Interview with the founder of USDC and Bridge on the rise of stablecoins

12月 12, 2025 12:24:58

Share to

Circle co-founder and USDC co-creator Sean Neville, along with Bridge founder (now part of Stripe) Zach Abrams, deconstruct the true story of stablecoins and the future of internet-native currency.

They share how to find product-market fit (PMF), how to survive early chaos, and how to build financial infrastructure that can scale massively. They also reflect on the earliest days of USDC—when stablecoins were completely overlooked and regulatory frameworks did not exist. Sean discusses how his belief in "money operating at internet speed" shaped a multi-billion dollar asset. Zach recalls the transition of Bridge from NFT products to stablecoin infrastructure—and the changes in business after joining Stripe.

Together, they dissect the following questions:

-- Why are stablecoins experiencing a moment of explosive growth now?
-- What constitutes true product-market fit (and why founders almost never "feel" it)?
-- The challenges of building crypto or AI products in a highly regulated environment.
-- Why the world may need new underlying chains like Arc and Tempo.
-- How decentralization, liquidity moats, and interoperability will shape blockchain adoption in the next decade.
-- What does the future of programmable money and AI-driven financial workflows look like?

Whether you are a founder, a builder interested in crypto, or someone trying to understand the direction of global payments, this conversation recorded at the a16z crypto founders summit offers deep insights from two operators at the center of industry transformation.

Takeaways：

1. Stablecoins are the only true applications that have achieved "product-market fit" in the crypto space.

Stablecoins are the only products in the crypto industry that have truly found PMF.
Assets and exchanges do not count as "products," while stablecoins meet real needs: liquidity, cross-border capital movement, demand for dollars, etc.
The success of USDC comes from the liquidity moat of DeFi, while USDT comes from the liquidity moat of CEX.
New stablecoins face extreme difficulty in breaking through this moat.
In the future, there will be thousands of "tokenized deposits + stablecoins," but only a few will have a truly global impact.

2. Why are stablecoins only now exploding? (It's not a technical issue, but a regulatory and distribution issue.)

The biggest challenge has never been technical, but regulatory approval, partner ecosystems, and compliance frameworks.
Circle spent years educating regulators, and only by 2023-2024 did the market and policies mature in alignment.
Stablecoins are shedding the negative label of crypto, especially in the field of AI engineers, where "stablecoins ≠ crypto speculation."
The growth of stablecoins still primarily comes from crypto capital markets, but "real payment use cases" are emerging:
Remittances
Export settlements
Wallet savings (people in developing countries holding dollars)
But the scale remains very small (a few basis points of global payment volume).

3. Why is the new generation of base layers (Arc, Tempo, etc.) necessary?

Existing blockchains are fundamentally not optimized for payments.
Issues include:
Wallets needing pre-funded gas (high cost, poor experience)
Transaction congestion during peak times
Gas volatility
Default public, not private
Payment chains must have:
\<300ms final settlement
Privacy
Dedicated channels isolated from meme coin/DeFi transactions
Using the same currency to pay gas (e.g., paying USDC gas with USDC)
AI agents driven "automated payments" require entirely new infrastructure.
It is unrealistic to have AI agents manage a bunch of on-chain tokens just to pay gas.
Arc (Circle) and Tempo (Stripe) are independent paths, but the common goal is to build an "internet-grade payment chain."
Who succeeds is not important, as long as payments can shift from being a "byproduct of blockchain" to a "native capability."

4. The true nature of PMF: not an "epiphany," but an endless stairmaster.

Entrepreneurship is not a moment of "no PMF → PMF," but an ever-rising stairmaster:
One customer ≠ PMF
Five customers ≠ PMF
One million in revenue ≠ PMF
Funding ≠ PMF
Even a publicly traded company like MongoDB is constantly iterating, abandoning old businesses, and embracing new products.
Success is a series of leaps, not a destination.
Founders will never feel that their company is "destined for success."
Only employees may feel that way.

