Podcast 162: Stablecoins are the pathfinder for digital assets, with David Birch

Stablecoins have been circulating in the fintech conversation for two years. Blockchain has been around for more than ten. The signal-to-noise ratio in the discussion is close to zero.

In this episode of the Fintech Garden Podcast, recorded at the Fintech Garden Budapest studio, host Igor Tomych sits down with David G.W. Birch — author, consultant, and one of the longest-standing pattern recognisers in digital money and digital identity. Birch steps back from the current news cycle and applies 40 years of experience to the question: what is actually happening, and what does history suggest happens next?

The answer is more interesting than the industry conversation would suggest.

The Industrial Revolution Had Stablecoins Too

David opens with the historical parallel most of the industry has missed.

In late-1700s England, as the country transitioned from a rural to an industrial economy, factory owners needed money to pay workers.

The state had not yet produced enough of it in the new form the economy required.

So private actors stepped in.

Factory owners minted their own tokens — usually in copper — and those tokens circulated as money. Businesses grew. Supply chains formed. New industries came into existence.

These were, functionally, stablecoins.

Private-issued instruments filling a gap the state could not yet fill.

Over time, the state absorbed the function. The tokens went away.

But something more important happened alongside.

Entirely new institutions — banks, the banking system as we recognise it today — emerged to interface with the new economy.

The tokens themselves disappeared. The institutions they created stayed.

That is the pattern. And it is the pattern that is likely to repeat now.

Which Means: The Stablecoins Are Not the Durable Object

David’s read of the current moment applies the historical lesson directly.

There is now an explosion of stablecoin issuers.

Over time, the state will step in — either by absorbing the function through central bank digital currencies, or by regulating the space down to a small number of dominant issuers.

The margins on the medium of exchange are structurally thin. You cannot charge much more than a dollar for a dollar.

Most current stablecoin issuers will not survive as independent businesses.

What will survive is the new class of institutions being built to interface with, manage, and control this infrastructure.

Companies like Fireblocks. Increasingly, the incumbents who understand what game is actually being played.

The stablecoins are not the durable object.

The institutions being created around them are.

If you are building in this space, that reframes what you are actually building.

The New Economy Needs a New Kind of Money

The specific reason stablecoins are appearing now is that a new kind of economy is emerging that needs money capable of operating within it.

Distributed ledgers. AI agents. Networks of interacting bots.

All of them need money that functions in their environment. Traditional bank money does not.

David is careful about what stablecoins are actually being used for today. Not much of it is genuine payments yet. A large share is still cryptocurrency speculation, wash trading, and adjacent activity.

But real-world usage is growing. Payment volume will follow.

The demand behind this is worth naming precisely.

It is not demand for cryptocurrency. It is demand for dollars.

In most emerging markets, and even much of the developed world, cryptocurrency was an imperfect way of getting dollars. Stablecoins are simply a better way to get them.

That is the demand curve underneath the market.

Not speculation. Not ideology. Practical demand for USD-denominated value that moves easily across borders.

AI Is Likely to Be the Dominant User

David endorses the argument — attributed to Mike Novogratz and others — that end-users are probably not the audience fintech operators should be optimising for over the next five years.

The dominant users of stablecoins may not be humans at all.

Networks of AI agents, acting on behalf of people and businesses. Booking. Buying. Negotiating. Paying each other.

Agents need a form of money that operates natively in digital environments, without the friction of traditional banking rails.

The stablecoin architecture answers this need in a way legacy infrastructure does not.

This shifts the strategic question.

From: “How do consumers use stablecoins?” To: “How do agents use them, and what emerges from that?”

The second question is bigger.

The second question is also the one most fintech product roadmaps are not yet designed around.

Regulation and Usability Are the Two Maturity Gaps

Two structural gaps still need to close before stablecoins move fully into the mainstream.

The first is regulatory. The US has moved aggressively with legislative infrastructure — a fit for an environment where most people already want dollars. Europe has leaned harder into consumer protection. The UK sits somewhere between. Global synchronisation will take time.

The second, which David identifies as more important, is usability.

In a Substack piece David references, Jeremy Light illustrates the problem.

You want to send someone USDT.

There are around ten different chains offering it.

Which one do you use? How? What UX works best in that environment?

The industry is nowhere near settled on these answers.

Which is why stablecoins still feel like the grey zone to mainstream users — not because they lack utility, but because the utility is buried under implementation friction most users cannot navigate.

AI Changes What Interoperability Means

Here is a reframe worth pausing on.

The historical fintech assumption has been that interoperability requires common standards, established protocols, shared APIs. The tedious, multi-year work of getting a fragmented ecosystem to agree.

David recently saw a demonstration of an AI querying open banking APIs autonomously — probing, learning, figuring out which endpoint did what without any prior standardisation.

Within a short period, it could talk to all the banks in scope.

If AI can bridge fragmented infrastructure at query time, the pressure for formal interoperability diminishes.

Interoperability doesn’t mean what it used to mean.

This is a genuine reframe of a problem the industry has been solving for two decades. It suggests that some of the standards work currently in flight may be redundant before it is complete — not because it fails, but because the substrate underneath has changed.

Why BlackRock Actually Cares — And It’s Not Payments

This is the strategic insight most fintech commentary is missing.

Large institutional players like BlackRock are interested in stablecoins.

Not because stablecoins are a better way to move money.

Because stablecoins are the pathfinder for digital assets.

The technical difference is worth spelling out precisely.

Swift is a messaging network. It sends messages, not money. Clearing and settlement happen separately, underneath, through bank relationships. That separation is expensive.

