Coinbase CEO Brian Armstrong on bitcoin going to $1 million, electing a pro-crypto Congress, and Jamie Dimon

John Collison:
It feels like a good time to do this and take stock. Stripe—we launched in 2011. Coinbase was founded in 2012 and grew up at the same time. Coinbase just got added to the S&P 500, The GENIUS Act just passed, so a good moment to step back and take stock.

Brian Armstrong:
Yeah.

John Collison:
As you look back on it, when I think about crypto in, say, the 2014–2015 era, it was such a red ocean. There were so many different competitors, and you guys beat so many people to get here. As you look back on it, how did Coinbase win?

Brian Armstrong:
It has been a long journey, and I remember seeing you guys go through YC and everything was very inspiring for me. It was, “Okay, maybe a fintech company could be built in this area.” So I had a lot of aspiration to go through YC, in part by seeing what Stripe did. And you’re right, in those early days, it was the Wild West. I’d go to the early Bitcoin meetups, and there were brilliant cryptography PhDs, and also anarchists. One of the first meetups I went to in San Francisco for Bitcoin, a guy there told me he was starting his own religion. I thought it was a little off‑topic, but somehow it all worked together in his mind. So we’ve come together from these early days, which reminds me of the Chaos Computer Club equivalent. And I think Coinbase made a couple decisions there that helped us ride through. One was that we decided to follow a regulated approach, which meant working with the United States government and getting money transmitter licenses. Which I remember the anarchist crowd at that time—

John Collison:
This was considered selling out?

Brian Armstrong:
It was a little bit. One guy—I remember—I had just gone to meet with some banks or something like that, so I was wearing a suit. I came to the Bitcoin meetup and the guy was—

John Collison:
You’re literally the suit!

Brian Armstrong:
Yeah. He said, “You look like a banker.” And I said, “No, no. I’m a software guy, trust me.” I think the other thing that we did was, we tried to stack up as many credibility indicators as we could. One of them was getting accepted into Y Combinator. That led to us getting a bank relationship in the U.S. at the time when that was very difficult to do, with Silicon Valley Bank of all places. Getting the money transmitter licenses. There was a period in the U.S. where we were, I think, the only U.S. company that had a bank partnership, so it was easy to connect your bank account to buy bitcoin. That really allowed us to get started in a way that nobody else could catch up.

John Collison:
So you were more credentialed versus... Who were the competitors at the time? Was it Mt. Gox? Who were you competing against for consumer share?

Brian Armstrong:
It’s been so long. Mt. Gox was in Japan, which of course blew up. There was a company called Tradehill in San Francisco, which probably could’ve been a big exchange, but it didn’t hit some of those milestones. There were so many of them along the way. I couldn’t even name all of them. There was a Bitcoin conference in San Jose, California in 2012, that was kind of iconic in hindsight. The Winklevoss brothers showed up, and I met Vitalik Buterin there for the first time. He had not invented Ethereum yet. I remember people looking around the room and it felt so small‑time. Some of the talks only had five people show up to it. We had this little booth with T‑shirts.

John Collison:
What was Vitalik working on before Ethereum?

Brian Armstrong:
Oh, he was a writer at Bitcoin Magazine. He was actually a very good writer.

John Collison:
You see it in his—how he leads the Ethereum movement. Yeah, exactly. From the philosopher king.

Brian Armstrong:
Yeah. I don’t know if there’s a lesson to draw from this, but I think being early on some of these tech trends—you don’t want to jump in and be the 42nd AI company or something. You want to jump into something when you think it’s cool, nobody else thinks it’s cool. I guess Paul Graham has that thing. You’d rather have 1,000 people who really love your product than a bunch of people who don’t care. That’s where we were. Nobody took seriously what we were doing—us at Coinbase or the entire industry at that time.

John Collison:
Can you describe being the more buttoned‑up player? Having the money transmitter license, having bank partnerships. Did that help you win because you were able to build products that other people weren’t able to follow you with? Or did it help you win because, for consumers, it gave you a brand and a trust that led them to choose you?

Brian Armstrong:
I think it was both. It allowed us to launch products that others couldn’t. It allowed us to not get shut down.

John Collison:
Did other people get shut down?

Brian Armstrong:
Yeah, they either just—They got cease and desist letters. They couldn’t afford the legal bills. They got shut down by getting hacked, basically.

Brian Armstrong:
We had a couple of close calls in the early days where we were not the mature company we are today, and we had people trying to break into our systems. I can tell you about some of those close calls if you want. We were able to get just enough good talent in the door due to the credibility indicators. By the way, some of the other companies at that time were totally anonymous. These people were telling me, “I don’t want to put my name on this website because it’s against the ethos of crypto.”

In my view, “if we’re going to raise money and be regulated, how are you going to be anonymous forever? If this gets big enough, someone’s going to come knocking on your door.” I wasn’t afraid to put my name on it and say, “I’m a U.S. citizen. I live in the United States. I’m doing this for the right reason. I’m here for the long term.” Companies are a reflection of their founders. Stripe’s very much a reflection of you and Patrick—bits and pieces of you in many ways you probably don’t even realize.

John Collison:
We are in a pub, after all.

Brian Armstrong:
An Irish pub. This is one small example. The same thing was true of Coinbase with Fred Ehrsam and I. It was really a derivation of our DNA—which was many things—but one of them was, we were legitimate and trustworthy and we were doing it for the right reasons for the long term. That counted for a lot.

John Collison:
Fred had worked in the financial world before, had you? No, you were at Airbnb before. Had you worked in finance at all before?

Brian Armstrong:
No. I had studied computer science and economics. I’d worked pretty much as a software engineer. I had tried doing another startup that was more of a base hit—didn’t work. Fred got a computer science degree and economics too at Duke, but he went into FX trading at Goldman. When I first met him, I thought, “okay, the guy knows how to write code, he has a CS degree, but he knows something about finance.” Within the first few weeks of us working together, he came back to me at one point and said, “I’m pretty sure we’re losing money every time someone buys bitcoin.” I said, “How could that possibly be? He walked me through it on the whiteboard, and he was right. It was, “Okay, complementary skill sets.”