5. Why did Stripe acquire Bridge? (One of the most important signals in the industry.)

Stripe is one of the few companies in the industry that truly understands the "future scale of stablecoins."
Zach originally thought "no one would want to acquire us," until Stripe demonstrated the same level of belief.
The experience of Bridge post-acquisition is a "dual-speed world":
Daily operations are more complex (processes, systems, compliance alignment)
Strategically, speed is accelerated by 10x (bank partnerships, issuance platforms, enterprise clients)
Once Stripe enters stablecoin issuance and payments, it marks a significant turning point for stablecoin adoption.

6. Stablecoin → AI Money: The next round of infrastructure revolution.

AI will execute financial workflows, with humans only responsible for oversight and optimization.
Examples:
Circle's compliance team originally handled AML monitoring
Sean's new company uses AI agents to fully execute workflows
Humans are responsible for review and decision-making
In the future, financial institutions will not need teams larger than a thousand; AI will reshape organizational size and operations.

Main Text:

Host: Many people in the audience have likely used the products you created, Sean. When you designed USDC years ago, it was hard to see it as a good idea: zero interest rates, no business model, and the market had no clear demand. What gave you the belief that USDC could succeed?

Sean:
Actually, this should have been seen much earlier, not just in 2017—I wrote the first version of the white paper that year and started coding. We founded Circle in 2013, and we already had a strong belief then—we always believed in "democratizing global finance on the internet track."
We started with Bitcoin but later realized what we really needed was a digital expression of government fiat currency—trustworthy, but able to operate at internet speed.

So we chose to put the dollar on Ethereum first. But this concept was never "single-chain," but rather "multi-chain, multiple issuers," opening up a "new world where currency operates at internet speed." The mission of USDC is to turn this vision into reality.

Host:
So why now? Why are stablecoins experiencing an explosion today? When you were doing this, almost no one understood its value.

Sean:
Because initially, it was a very difficult problem.
The difficulty was not technical—although the infrastructure at the time was indeed not mature enough, the hardest part was actually the regulatory system:

We had to clear regulatory hurdles
We needed to find suitable partners
We needed to build distribution channels

These factors were more critical than technical challenges.

We did not shy away from the problem; we knew it ultimately had to be written into law, but that would take time. So we invested a lot of time educating regulators, explaining how this mechanism works. From an economic perspective, for example, the "Chicago Plan"—which advocates separating money creation from the credit market—stablecoins actually turned this idea into reality. We often discussed these issues with regulators, but it was destined to be a long-term strategy.

At that time, interest rates were very low; for example, people like Balaji predicted when helping us push USDC on Coinbase that it might take 10 years to become a real business model. Fortunately, it happened faster than that.

Host:
When I was working with you on USDC at Coinbase, it felt more like a scientific experiment. No one knew what the business model was or how it would be used. The entire industry was far from mature compared to today; DeFi barely existed. CryptoKitties might have just emerged or might not have appeared yet. The entire ecosystem was very immature, and the development over the next six or seven years revolved almost entirely around stablecoins. Your approach at that time was completely counter to the trend.

Sean:
Even within Circle, there weren't many people supporting the development of USDC. We were doing a lot of things at that time. Although we had a strong conviction about "internet finance," we were very flexible about the specific path, especially in the early days when we were struggling to communicate with various states in the U.S. to obtain payment licenses.

We tried multiple product experiments, starting with Bitcoin, then moving to Ethereum (USDC launched on Ethereum), and we maintained relative neutrality regarding the underlying chains, only requiring them to meet some key characteristics.
The core focus was always on: a programmable currency layer that operates across multiple chains.

Host:
I have to tell a story. The day I visited your office in 2017, the market was crashing.
(Zach: Can you be more specific?)
That day was a "bloodbath," but looking back today, it seems trivial. Bitcoin dropped from $20,000 to about $6,000. I thought you would cancel the visit or that there would be panic on-site. But when I arrived at the office, everything was normal. You said you had an OTC trading desk, and as long as prices fluctuated, you could make money.