In a stablecoin architecture, money and data flow together in the same transaction.

There is no separate clearing and settlement layer.

Once the networks are in place to move stablecoins around, they can move digital assets.

Initially, digital assets get exchanged for stablecoins. Downstream, digital assets get exchanged for other digital assets.

The cost reduction on financial intermediation, David estimates, is roughly 80%.

That is not an optimisation. That is the game.

Understanding this reframes what BlackRock, Fidelity, Franklin Templeton, and the other large asset managers are doing in stablecoins. It is not about payments. It is about the underlying rails for the next generation of financial market infrastructure.

The payments story is the cover. The infrastructure story is the strategy.

The Missing Layer Is Digital Identity

The BIS calls the emerging infrastructure the “next generation FMI” — Financial Market Infrastructure — describing it as the internet of value.

The missing layer is digital identity.

Money moves. The trust framework around who is moving it, and whether they are authorised to, has not been built to the same standard.

This becomes especially consequential in an agentic environment.

Agents themselves need identities.

China has just passed a regulation requiring all robots to have identities. Literally, robot passports.

The open design question David flags: do we build agent identity by extending existing human digital identity infrastructures, or by creating new identity systems for agents that interoperate with the human ones?

This is the design decision that will shape the next decade of financial services.

Whoever solves it well owns the layer.

The Trust Layer vs the Execution Layer

There is a related architectural distinction David draws.

The trust layer. Is this really Dave Birch? Is this really Barclays Bank? Am I really talking to Walmart? Am I authorised to spend this amount on these things?

The execution layer. Now I press the button and send the money.

Protocols in the trust space are coalescing. Not fixed yet, but forming.

The gap David identifies is in the guardrails between the two layers.

Decision → execution needs something more between them than we currently have.

The engineering work to fill that gap is where the practical, near-term product opportunity sits.

Not glamorous. Not the AI story most operators are chasing. But structurally more important than most of the current product-roadmap conversation.

The Five-Year View: Financial Services Disappear Behind Bots

Asked where this goes over the next five years, David’s answer is direct.

Most people will stop interacting with the underlying financial services altogether.

AI bots will handle those interactions on their behalf.

David’s own example is car insurance. “I never want to talk to my car insurance company again.”

Once a bot can handle it, most people will hand it off.

The financial services layer does not vanish. It just retreats behind an interface most users never touch.

This has consequences for how banks and fintechs should think about brand, distribution, and customer experience.

The customer, over five years, may increasingly be another bot rather than a human.

If your customer acquisition strategy is optimised for human attention, you are optimising for a shrinking market.

The Longer View: Money Itself May Disappear

This is the philosophically striking point that closes the conversation.

If digital assets can be traded continuously in globally liquid markets 24/7, then any transaction between two parties can be resolved as an exchange of digital assets.

You never need to convert into fiat to settle.

You just exchange baskets of assets. What do you have. What do I want. What can we agree on.

In that world, money as we currently understand it becomes redundant — an abstraction layer no one needs anymore.

David acknowledges this sounds radical.

He also points out that it feels crazy only because humans are used to negotiating in currency.

For bots, exchanging half an Apple share for a flight ticket to New York is trivial.

Igor’s engineering reframe here — that money is a derivative of the underlying value, and if you can exchange the primary objects directly you no longer need the derivative — David explicitly endorses.

Exchange primary objects, not derivatives.

This is the sentence that will still be true in ten years.

Everything else in the current stablecoin conversation is scaffolding around it.

Where Innovation Is Actually Flowing

Two closing observations worth noting.

One: The entrepreneurs building AI and bot-based services today are not asking to be given bank accounts or credit cards. They are simply building on stablecoins.

This is where innovation is being drawn.

Unexpected new services will emerge from this space. Not the simple DeFi replicas the industry has already discussed. Something deeper.

If you are looking for where the next wave of financial services innovation appears, it appears here, and it appears first.

Two: NFTs are quietly becoming useful again.

After the “chimpanzees with sunglasses” era discredited the category, the durable use cases are re-emerging.

Ticketing is the clearest example. Tickets are perfect non-fungible tokens — transferrable but uncopyable.

When you can trade fungible tokens (stablecoins) for non-fungible tokens (NFTs), you have the underlying structure of an entirely new class of economic activity.

The two categories that seemed to be fads separately look meaningful together.

What This Means for Operators

The episode does not deliver a checklist. It delivers a framework.

Pulled out into operational implications, five things stand out for anyone building in fintech now.

One: The stablecoins are not the durable object. The institutions being built around them are. If you are building in this space, build the institution that will still be relevant when the state absorbs the medium.

Two: The 80% cost reduction on financial intermediation is the strategic prize. Payments is the cover story. Digital assets infrastructure is the play. Align your product strategy accordingly.

Three: Agents are the customer over a five-year horizon. Human UX still matters, but the emerging demand curve is agentic. Product roadmaps built purely for human interaction are optimising for a shrinking market.

Four: Digital identity — for humans and for agents — is the missing layer. Whoever solves it well owns significant value. This is where infrastructure investment should concentrate.

Five: Exchange primary objects, not derivatives. This is the sentence to keep returning to as the industry evolves. It is the framing that survives the current cycle.

David’s closing point is understated but substantive.

The invention is happening. The unexpected consequences will follow.

The operators who understand what game is actually being played will be positioned when the current stablecoin conversation gives way to whatever comes after.

The ones who are still debating whether payments will move to stablecoins by 2027 are asking the wrong question.