John Collison:
V1 of the business model is not going to work. You need some optimizations.

Brian Armstrong:
Yeah.

John Collison:
That’s wild. You’re saying some of the close calls. I think that one thing people don’t realize is, they think when a startup isn’t working that it’s bad and stressful because no one’s buying your product, and that is true. But when a startup really is working, it is also bad and stressful in a way, because you’re being pulled along by forces, and it’s not obvious that you can keep up. What were some of the close calls?

Brian Armstrong:
I’ll give you a couple examples. On the first version of the app, I had created this simple hot wallet that was live on the servers. I remember telling people, “This app is an alpha, don’t store any money here you’re not willing to lose.” The company had raised about $150,000 or so at that point from Y Combinator. As early users came on the site, they kept depositing more money despite these warnings all over the app, and it was getting closer and closer to being $150,000. I felt like, if we lost all the money, I would want to be able to pay it back, or else we’d be insolvent.

John Collison:
You were tracking corp cash—how much money the company had—versus user deposits, and you didn’t want user deposits to outstrip how much money Coinbase had.

Brian Armstrong:
Exactly. I calculated at one point the growth of the deposits and how much money we had, and I said, “Hey, we have about eight weeks to put this onto some new system.” I had the vague sense we needed to move the money off the internet to cold storage. I didn’t know how to architect one. I’d studied computer science, I took a security class, I had basic understandings of cryptography, but I’d never built a real key storage solution or anything like that. So I called up two friends of mine. I said, “I need a crash course—how to do this. How would you architect it?” I remember asking one of them, “How long do you think it’ll take to build this?” He said, “A team of 10 people would take probably a couple years to really validate it and everything.”
I said, “We have eight weeks.” He said, “Oh, you’re in trouble.” I tagged one of the other engineers on our team, and we basically got to work coding up the first generation of the cold storage architecture. We were sleep‑deprived, and we made some trade‑offs that we thought were reasonable to get it there in time, but it worked. It was one of the most do‑or‑die moments, where if we had waited longer and the company had gotten hacked, we wouldn’t exist here today.

There was another example like that, where we were sitting at lunch one day in San Francisco, and someone on their computer noticed there were a lot of refunds going out. We quickly realized someone had hacked into one of our customer support accounts and was issuing themselves refunds as fast as they could. I remember we shut down the whole website, figured out how they accessed that account, managed to block it, came back online maybe 12 or 24 hours later. We patched it, we went back to work, and I realized, if we had been asleep when that hacker started doing it, we would’ve been insolvent by morning. We only lost $50,000 or so in that incident, but it was pure luck that we were in Pacific time, San Francisco—they started doing it in a place where we noticed it.

There were moments like that in hindsight—very scary. They were a coin flip. There were three or four coin flips like that, not all of them cyber‑related. You’re right. Once it started to really work, there were other challenges about keeping up. One was just the level of customer support tickets coming in and we didn’t have a customer support team. So nights and weekends, 9:00 PM to midnight, we’d answer all the support tickets after doing the other work. At some point there were 10,000, 20,000, 30,000 unread support tickets, and people were starting to get really angry and frustrated with us. We realized, “Okay, we have to hyperscale here. How do we build a customer support team in the next seven days or something?”
Somebody on our team made this 10‑question quiz, because we couldn’t interview all these people that fast. We had people apply online, take this quiz. It was a really hard quiz about knowledge of crypto and other things. We didfive‑minute interviews with a bunch of them and hired them in, and we started to get up to speed. There were a lot of problems like that that we were figuring out along the way, by the seat of our pants.

John Collison:
What does the general tech public not appreciate about the cybercrime landscape?

Brian Armstrong:
One is that there are a lot of North Korean agents trying to work at these companies.

John Collison:
In the U.S. or remotely from North Korea?

Brian Armstrong:
For the most part remotely, as far as we can tell.

John Collison:
The remote work revolution.

Brian Armstrong:
Yeah. It does bring another vector. DPRK is very interested in stealing crypto. You’d think that we can collaborate with law enforcement—and we get these dossiers of “Okay, this is a known actor,” that we share sometimes with other companies. But it feels like there are 500 new people graduating every quarter from some school they have whose whole job is this. In many of these cases, it’s not the individual person’s fault. Their families will be coerced or detained if they don’t cooperate. So actually, they’re the victim as well in many cases. But they’re interviewing for these roles.

We’ve had to do a number of things along the way. First, these people are being coached offline as they’re on the camera, so we force them to turn on the camera and prove they’re not AI. We also started requiring everybody to come to the U.S. for orientation to access any sensitive system. Fingerprinting them, making sure anybody with sensitive access has U.S. citizenship and family in‑country. You don’t want someone to feel like they can flee and then have no fear of extradition or these kinds of things. We’ve had to adapt all of this on the DPRK side.

The other thing that was surprising to me is the willingness of these threat actors to try to bribe our customer support agents. Our customer support agents work in facilities that are pretty locked down, and they have a Chromebook that is pretty locked down. In some cases they’ve been offered hundreds of thousands of dollars to smuggle in a personal phone and take photos of a screen or something. You’d think, “Wow. Okay.” We’ve really had to lock down the access that these agents have. We’ve started to move more of it to the U.S. and Europe. We just opened a new customer support facility in Charlotte, North Carolina, and really started to make a deterrent in the sense of, when we catch people doing this—and we red‑team it consistently—we don’t walk them out the door—they go to jail. We try to make it very clear that you’re destroying the rest of your life by taking this, even if you think it's some life‑changing amount of money, it’s not worth going to jail. These are the kinds of things we’ve had to deal with now, and I’m sure the stakes will just keep getting higher.

John Collison:
So more proof of physical presence, more compartmentalization, and more deterrent effect through aggressive prosecution.

Brian Armstrong:
Good summary. Yeah.

John Collison:
It feels like the proof of physical presence is going to become a bigger deal in a world of AI deepfakes—just higher stakes for all this kind of cybercrime—where in a weird way maybe you see certain areas where remote work goes backwards. People are saying, “Oh no, we want to do our customer service in the United States,” or other examples like that. It feels like the physical presence is becoming more relevant and not less.