Sean:
By 2017, we had already been through four or five years, with countless small crises in between. For example, in the early days, we couldn't find banking partners, and only Silicon Valley Bank was willing to open an account for us—on that very day, I went to a panel at SVB, and the news of Mt. Gox's collapse broke on the same day; the bankers' reaction was to immediately withdraw. There were many small crises, but if you have faith in something, you persist and remain resilient; as long as you survive long enough, you can succeed.

Host:
I want to talk about those difficult moments. But first, let me ask Zach, your path has also been full of iterations. Bridge started with NFTs and later transitioned to stablecoin infrastructure. How did that happen?

Zach:
Our journey has been shorter than Circle's. Circle spent four or five years exploring before finding the final direction. We started our business during the NFT boom in 2022—wanting users to buy NFTs with bank accounts. The process was: deduct money from the bank → convert to stablecoin → buy NFTs. It turned out to be a terrible idea. But in the process, we discovered: "Building products with stablecoins is very difficult."

As time went on, we became increasingly convinced that stablecoins would become core payment infrastructure, and someone needed to solve the complexity of "connecting the banking world with the stablecoin world."

But even as we began to gain customers, we still constantly questioned whether we had found PMF. Two or three years passed, and we still asked, "Have we really achieved PMF?"

Zach:
I'm curious—when did Circle feel that stablecoins truly achieved product-market fit (PMF)? Because at that time, although there was a lot of usage, there was no business model, so it shouldn't count as PMF. Now that the company is publicly traded, perhaps you feel you have PMF.

Sean:
I have a possibly provocative viewpoint:
So far, in the entire crypto space (although in some sense it is still early), I believe the only one that has truly found "significant PMF" is stablecoins.

I don't think "assets" or "exchanges" count as products. I know many people disagree, but that is my view. However, stablecoins did not achieve this from the start. It did take time.

The largest stablecoin in the world (Tether) derives its liquidity moat from the centralized exchange's crypto capital market. Meanwhile, USDC's liquidity moat comes from DeFi.

But this was not formed early on. I think it began to accelerate around 2020 and has continued to grow. Now, while regulatory barriers still exist, at least "how to overcome them" has begun to emerge. The next two biggest challenges for stablecoin adoption are:

Liquidity moat
Interoperability

Given the current ecosystem, it is hard to imagine a new dollar stablecoin suddenly crossing the existing moat unless there is a significant change in the industry. This may be a minority viewpoint, but it is indeed my current perspective.

Zach:
I believe there will be many, many different types of stablecoins in the future. I don't know how many will break through the moat, but most actually don't need to.

Sean:
I agree. I think in the future we will see: any system that can keep accounts—like digital account balances in traditional financial databases—could be "tokenized." So there will be thousands of different forms of "tokenized deposits" and "stablecoins." But I also believe that the truly "important" ones may only be a few—those that can cross the liquidity moat and establish real scale effects.

Beyond the crypto capital markets, there will also be people pushing for payments and other use cases. We have discussed these issues for many years, but they are still not mainstream use cases. However—we are now very close to realizing them. I have been saying this for the past seven years (laugh), but now it is truly close.

Zach:
I also feel we are very close, but it is still hard to express "how early we really are." It seems that the situation has stabilized a bit now, with public companies emerging, and the industry appears to be moving from past chaotic cycles to some form of linear growth. But in reality:

Regulatory requirements are still not fully defined
Stablecoins account for an extremely small portion of global payment volume (a few basis points)
In the U.S. banking system, the "deposit scale" of stablecoins is almost negligible
Non-dollar stablecoins are virtually non-existent
Most banks and financial institutions still have no contact with stablecoins
The blockchain itself has not yet been optimized for truly scalable payments

So we are still very early.

Sean:
I completely agree.

Host:
Sean, you mentioned that you believe there will be multiple stablecoins in the future. Currently, Tether and Circle have first-mover advantages, capturing nearly the entire market. Do you think this dominant landscape will remain?

