Fintech Futures add that the bank is focused on not just external transformation but internal enablement, with custom copilots for employees and developer tooling designed around OpenAI’s ecosystem. It's smart: set up the infrastructure, set up the governance and oversight, and let teams build on top of these trusted (and vetted) application layers that you've tee'd up for them. I wouldn't bet against my friends in Madrid to teach the industry a thing or two here.

Global Wave: AI Moves from Pilot to Platform

These certainly aren't isolated bets to the Americas or EMEA either. Commonwealth Bank of Australia is rolling out AI copilots built on Microsoft Azure OpenAI, targeting productivity gains for both internal teams and retail banking clients. This is similar to what we were doing at SMBC, just likely with a bit more buy-in and support from leadership (that matters almost more than anything).

Canadian-based TD Bank is similarly leaning in, with a strategy focused on AI-as-a-service across fraud prevention, onboarding, and product personalization. TD’s Chief Innovation Officer Rizwan Khalfan calls it “a turning point in banking’s digital maturity.”

And Arizona State University is already preparing the next generation of banking professionals for this new reality, with coursework and partnerships focused on prompt engineering, AI ethics, and human-AI collaboration.

How are you getting your employees prepared and engaged with AI? Having helped lead both AI training and use case discovery the past few years, it takes a very dedicated team with a focus on building what we call at Darrery Capital tradecraft AI.

This is delivering craft mastery in a way that scales… understanding the nuances of the domain, the weight of different signals, and the rhythm of expert decision-making. It's building software with expertise embodied through unsexy but critical features that surround the AI: the data pipelines that automatically pull in the right context, the interface design that anticipates a user's next three clicks, the integration architecture that embeds seamlessly into existing workflows, and the feedback systems that improve recommendations over time to continuously learn as a craft master does. -- Mike Degnan, Darrery Capital

Stablecoin Rails and the Regulation Moment

While AI grabs the headlines, stablecoin infrastructure is quietly reshaping the financial plumbing underneath. With new U.S. legislation normalizing digital asset and M&A moves like Ripple x Fortress, we’re entering a new era of programmable value transfer. As The Financial Brand explains, the U.S. now leads Europe in defining a regulatory perimeter that may accelerate adoption across cross-border payments, remittances, and B2B liquidity flows. Stablecoins are quickly becoming a new layer of financial plumbing: programmable, always-on, and settlement-native. This unlocks entirely new service models and challenges for incumbent banks, fintechs, and DeFi-native players. Highly suggest you follow Ron Shevlin , Simon Taylor , Alex Johnson and Jason Mikula in addition to subscribing to our substack run by Michael Degnan .

Infrastructure Shifts: M&A and Middleware

Ripple’s $200M acquisition of Fortress is a case study in the convergence of crypto and traditional payments infrastructure. Fortress offers APIs and middleware that connect banks and fintechs to the tokenized asset ecosystem, precisely the layer needed to make stablecoin adoption real for institutional clients. This is part of a broader land grab as incumbents try to “buy” their way into relevance as crypto-native firms mature.

Significant changes are coming folks, and it's no longer in drips, it will be in waves. Middleware providers like Fortress are being scooped up for a reason: stablecoin adoption, API layers, and tokenized asset connectivity are quickly becoming essential infrastructure.

And don't get me started talking about Circle as they are just killing it. Having met their team back in 2015 (around the same time Santander invested in Ripple) while our investment thesis was coming into form, they have both come full circle (pun intended) and I can't be more happy for each of their teams. The progression of Circle and the evolution of tokenization is going to be a great one to watch and invest and partner into.

Wildcards and Future States

Finally, another read that was really thought provoking. A piece in Fintech Futures by a banking OG I haven't caught up with in some time ( Dave Wallace ) is speculative, strange, and brilliant. “The Agentic Three-Body Problem” isn’t just speculative fiction — it’s a provocation. It explores what happens when AI agents evolve from tools into autonomous actors, each operating with distinct goals around optimization, compliance, and human alignment. It’s less about singularity and more about multiplicity, systems that make decisions on our behalf but aren’t always pulling in the same direction.

In a future defined by agent swarms, some customer-facing, some internal, some adversarial, the challenge won’t be technical performance, but coordination, ethics, and control. Think of it as financial services crossed with the social dynamics of competing civilizations. It’s a scenario that feels far out… until you realize fragments of it are already here (and it kinds of reminds me of the AI paperclip scenario with AGI).

What happens when your customer service agent, risk model, and compliance copilot start “negotiating” trade-offs across objectives? It’s the kind of question that keeps technologists, ethicists, and regulators up at night — and it’s exactly why this piece is required reading for anyone building AI systems with real-world consequences.

You can find this story in the comments on LinkedIn. Dave's article also reminds me about the upcoming season 2 of Apple's Three Body Problem. So good.

Staying in the loop

Along with my work with the team from Darrery Capital , I am looking to write more across platforms (now that the initial generation of social media is mostly dead). Between my recent post on what libraries taught me about AI in financial services and sharing Michael Degnan's thoughts on the tradecraft moat, I plan on making up some lost ground.

If you’d like a regular dose of fintech x AI x infrastructure insights in the Make Banking Better newsletter, hit that subscribe button. Want to help shape it? Reach out to tell me what's on your mind and what you're working on. I'll be sure to cover it in future editions.

Also be sure to visit Darrery Capital and opt in to our Substack.