Brian Armstrong:
Yeah. By the way, people may have seen this too, but we’re turning the tables and we’re actually going after threat actors too. We put out a $20 million bounty for information leading to the arrest or conviction of folks who recently conducted an attack on our customers—who we made whole, by the way. We’ve gotten a lot of inbound tips from that, and we’re having a fun time working with law enforcement on tracking folks down. I think that’s the ultimate deterrent—not going after insiders, but after the threat actors themselves. We need to be a hard target.

John Collison:
Who, after the DPRK, are the biggest troublemakers?

Brian Armstrong:
There are some groups in the U.S., and they’re usually younger people in their twenties caught up in hacker groups. Some of them are more mobile, they’re in Eastern Europe. Sometimes it’s about getting them into a U.S. extradition country, or sometimes they’re already here.

John Collison:
I want to take a bit of an inventory of everything that’s happening in crypto, and that’ll probably be helpful for the discussion. As I look at what’s working: store of value is for sure working in a big way. 24/7 trading of assets—I would include meme coins. There’s some success in payments so far, but I think both our companies are seeing that really ramp up. We’re very optimistic on payments volumes in crypto, especially stablecoins, over the next five years. Global access to the dollar is a big thing. What else would you include in that list?

Brian Armstrong:
Yeah, I think you hit the big ones. Prediction markets did, I’d say, enter mainstream consciousness during the last election.

John Collison:
For sure. That’s a good one.

Brian Armstrong:
And the volumes have continued. There are more things on the horizon we could talk about, but I think you nailed the big ones.

John Collison:
Crypto is, I guess, load‑bearing for prediction markets. We just didn’t have them pre‑crypto.

Brian Armstrong:
Yes and no. There were prediction markets before crypto, and there’s an open question currently legally in the U.S. about: if you build them entirely on‑chain with self‑custodial wallets, are they not financial services? In which case they’re much more permissionless about how they can be created—which is true in self‑custodial wallets and other areas. That’s another great one.

A cool thing about the trading use case too, by the way, is that every asset class is now coming on‑chain. It’s not people just trading crypto; they’re going to be trading stocks soon. Private companies are doing capital formation, and commodities and FX markets, debt instruments, Treasuries. I think every asset class is coming on‑chain, and we’re leaning into that with what we... We coined this term, “the everything exchange.” We’re trying to be a liquid market that’s global for every asset class. That’ll be exciting in a variety of ways, including more international reach. There are a lot of people in countries where they have high demand for U.S. assets. They want to invest in NVIDIA and these things, but unless you’re a wealthy person—usually in some of these markets—you can’t get a U.S. brokerage account opened. Then there’s the 24/7 trading, the fractional shares, and you can start to do things with perpetual futures and other interesting things.

John Collison:
So you think for, say, stocks, these ADRs—American Depositary Receipts—where if you’re an American and you want to buy a German stock or a Canadian stock, it’s weirdly tricky to do so. So you end up buying a derivative. You think on‑chain is just a better derivative of that?

Brian Armstrong:
Yeah. I think a lot of those will get tokenized. They are already in some ways. Then private‑company capital formation—or even people doing anything, they want to raise money for real estate development or to make a film. I think people will eventually be able to do novel things around governance too, where, let’s say, you really want to have only long‑term holders of your stock be voters. You can put that in the smart contract and say you had to have held this for at least a year to be able to vote. There’s interesting things you could do around that as well.

John Collison:
Is tokenization of access... I was going to ask you what the next big use case that starts working is. Would that be your vote? Or are there other next big use cases that you would point to?

Brian Armstrong:
Tokenization, yeah. The big ones right now are trading and payments. I think bitcoin as a store of value that’s inflation‑resistant. It’s not to be underestimated. That’s also a $20 trillion opportunity with gold as a comparable—but better than gold, I think.

John Collison:
Gold is a $20 trillion market cap?

Brian Armstrong:
Yeah.

John Collison:
And bitcoin is one and a half?

Brian Armstrong:
It’s two or three or something, yeah.

John Collison:
Oh, okay. I’m out of date.

Brian Armstrong:
Yeah. I think it’ll eventually be bigger than gold. If that’s all crypto ever was, that’s already enormous. But we’re starting to see borrowing and lending, capital formation. People doing decentralized social media too is pretty interesting. We just launched a prototype of this, or a beta of—

John Collison:
I watched the Base event.

Brian Armstrong:
Yeah. That was cool. Now people are posting content, and every post on social media is its own coin, and you, the creator, have your own coin, whose value flows up. People can buy it. If they like the post, they can reshare it and earn a share of the economics. They can remix it. You see sometimes on the internet people have a meme template, and there are a thousand—and all the value can flow back to the content creator. That’s an interesting primordial soup where some cool ideas are happening. I don’t know exactly how it’ll play out, but the Base app’s been fun. There’s a bunch of people on the waitlist.

John Collison:
There are some good new use cases. I think both of us would say payments actually qualifies for a major new use case, because over the last 10 years, it’s been pretty much just the crypto‑enthusiast use case, whereas now at least we are seeing a bunch of mainstream use.

Brian Armstrong:
Yeah. I was excited to see you guys come into the space, because it was super‑legitimizing, frankly. The tech was slowly getting there, and there was a scalability issue. There was a usability issue with, are you sending some random string of characters, or is it a human‑readable name? The wallets need to get better. Crypto should be the easiest way to pay. What’s your current thinking on what’s the next blocker to stablecoin adoption?

John Collison:
I think training consumer familiarity is going to be the big one. Right now, one of the places where stablecoin usage is most interesting is in cross‑border money movement, where it solves a real‑world problem. If you want to send $2 from the U.S. to Turkey, that’s not really a thing you can do. There is no product that historically fills that space. Crypto makes that easy. In a lot of creator‑economy use cases or things like that, we’re seeing a lot of pull. What’s interesting is, you get this cycle of the apps getting better and consumer behavior coming along with it. An example of that might be QR codes. The QR code as a technology has existed for decades—it’s a super‑old technology—but it required two things. It required Apple to natively include QR‑code scanning in the camera app on the iPhone—so if you just point the camera app on the iPhone towards a QR code, it did the right thing. That happened in the late 2010s. Then with COVID, there was lots more focus on touchless experiences and things like that. I feel like in the U.S., people really got familiar with QR codes during COVID.