Sean:
Well, I don't have any special insights into the internal strategies of other companies (of course not). But I believe that the current trading volume of stablecoins still primarily comes from crypto capital markets. Although we are now indeed seeing some real cross-border payment use cases, such as:

Remittance payments
Import and export scenarios
People holding dollars on their phones (stablecoins are the best way)

These use cases have emerged, but the scale is still not large. So today's market landscape is still determined by the crypto capital markets. What is more interesting is that when stablecoins truly open up the core payment market, new players will emerge, possessing different distribution channels, partnerships, and capabilities. These new stablecoins will enter new markets in completely different ways, bringing about a new landscape.

I remain optimistic about this. Every company has a different view of stablecoins:

Some focus on underlying assets and investment returns
Some focus on product experience
Some "accidentally became stablecoin companies," even disliking the term "stablecoin"

These are the people I enjoy talking to the most; they refer to it as: "dollars" (laugh) or "streaming dollars."

Sean:
When I started my new venture (AI + finance), I long avoided using the term "stablecoin" because AI engineers have an extremely negative reaction to "crypto." I even think they are more averse to crypto than other developers. However, after this summer, the situation began to change; stablecoins gradually shed the negative impression of crypto and gained independent recognition. This is quite interesting. We may not call it stablecoin in the future, but the term seems to have become a brand.

Zach:
I think the term has already "solidified." (laugh) But you can continue to fight against it.

Host:
Zach, I want to return to the topic of PMF. Many entrepreneurs in the audience must be curious: how did you persevere step by step? How did you know "we finally got it right"?

Zach:
I want to say that people often describe PMF as a dividing line:

Before is "a desert without PMF"
After is a "paradise"

But in reality, it is not like that at all. For example, Bridge:

We conceived the product
Built the API
Launched the API

At that moment, we did not feel PMF. Later came the first customer—we didn't know if they would grow, and we did not feel PMF. Then 5 customers, 10 customers… every month there were new inbound requests, but they were all small teams from Latin America and Africa, and we did not know how big these customers could become. Still did not feel like PMF. We reached $1 million in revenue. I looked at the customer profile: it was a strange and chaotic mix of companies. I was completely unsure if this could constitute a sustainable business. Still not PMF. We raised Series A. This made me even more anxious—because I still did not feel we had PMF.

Until I attended an event and heard the CEO of MongoDB share:

Their company had $100 million in revenue when preparing to go public
Amazon launched a competing product, and the entire market thought they would "die"
They watched their original core business gradually decline to zero
The reason the company survived was a small product that only had $1 million in revenue at the time
Later grew into a billion-dollar business

This made me realize:

Even publicly traded companies are still experiencing the same "uncertainty and iteration" as entrepreneurs—just on a larger scale.

At that moment, I understood: entrepreneurship is not "reaching a node of PMF," but an endless stairmaster. You are always asking:

How do we go from $1 million to $5 million?
How do we go from 1 million users to 5 million?
Each stage's difficulty is higher than the previous one
Every step feels more "life-and-death"

It sounds a bit masochistic, but it has actually helped me a lot. Because it made me realize: "Uncertainty" is not a signal of failure, but the essence of entrepreneurship. Even today, with Bridge inside Stripe, I still feel that pressure every day.

Host:
So there is no easy path? It's all uphill?

Zach:
Yes. Maybe others have different experiences, but that has been mine. I have previously joined many companies that later succeeded:

Square (very early)
Coinbase (very early)
Brex (very early)

When I joined these companies as an employee, I felt that they were "destined to succeed." So I thought: there would also be a moment when I would feel that Bridge was destined to succeed.

But now that Bridge is part of Stripe, if a newcomer joins, they might feel that Bridge is "destined to succeed." But as a founder, I will never have that feeling. Because founders are always thinking: "What is the next life-and-death issue that will take us down?"

Host:
So how did you decide to jump off your "stairmaster" and switch to Stripe's "stairmaster"? What made you decide to be acquired?

