Hi everyone, 

I’m Julie VerHage-Greenberg with This Week in Fintech. Hope you all had a lovely Holiday and were able to take some time to relax and reflect. 

We’re back with another episode of The Architects, our special series co-created with Flourish Ventures, where we spotlight the people and ideas reshaping the U.S. financial ecosystem.

Today’s conversation is about a big, infrastructure-level question: what happens when money and assets become programmable, interoperable, and increasingly tokenized. And what a true “internet for finance” could look like.

I’m joined by Flourish Ventures’ Kabir Kumar, along with two guests at the heart of this transformation: Siddharth Shetty, CEO of Finternet Labs and one of the architects behind India’s open financial ecosystem, and Michael Lee, a financial research economist at the Federal Reserve Bank of New York, where he studies digital money, payment systems, and how innovation intersects with financial stability.

We dig into tokenization, trust, privacy and compliance, and why AI may become the interface layer for how value moves in the future. Let’s get into it.

Episode Summary: The Finternet: Tokenization, Digital Money & the Next Financial Infrastructure

[00:00 – 02:37] Pre-Flight Check: Ambitious Agenda, Big ConceptsJulie opens by checking whether the guests are comfortable with the question set and acknowledging they likely won’t get to everything. The tone is collaborative and candid, setting expectations for a wide-ranging conversation about infrastructure, regulation, and what “next-gen finance” could actually look like.

[02:37 – 08:02] Building the Finternet: Lessons from India (and Beyond)Siddharth Shetty introduces the origin story: years of designing digital public infrastructure at population scale, starting in India, then applying learnings across dozens of countries. The core observation: most financial infrastructure is still purpose-built (country-specific, rail-specific, use-case-specific), which makes global interoperability painfully slow. The Finternet is presented as a response: a new architecture meant to reduce the friction of connecting systems across jurisdictions.

[08:02 – 09:24] The Big Idea: “Imagine an Internet for Finance”The episode crystallizes the “Finternet” as open, programmable, and interoperable by design, not a single product, but an infrastructure model. The promise is that value transfer could become more universal, like information transfer did on the internet.

[09:24 – 14:42] The Fed’s Lens: Evolution, Not a Single Master PlanMichael Lee frames today’s financial system as an evolved artifact, shaped by constraints, tradeoffs, and history, not the output of a single design. That perspective matters because it tempers “reinvent everything” thinking. He also draws a key distinction: the future isn’t just about new technology, it’s about aligning technical capabilities with the goals we care about (stability, efficiency, fairness).

[14:42 – 19:54] A Future Built Around Agency and EmpowermentA major thematic pivot: both guests talk about agency as the north star. The future system should give people more control over identity, credentials, data, money movement, and access to markets. Siddharth pushes the idea that better rails and standards can reduce transaction costs and open participation beyond the top slice of the population, turning “ownership” and “access” into default capabilities rather than privileged ones.

[19:54 – 24:44] Why Tokenization Matters to the Fed (and Why Consensus Isn’t the Point)Kabir asks why tokenization is a serious area of analysis inside a central bank context. Michael makes two important points:

• There isn’t one “Fed view”, diverse internal viewpoints are a feature, not a bug.
• Tokenization is interesting because it may enable new arrangements that are difficult in legacy systems, but the hard work is as much legal and regulatory as it is technical.

[24:44 – 31:33] Tokenization 101: Verifiability + TransactabilitySiddharth offers a clean mental model for why tokenization is different from “just digitization”:

• Verifiability: proving identity/credentials/asset provenance cheaply and reliably (critical for compliance, sanctions screening, investor eligibility, etc.).
• Transactability: transferring claims/property rights more universally, without endless bespoke integrations and bilateral agreements.This becomes the bridge to why tokenized systems could unlock broader interoperability across asset classes (money, deposits, securities, registered assets like property, and more).

[31:33 – 36:23] Mass Adoption Depends on Legal Frameworks, Not Just Better TechThe conversation gets practical: tokenization can work in pilots and corridors, but scaling requires legal recognition and regulatory clarity. Michael reinforces the idea that many barriers in finance are not “we can’t,” but “we intentionally restrict”, because safeguards matter. The question becomes: how do you preserve the safeguards while still capturing efficiency gains?

[34:32 – 38:06] Privacy + Compliance: Finding the Line (and Why New Tech Changes the Equation)Michael unpacks what’s newly possible: systems where transaction contents can be hidden from everyone except the parties involved, potentially even hidden from the system operator, while still enabling compliance under defined conditions. The core tension: privacy technology is getting stronger, but “bad actors love anonymity,” so the design challenge is balancing privacy guarantees with enforceable controls.

[38:06 – 42:57] Tokenization’s Real Unlock: Democratizing “Enterprise-Grade” Financial ToolsMichael describes a powerful implication of verifiable ownership and commitments: individuals may be able to access arrangements that were historically bespoke and reserved for sophisticated players (e.g., committing future cash flows or proving ownership/rights in ways that reduce underwriting friction). The underlying point: tokenization could compress the cost of trust.

[39:39 – 46:28] AI + Money Movement: A New Retail Payments ExperienceJulie flags the buzzword they haven’t hit yet: GenAI. Siddharth argues AI changes payments in two ways:

1. Experience layer: moving beyond today’s fragmented UIs and “checkout pages” toward delegated, agentic flows.
2. Optimization layer: agents can search liquidity/routes/fees and execute within constraints (limits, permissions, observability), reducing the cognitive load on humans.A key connective claim: AI’s upside is much bigger in a tokenized world because programmable assets allow agents not just to recommend, but to execute safely.

[46:28 – 50:09] Reality Check: The Journey Is Just StartingMichael closes the main discussion with a sober warning: mainstream adoption is hard because consumers have limited headspace for new financial behavior. For tokenization and agentic AI to matter, they must be safer, simpler, and measurably better, otherwise the benefits accrue only to sophisticated, already-advantaged users.

[50:09 – End] Rapid Fire: Dinner Guests, 2025 Highlights, and De-Stress RitualsThe episode wraps with a light rapid-fire round (including the required disclaimer that Michael’s views are his own). It’s a clean tonal contrast to a dense, systems-level conversation, and a reminder that behind the infrastructure debates are real people trying to design something that lasts.