So again, that technology existed for a long time, but it required the consumer apps to get good in the form of the iPhone supporting it, and consumers to be wired to expect it and use it. I feel like we still don’t have both of those for the stablecoin use case. Because we wanted to encourage stablecoin adoption, we have our coffee station downstairs on the first floor, and you can pay with crypto. It’s a super janky experience, because you always run into a bug or something like that. That will get good. As consumers get familiar and as all the app experiences get really smooth, that feels to us like it’s going to drive a sharp rise in usage. I don’t know—does that scan?

Brian Armstrong:
Totally. I think you’re right. It’ll be interesting to see if QR is what’s needed here, or if it’s a tap‑to‑pay NFC thing.

John Collison:
I’m using QR just as a metaphor for this wave of—

Brian Armstrong:
Sure.

John Collison:
But I agree. Tap‑to‑pay NFC feels like it’d be a very plausible ingredient—some better OS support. I guess now with the new iOS rules around NFC, maybe you can use the NFC APIs in the Coinbase app, question mark?

Brian Armstrong:
We’ve seen some good demos of tap‑to‑pay. I think that part is opening up. The Secure Enclave on the device is another story.

John Collison:
You still can’t get... Yeah.

Brian Armstrong:
We’ll keep working on that one. I think you’re right. We debated mailing these stickers to flood the zone of a city to test it out—seeing which merchants want to accept it. A lot of the use case is happening across borders, happening in more internet‑native applications where crypto is the only way to pay. Because it’s a global group of people coming together into some community. It’d be a little weird to use one country’s currency. Brick‑and‑mortar will probably be a late adopter, is my guess, but still, they don’t want to pay 2%–3% on the fees, they can save money, and give some of it back to the customer too.

John Collison:
Yeah, I agree. It feels like the digital use case is where it’ll roll out first, and we’re starting to roll out in more and more of the Stripe Checkout experiences—pay with crypto as an option. There are a lot of brick‑and‑mortar retailers that accept Alipay or accept JCB. Obviously that’s a niche use case—accepting JCB, the Japanese card brand, in the United States. But it’s a large enough constituency to be worth it. It feels analogous here once you get above some minimum penetration, not everyone’s paying with it, but it’s just not worth not accepting it.

What do you make of all the banks starting to embrace crypto? Jamie Dimon said previously, “Bitcoin is a fraud. It’s worse than tulip bulbs. If I was the government, I’d close it down now.” And now of course they’re launching JPMD, their tokenized dollar. It’s been an interesting reversal. I’m curious what you make of all that, across banks generally.

Brian Armstrong:
Ultimately they’re responding to customer demand. If their clients want it, they’re going to support it. I feel like the banks have had this back‑and‑forth with crypto, where they’ll say, “We’re not so sure about Bitcoin, but we like blockchain technology. Maybe we could make a closed network for interbank settlement.” Or maybe they don’t like paying SWIFT fees. They are excited about it in the abstract, and they’ve run a lot of pilots, but I haven’t seen them full‑on embrace it. It’s a little bit of that Innovator’s Dilemma—Clay Christensen—where culturally it’s so hard to get an organization, where they’re making tons of money off the traditional system, anything they do in crypto would just be this fraction of a share but it comes with all this risk and complexity. They don’t get a lot of people who come in with that mindset. They don’t have a lot of DNA inside the company that wants to build crypto.

Now I think, just like The New York Times when the internet came out, some of these local newspapers are not going to adapt to the new system and they’re going to fade away, but a lot of the best ones will adapt to it. In the same way that there are websites now for these media companies, the banks are going to embrace crypto. The payment companies—Visa and Mastercard—are running good pilots on stablecoins. Obviously I think Stripe jumped in with both feet, which is really smart, and you’ve caused a lot of other people to wake up and do the same thing. I think the smartest ones will adapt. That’s good. From a Coinbase point of view, we increasingly want to be people’s primary financial account. As crypto eats financial services, for some of our customers we can be a bank replacement, and they can manage their trading, their payments, their direct deposit, they have a credit card, they have a loan. A lot of them, by the way, won’t even really care that it’s crypto. They just want the best financial services. If it’s the cheapest way to send money to their family overseas, or they can get the best rewards on this card, or whatever it is, that’s what we ultimately want to provide to them. The banks will have to compete in that new environment. Are they going to be an infrastructure layer, powering some of these new fintech apps that the next generation of kids thinks of as their primary financial account? Or are they going to adapt to the new world and build their own apps or embrace crypto? As I said, the smartest banks will do it, and many will get left behind. That’s good. That’s free‑market competition.

John Collison:
Totally. And companies don’t disrupt; consumers cause disruption by voting with their feet to use something else. Has anyone impressed you on the banking side of things, just in terms of being forward‑thinking?

Brian Armstrong:
Yeah, a number of them. I actually think Jamie Dimon is a great leader—very smart. His comments on Bitcoin—

John Collison:
He’s the GOAT. Yeah.

Brian Armstrong:
—I don’t really agree with, but he’s a great guy, and we work with them a lot. I think Santander has been really great. There are some banks that have really embraced crypto, like Citizens Bank. Cross River and Silicon Valley Bank had a love‑hate relationship. There’s a bunch.

John Collison:
What does Coinbase becoming people’s primary financial account look like?

Brian Armstrong:
It means that we have a bigger responsibility and share of people’s financial lives. It’s not just coming in to trade crypto. It’s like, “Okay, maybe I started with that, but I have a lot of crypto here. Okay, I can get the best rate on a loan to do a mortgage because I have a lot of assets here as collateral. Or it could mean that I can use a bitcoin credit card and get 4% back in bitcoin as rewards when I spend. Or it could mean that it’s very easy to send money overseas instantly for less than a cent.