Zach:
First of all, we were not "looking to sell the company." I believe good companies are not "selling themselves," but rather "being bought by others." And frankly, I once believed 100% that our company "could not be acquired." Because we had great faith in the future of stablecoins and believed they would be enormous. And I could not imagine:

There being another company
Large enough, with enough resources
That shares the same strong belief in stablecoins
And is willing to invest heavily in this frontier field

I felt I could not find such a company—even Stripe did not seem possible to me. Initially, I was trying to get Stripe to become a customer. I tried many times to get them to sign contracts and use our products. But they kept delaying.

However, as I interacted more with John and Patrick (Stripe co-founders), I gradually realized: their understanding of this field was astonishingly aligned with ours. They believed it was an important component of future payment infrastructure. They were also willing to invest significant resources.

At the same time, I found that if we did this with Stripe, our chances of success would significantly increase, and we would move faster—faster than I expected. For example:

Once the acquisition was completed, it pushed the entire industry forward significantly on the adoption curve
The new products we launched showcased what stablecoins could do to the world
The "open issuance platform" we launched could not attract similar clients when Bridge was an independent company
But after becoming part of Stripe, we could collaborate with large enterprises and banks

All of this proves: Joining Stripe accelerated our efforts far beyond my expectations.

Host:
What changes have occurred in your life post-acquisition? How does the work style and company culture of Bridge differ after joining a large company?

Zach:
They do indeed force me to have a "cheeky pint." This is some sort of "initiation ritual." Back to the point, I summarize two phenomena:

(1) Daily operations are actually more difficult.

Although we operate quite independently:

We have our own office
Our own Slack
Stripe employees need to be introduced by someone to DM me
Google Doc permissions cannot be added casually

We do indeed "operate like an independent company." But daily operations are more challenging than when we were independent because, in addition to continuing to:

Find banking partners
Build crypto infrastructure
Develop wallets
Do product and customer support

We also have to do a lot of "alignment work with Stripe," such as:

Our CRM needs to integrate with Stripe's CRM
Customers need to be identifiable by both teams from which entry they come
Internal system permissions and processes need to align
Many operational processes must be bi-directionally integrated

All of these make daily operations more complex. So at the daily level, the pressure is just as high, if not higher.

(2) But from a macro perspective, the "acceleration" is immense.

Some things that would take years for Bridge to achieve independently are now accelerated by Stripe at key points:
For example:

The launch of our card product
The ongoing push for enterprise adoption
Collaborations with large banks
The stablecoin issuance platform

These are things that could only be accomplished after joining Stripe. So on a macro level, the acquisition has put us on a faster track.

Host:
So you are still on the stairmaster.

Zach:
Absolutely correct.

Host:
Sean, you founded Circle over a decade ago, and now you are starting a new company in the AI era. How has your approach to entrepreneurship changed compared to back then?

Sean:
Many things are different. But first, let me mention one "unchanged" aspect: when we founded Circle, we knew finance was a highly regulated industry, so:

We had to invest heavily in educating regulatory bodies
We had to understand the international regulatory landscape
We had to know what would ultimately be written into law

The same applies in the AI + finance space:

You must consider how AI workflows will impact global finance
You must think about which AI safety standards will become law
You must navigate the political dynamics of various countries regarding foundational models (LLMs)

These patterns are very similar to the early days of Circle.

But the truly different aspect is that the operational methods at the "tactical level" of the company have changed dramatically.

For example, at Circle, we had many risk control and compliance analysts:

Monitoring for money laundering
Opening cases
Collecting evidence
Submitting suspicious activity reports (SARs)

Now, in my new company, these processes are fully executed by AI agents,
with humans only responsible for: oversight, adjusting workflows, and optimizing efficiency. So the role of humans has changed, but the process efficiency is completely different. If we look at the scale of manpower—Circle now has about 1,200 people—I cannot imagine needing such a large team to do the same things today.