We need to do a good job of leveraging the benefits of crypto to update the financial system, and not make it just crypto for crypto’s sake. The early users of Coinbase were crypto diehards, and that’s great. We’ve gotten to maybe 6% or 7% of the world having used crypto now—like the early 2000s for the internet. We now need to get to about a billion, or half the world, eventually using it, to really increase economic freedom globally. It’s got to be about something more than the technology. It’s got to be about: it’s faster, it’s cheaper, it’s better, it helps me get done what I want to get done.

John Collison:
The reason I ask is because I observe the growth of neobanks outside the U.S. Nubank is taking over in Brazil. Revolut is the fastest‑growing bank in Europe. In the United States, the largest consumer banks are basically the same—very similar, recognizable set of characters from the 1970s. Maybe that’ll still be true in ten years’ time, or maybe things will really have shaken up. I’m very curious what the neobank revolution in the United States looks like.

Brian Armstrong:
That’s a great way to phrase—I probably should have phrased it that way when I was describing it. Some people are calling them stablecoin neobanks, or superapps. There are a lot of terms you could come up with for this. I think that’s our opportunity. By the way, as a technicality, I don’t think we’re going to go get a banking license, because—I don’t know how deep people want to go on the details here—but the core thing a banking license allows you to do is not have all the money there. It’s called fractional reserve.

John Collison:
Right.

Brian Armstrong:
Which is an interesting business model, but it also comes with such heavyweight regulation that it makes it much harder to really innovate on products and ship quickly. We actually don’t want to be a bank.

John Collison:
You want to be fully backed, as opposed to doing fractional‑reserve banking.

Brian Armstrong:
I want to be—100% reserve, not fractional reserve—which is actually safer for the customer because you can’t have a bank run. You can even store the assets, like the stablecoins especially, backed in U.S. Treasuries and things that are bankruptcy‑remote, et cetera. There are really strong protections you can put in there. That’ll allow us to continue to innovate and provide the best customer experience. Ultimately, I think it could be a bank replacement for many people—or let’s call it a... I like your term. The U.S. needs a neobank, since the rest of the world seems to keep having them.

John Collison:
What is your prediction for the ten‑year rate of growth—average annual growth—in the bitcoin price?

Brian Armstrong:
The rough idea I have in my head is we’ll see a million‑dollar bitcoin by 2030. There are high error bars around these things, but a couple of data points: We’re starting to see regulatory clarity emerge in the United States, which I think is a bellwether for the rest of the G20. You have The GENIUS Act passed now for stablecoins, but this market‑structure bill is being debated in the Senate. Fingers crossed, something could happen by the end of this year. That would be a huge milestone. The United States government now has a strategic bitcoin reserve. If you’d asked me five years ago, that would’ve been vision‑board stuff—someone would’ve said, “You’re crazy. The United States government’s not going to officially hold bitcoin.” But if they do that, which there’s now an executive order to do, I think a bunch of other countries are going to do that.

We’ve seen a lot of interest from sovereigns. Coinbase provides crypto services to about 140 government entities, at the federal, state, and local levels, and internationally. Governments are getting more and more engaged in this. You can imagine the big pools of capital around the world. I don’t see the regulatory thing going away. That was one of the big risks—“Is the government going to shut this down?” I think that risk has been severely diminished. Then, “Is there going to be some flaw found in the Bitcoin protocol?” I think that’s been severely combed over at this point. We need to make sure we upgrade it to post‑quantum cryptography, but—

John Collison:
Aren’t elliptic curves already post‑quantum, or no?

Brian Armstrong:
They are, theoretically, I believe. No one really—

John Collison:
It would be really good to nail that down.

Brian Armstrong:
It’s a little untested, but I believe there is a path where the Bitcoin core team and Ethereum and so on—everyone—they’re looking at proposals now to upgrade to—

John Collison:
Definitely post‑quantum.

Brian Armstrong:
Yes. I think the list of risks is getting diminished, and at the same time, the amount of demand—the amount of institutional money just sitting on the sidelines waiting for this next bill to land. The big institutions I talk to are holding 1% of the portfolio in bitcoin. I’m like, “What would it be to get you to 5%–10%?” They say, “Regulatory clarity. That’s it.” I think we’ll continue to see huge inflows of capital. The ETFs have been huge.

John Collison:
When people talk about the institutions on the sidelines, I think the gold comparison is a very apt one, in that it’s a non‑productive store of wealth—which I mean as a compliment, not an insult. But depending on what you mean by institutions, I don’t think the major sovereign wealth funds or mutual funds hold a lot of gold today, I guess. I don’t know. Doesn’t it seem like you’re substituting the gold buying, but it would be unusual for those large institutions to buy a non-cash‑flowing asset?

Brian Armstrong:
It depends on their strategy, but there is a theory that in a diversified portfolio, you should have some percentage of it—5% or 10%—in commodities of species or whatever. I think there’s a good bet to be... BlackRock has published reports and research on this—how they believe crypto should be a part of every healthy, diversified portfolio at this point, because it has interesting inversely correlated aspects with other assets. It’s changing over time, by the way, as it evolves. I think that we’ll be sitting here in five or 10 years, and most wealth managers—or maybe sovereigns—who hold a traditional diversified portfolio will include 1%–10% crypto.

John Collison:
If you did want to put a fraction of your portfolio into crypto, should you dollar‑cost average into bitcoin—and maybe the other major coins, but not the meme coins—in a market‑cap‑weighted way? Or, what’s the right investment strategy if you want to put some money into crypto?

Brian Armstrong:
Well, with the caveat that I’m not giving investment advice and all that—I think it’s reasonable for people, if they’re just learning about any new thing, whether it’s crypto—if you like it, put 1% of your net worth in—something you’re willing to lose and learn about, right?

John Collison:
But within crypto, what should they put it into?

Brian Armstrong:
Oh. I think bitcoin is a great place to start. We have something called the COIN50 Index, which is the top 50 coins by market cap.

John Collison:
Market‑cap weighted, I presume.

Brian Armstrong:
Yeah, market‑cap weighted. I’m not going to go down and list specific coins, but I think you should hold bitcoin. You can hold the index if you want. Then, go try and use crypto. Spend it at a Stripe merchant. Use your Coinbase credit card, and post some content onto the Base app and start to earn money in crypto. Shop at a store that accepts it—whatever. I think people should go use crypto. There’s a part of it that’s an investment. Eventually, as these stocks get tokenized and people want to get a loan, they’re going to be using crypto without even really knowing it underneath. It’s like they may not know how electricity works, but they can turn on a light switch.

John Collison:
That makes sense. So, we have The GENIUS Act now. What does it mean in particular? Because stablecoins were legal before, they’re more enshrined now, but what can we do now that we have The GENIUS Act that we couldn’t before?

Brian Armstrong:
Okay. I’ll start with stablecoins and then what The GENIUS Act—so, stablecoins: basically fast, cheap, global payments. We can now do it in under one second, one cent, anywhere in the world. That’s unique. There’s no other payment rail that gets all three of those—fast, cheap, global.

John Collison:
This is not a GENIUS Act thing. You’re just saying this is a thing you can now do with stablecoins.

Brian Armstrong:
Yes. I’m getting to your question. There’s no other payment rail that gets all three of those boxes checked. Some are fast and cheap and not global, et cetera. That’s a huge deal. Now, what The GENIUS Act did was it said, “If you want to call yourself a stablecoin and operate in the United States, there are certain requirements you have to meet that make it more safe and trusted.” One of them is, 100% of the reserves have to be backed in U.S. dollars or short‑term U.S. Treasuries. You can’t hold it in other risky things. You have to pass an audit once in a while and prove you’re doing it. It’s basic hygiene.

These are the tactical things, but the bigger‑picture thing it did was put a federal law on the books—a stamp of approval—and said, “This is going to be trusted, legal, allowed, and built in the United States.” That created huge demand from every company that deals with payments, which is 100% of them. It said, “We need to have a stablecoin strategy. We need to figure out what we’re going to do here, because it’s now going to be a thing. The U.S. government blessed it.”

John Collison:
I’ve noticed so much more interest in the past two or three months.

Brian Armstrong:
Oh, yeah.

John Collison:
It’s incredible. You’ve probably noticed something similar.

Brian Armstrong:
Yeah, it’s like a gold rush. Everyone’s coming in, trying to figure out what is going on with this and how it can save their company money. There are lots of companies trying to pay out developers in markets all over the world. In the U.S. and Europe and a few other countries, it works pretty well. Then there’s a whole long tail of countries where people just don’t have access to good financial services. It’s kind of a black box—how much money shows up on the other side of this payment rail, with huge fees taken out. If you have a global payment rail that’s fast and cheap, it democratizes access to financial services, where anybody with a smartphone is on a more level playing field. Their wealth can’t be taken away from them, eroded away via inflation. It’s an incredible tool for progress and property rights and sound money—economic freedom, basically. That’s what opened up with stablecoins. Now we need to get the market‑structure bill done, which is all the non‑stablecoin crypto assets.

John Collison:
Got it. So there’s this “what’s a security, what’s not” and all that kind of stuff.

Brian Armstrong:
Yeah.

John Collison:
You must have some funny stories from the passage of The GENIUS Act, because it was such a winding—

Brian Armstrong:
Rollercoaster, yeah. One thing I realized was: For a long time we were trying to make progress in D.C. and advocate for legislation, and nothing was happening. People educated me about this a bit—it’s actually kind of rare for Congress to act. In fact, somebody told me, “Congress is really good at doing two things: nothing, and overreacting in some crisis moment.” We realized that at a certain point we had to generate political will to do this. We had 50 million people in the U.S. who had used crypto, and we decided, “Let’s try to get them organized.” We funded this 501(c)(4), called standwithcrypto.org, and we got 2 million folks in the U.S. to raise their hand and say they wanted to elect pro‑crypto candidates.

I remember talking to our policy team and saying, “Let’s put a scorecard, A to F, of every politician in this upcoming election last November.” Our policy team—their whole job is to build relationships with politicians—said, “Why don’t we put a list of our crypto champions?” I said, “No, I want to put a list of the crypto enemies too, with an F. Who has an F?” I could see them nervously perspiring. I said, “Let’s put the scorecards. Someone needs to win an election, someone needs to lose an election, because of the crypto vote.”

John Collison:
Crypto famously, in terms of how it shows up in D.C., is much more pugnacious than, say, tech was historically—where it was, we have our friends and we have our enemies and things like that. So did that and Stand With Crypto and Fairshake... I guess Fairshake was a more relevant one in some of the recent battles. Did that come from you, essentially—that we’re not going to be pussyfooting around here, we’re just going to make clear how we feel?

Brian Armstrong:
There were a lot of people involved, so I can’t claim full credit. But what I saw shift in the crypto industry was a lot of the traditional policy advice we got from people who worked in tech said, “You need to go in there and build relationships, so be the good cop. Then form a trade organization who can be bad cop. They can fight the battles and write the nasty op‑eds...” For whatever reason, the people we kept getting into these trade groups wanted to also be good cop. I kept thinking—

John Collison:
When does the bad cop arrive?

Brian Armstrong:
—Who’s going to get into the political bloodsport on X and be tearing down these people that are doing bad things? They just wanted to have polite meetings behind closed doors. I realized as the election was coming up, “If they don’t do it, maybe no one’s going to do it, so maybe we should do it.” I don’t think we did anything too crazy, but for D.C. standards, this was crazy. We came out and said, “We’re unequivocally pro‑crypto. We don’t care if you’re on the left or the right. We want to elect pro‑crypto candidates. And by the way, we want to get anti‑crypto candidates out of office.”

This really blew people’s minds in D.C., because everything there is partisan. There was a feeding frenzy coming up to the election where I was getting angry phone calls from both sides: “How could you have given money to this person?” We’re like, “Well, they’re pro‑crypto.” Then I’d get a call from the other side: “How could you have written to that person?” “Well, they’re pro‑crypto.”

John Collison:
You’re really a single‑issue voter. Yeah.