Host:
What you just mentioned reminds me of something—in the early internet era, people always said the internet changed all industries and modes of production, but the only thing that didn't change was GDP (economic data). And now with AI: it clearly enhances individual productivity, but it is completely invisible in hiring data. This is especially evident in San Francisco—when we are hiring engineers, we can clearly feel:
The demand curve for engineers is simply "shooting straight up to the sky."

Sean:
Yes.

Sean:
I do believe: you cannot build a financial institution "by just vibing code." It still requires very experienced engineers. But the way engineers work has changed completely:

The way principal engineers work is different
The way architects work is different
The development model is different
The toolchain is also different

I also believe that university curricula will be forced to change. But universities have not yet caught up with this change. So I worry about students currently in university:
They are actually learning a lot of content that is "irrelevant to current computer science."

Host:
So what would you suggest they study?

Sean:
Drop out and start a business.

Host:
What if they don't drop out? What should they study?

Sean:
Humanities, poetry. Back to: philosophy, political science, English. I'm not saying this casually—it's genuinely a difficult question. And this generation of students currently in university will face the most direct impact of this era.

Host:
Before we wrap up, I must ask you both a question. You are both working on new base-layer chains:

Circle is working on Arc
Stripe is working on Tempo

Why does the world need new base-layer chains? Even two of them?

Zach:
I'll go first, and then I would love to hear Sean's thoughts. We have been building payment applications on blockchain for over two years, and from the very beginning, it was clear: existing blockchains are fundamentally not optimized for payments. This is reflected in many "big and small" ways. For example, one of the "big problems":

A large neobank with millions of users wants to use a specific chain.
If they want to open wallets for all users on that chain, they must pre-fund gas for each wallet to receive USDC. Just this alone would cost millions of dollars. Another example:

One of our earliest clients was a government agency that needed to distribute relief funds to a large number of individuals. The first distribution involved a few thousand transactions, which is actually very few, right? But it took us 18 hours to complete all transactions on-chain. These are all "small" problems that accumulate into "big" problems. Additionally, there are several "real big problems":**

Privacy
Blockchains are default public, and core financial institutions cannot accept all transactions being transparent and public.
Discrete payment lanes
If a chain experiences a clearing storm or a meme coin transaction surge, causing gas prices to spike, how can we possibly build "core payment infrastructure" on such a chain?

All these issues indicate that there is a need for a chain born for payments.

We hope Tempo, as an independent project, can solve this problem. But if Arc solves it, we would be happy too.

Sean:
I completely agree. If you aggregate all these demands, for example:

Must support \<300 milliseconds final settlement
Must ensure privacy
Must avoid competing for block space with meme coins
When sending USDC, gas should also be paid in USDC
AI agents should not have to manage multiple on-chain tokens just to pay gas

We have previously built on Base and Solana, but AI agents must hold the correct gas token, which is a reliability issue. So we ultimately realized: to truly unlock new AI + money use cases, we need a chain designed specifically for this purpose. The design centers of Arc and Tempo may differ, but the goals are aligned.

Zach:
The best part of the entire industry is: a few years ago, everyone wanted to solve the problems of "scalability, lowering gas, and increasing throughput," leading to the emergence of:

Various L2s
Solana
Aptos
Sui
Other new chains

This is a creativity-driven evolutionary process. The same creativity is now happening in the payment chain space:

Tempo
Arc
Plasma
And other independent teams

Ultimately, one or more chains will emerge victorious, which is good for the industry.

Host:
Last question: decentralization has always been a core value of crypto. Some are concerned that new payment chains may bring "centralization risks." If you had to convince them in one sentence, what would you say?

Zach:
One sentence: If these chains are centralized, they will not succeed. No bank would be willing to build its payment infrastructure on "another bank's chain." For Tempo and Arc to succeed, they must achieve true decentralization.

Sean:
We are essentially saying: entrusting the security of money to cryptographic technology is more reliable than entrusting it to a company or an individual. Decentralization vs. distributed, how to implement this can be discussed—but the principle cannot change. No one will use a "corporate chain." So the answer is: Crypto over corpo.

Host:
Thank you very much, Zach and Sean.

(Source)