Brian Armstrong:
Yeah. On the one hand I was like, “Maybe we’re just pissing off both sides and we’ll have no friends after the election.” On the other hand, if you’re taking flak, you’re over the target. We’re explicitly an apolitical company, but we are unapologetically pro‑crypto. To me, it was very consistent. By the way, I had to remind many of them: apolitical does not mean 50/50. It means we will support a pro‑crypto candidate regardless of what party they’re in. It’s not necessarily going to be 50/50. It might be 60/40 one way and 60/40 the other way the next election, or whatever.

That was a big mental shift, and it was kind of contrarian. It helped elect the most pro‑crypto Congress. The work we did—along with a bunch of other people; I don’t want to take credit for it. A bunch of other companies worked on this—set up the most pro‑crypto Congress, which is now getting this legislation passed. We made it clear to people in D.C. that, if you’re anti‑crypto, there’s no constituency for that. The American people want crypto, and if you are pro‑crypto, it can help you get elected. It’s just good politics. That was a novel idea.

John Collison:
Do you have issues now, once you get through the market‑structure bill—GENIUS Act just happened—do you have issues left? Are you done going to D.C.?

Brian Armstrong:
Yeah. The upside to all of this—there was a lot of downside. In the last administration, they were trying to kill the whole industry. But now that we’ve gotten to build a little bit of knowledge around policy, which can be dangerous—these things can go either way—hopefully we’ll get market structure done. It does start to make me wonder: what else can we do to help update the financial system? Our mission is around increasing economic freedom, so an example would be: I think the accredited‑investor laws are kind of unfair. Only rich people can invest money to get richer. That seems regressive. There could be a financial‑literacy test, instead of having a certain net worth or income, to become an accredited investor.
I like this idea—I’m very excited about this idea—of special economic zones. They’ve worked very well in China with Shenzhen, and in the UAE and these places. We have so much regulation; why not have a sandbox for one area to test out new ideas, where you could have one around crypto or biotech or drones or supersonic aircraft? I think it’d be great to see if we could get, I don’t know, 10 pieces of federal land in the United States set up as different special economic zones. There’s a company called Prospera, which we invested in, which has a prototype set up in South America, but they’re trying to do this now in the United States. There’s a lot of different areas. Do you have any favorite pet policy issues that you would push for in D.C.? Probably something with aviation.

John Collison:
Yeah, I was just going to say... Well, this one happened, actually. I’m super excited about these new MOSAIC rules that happened and—

Brian Armstrong:
What is that?

John Collison:
Okay. In general, there’s been a lot of discussion around the deregulatory agenda of the current admin. Getting rid of outdated regulation seems great, but you need to actually go do it and pass the new rules. They just did that in the aviation sector, where you can’t just build a plane and sell it to the general public—you have to go through a whole bunch of FAA approvals that were both a very cumbersome process, took multiple years, and also a very prescriptive process, where you can only have these kinds of engines or something like that.

Famously, electric aircraft couldn’t be certified, because they said an aircraft engine burns gasoline—does this and that and the other. They just came along—Sean Duffy announced last month—this new regime for certifying aircraft that is radically simpler and, I think, will allow for much more bottoms‑up innovation in that sector. That would’ve been on my list, but it happened. I think we’ll see much more innovation in the light‑aviation sector, which obviously the U.S. used to be a leader in, and stagnated over the past decades.

Brian Armstrong:
Yeah. I think you might’ve seen, by the way, there was an executive order instructing the FAA to allow supersonic aircraft over land in the U.S. that did not create an audible boom, by some decibel threshold.

John Collison:
Yes, yes.

Brian Armstrong:
That was really exciting. I’d love to fly places much faster. Another one I think would be great: It takes ten years and $2 billion, apparently, to get a drug to market through the FDA. They have three phases of trials around safety and efficacy. I think it’d be amazing—if you pass phase one for safety—why not let a doctor prescribe it, especially to somebody who is an end‑of‑life patient who has no other options? You know it’s safe; you just don’t know if it’s effective. Let a doctor make that call. This would give you data much sooner on getting these drugs to market, where every year that goes by, there are people dying. There are so many things like that that could be improved by making sensible deregulatory movements.

John Collison:
I think everyone agrees that the accredited‑investor rules are a bit silly, where they are both exclusionary if you don’t meet the minimum net‑worth test, and, as we have seen, wealthy people are not necessarily savvy when it comes to making investments. The list of Theranos investors was a Who’s Who of various wealthy and accredited people. At the same time, when it’s just a free‑for‑all, you get a bunch of scams and frauds. The U.S. has some of the best capital markets in the world. As I look at what’s happened in the crypto world—where various pockets are more liberal—I think the whole phenomena of rugging and people looking to engage in zero‑sum behaviors are some of the worst things we should not allow to perpetuate. As you think about replacing the accredited‑investor rules, what would your framework be for liberalizing investment and yet keeping the best things we have today—which is, I think, really high standards of propriety and really strict rules for honesty and fair dealing?

Brian Armstrong:
That’s a great question. Fraud—you’d want to prosecute to the full extent of the law. One thing you could do is have requirements about disclosures and say, “If you want to raise money for this thing, you have to provide the following information. If you materially misstate something, especially intentionally, and you defraud investors, you should go to jail for that.” It doesn’t mean only wealthy people should be allowed to invest, so I think that’s a nice way to divide up those issues. We also don’t want to create this false idea that the government’s stamp of approval means it’s a good investment. There are many publicly traded companies—in theory blessed by the SEC—that went down 85%–90% in the last few years. Some of the biotechs—I won’t name companies—but you can lose your shirt in the public markets with these officially—We need to teach a sense of personal responsibility in this too, which is: there is no return without some amount of risk, and the government is not here to tell you what is a good investment or not. You still have to make that judgment yourself. They can help by saying, “The information this person’s giving you is true. If they’re lying to you, there’s tort law for that where you can go have damages.” I probably lean a little more free‑market on this than most people, but the best consumer protection sometimes is competition. If you have a car company and it’s a bad car and it keeps breaking down and it’s expensive, in some ways the best solution is to have another company come and try to compete with them and provide a better alternative. A lot of people are actually very street‑smart—even if they’re not financially literate—they know what a scam is. They know that their three friends told them that that thing is not good.

John Collison:
Do you guys try to do anything on the Coinbase product side of things with some of the coin‑hype scams that happen? Or do you say people are grownups and they should evaluate their own investments?

Brian Armstrong:
We’ve had a lot of debate about this. It’s a little bit like the app stores, or maybe even Amazon, where—let’s say—someone’s trying to sell a fraudulent product on Amazon. They probably want to get rid of that. But let’s say that it’s a two‑ or three‑star product, and some people like it and a lot of people don’t, but it’s not a fraud. You want to allow people to make their own choices and see it got two stars—but you can buy this if you want. Maybe the provider can clean it up too, and you tell them… We’ve tried to take a similar philosophy where we want to list everything that is legal—and frauds would be illegal—and provide information to the customer to help them make better decisions. We’ve tried variations of this; we haven’t quite nailed it yet. I think about an on‑chain review system or reputation score. We released an API that allows anybody to look up a crypto address—which could be for an asset or a person—and have an on‑chain reputation score. It’s an early prototype. Eventually you’ll get a FICO‑score equivalent; you’ll get an Amazon‑ or Yelp‑rating equivalent. You want to make sure—if I’m trying to send a few USD Coin to John Collison for the pint—I need to make sure it’s the real John Collison, not an imposter squatting on your name or something like that. Those are some of the tools I think we can, in the private market, build that would help this, and let the government go after the fraudsters.

John Collison:
That makes sense. You mentioned crypto’s product‑market fit in emerging markets, and there is a very strong product‑market fit there that’s actually not new. People in markets where there’s historically been a very high rate of inflation have traditionally sought to get money out of that currency. People talk about buying hard assets as one approach, trying to get money in dollars. It’s already a behavior to try and get money out of that currency and into something that they think will hold its value better. This was not invented by crypto; this has existed for many decades. Maybe countries where there is a black‑market exchange rate—that’s a good indicator for “There will be demand there.” The governments of many of these countries will not always approve of this, where they would prefer that all money be kept in the country’s currencies, and there’s a gap between what the government wants and what the people want. Where does this go? Who wins?

Brian Armstrong:
You’re right—we are often walking this balance when we go into a new country, where we really want to work within the existing system, and we often will go get a license, set up a local entity for our regulated financial‑service products. I’ll come back to our self‑custodial wallet in a moment. There are usually portions of the government that are hesitant around crypto. Sometimes it’s the central bank, but there are other parts of the government that are very pro it. They want to digitize the country and create economic opportunity. You talk to different hands of the government, you’ll hear different things—but the people unequivocally want it.

In many parts of the world, we’re able to work within that system and offer what people want. Other parts of the world, it would be difficult to impossible for us to go in and create a regulated financial‑service business, because that country doesn’t have a clear regulatory framework. They only give out bank licenses to their cousins and friends who bribe them—which would be illegal for a U.S. company to do. You just can’t do business very effectively in many of these places. So we also have a self‑custodial wallet that is not regulated as a financial‑service business, because we never take possession of customer funds. It’s regulated more like a software product, like a messaging app. You can just launch it—

John Collison:
As if you kept your keys in 1Password, which is ultimately the software product.

Brian Armstrong:
Yeah. That’s allowing us to go into some of these other markets. You asked where this will go. In some of these markets, I’d say there’s the top five or 10 fiat currencies—government currencies, the majors—those will probably stay around. I don’t think they’re going anywhere.

John Collison:
U.S. dollar, Swiss franc, everything like that.

Brian Armstrong:
Yeah. But the long tail—the other 150 or so government currencies—I think they probably should get replaced. They’re not doing a good job. They’re often abused. They’re eroding the rights of the people. I think there’s a good case to be made that those should get replaced by bitcoin and USDC, et cetera. It was Ecuador, something like that, that pegged their currency to the dollar too. It’s a way of doing that in all these countries, almost as an act of civil disobedience, frankly. In some places like Venezuela, you may be technically breaking the law by introducing a self‑custodial wallet that brings dollars to the people. I think I’d be okay with that. That’s an act of civil... People are in terrible conditions there, due to the government stealing their wealth via inflation. I think that is a good thing.

Obviously a digital dollar is unequivocally good for the United States and keeping the reserve‑currency status and demand for Treasuries. It’s also a good thing for bitcoin to be used in the United States, because the democracies around the world are having a problem with deficit spending. It’s this issue that, to get elected, you can promise more free stuff, and that drives up the cost. How are we going to have the discipline to balance budgets? I think bitcoin is part of the answer to that. It’s a check‑and‑balance on deficit spending where, if it gets too out of control, people will flee to bitcoin in times of uncertainty. If deficit spending is kept under control, people will continue to use that fiat currency.

I don’t want to be hyperbolic about this, but I think Bitcoin is in some ways extending Western civilization and the American experiment. If the U.S. is going to lose the reserve‑currency status—which I hope it doesn’t, because I am an American; I believe America is a good thing for the world—but if we lose all discipline and we’re going to lose the reserve‑currency status, I’d rather people went to bitcoin than to the Chinese yuan. Thank goodness we have bitcoin as a check‑and‑balance in this new economy.

John Collison:
The deficit worriers look at the deficit math and say, “Interest payments are now greater than defense spending.” How does the U.S. avoid losing reserve‑currency status?"

Brian Armstrong:
One part of it is: don’t fight away the... The debt‑to‑GDP ratios are the thing to watch usually, I think. I forget where we’re at now—150 or 170 or something—but if you look historically, maybe when the British lost it or the Dutch, they were in the 200–250 range. So we’re within sight of a pretty dangerous threshold, historically. It’s hard to see where the political will is going to come from to do that, and I hope that bitcoin could be part of that solution. We’ll see.

John Collison:
Why haven’t non‑dollar stablecoins worked? The share of global stablecoins that is U.S. dollars is 95%+. It’s much higher than the U.S. dollar’s share of currency. Why is that and will it persist?

Brian Armstrong:
If you’re going to have permissionless access to anything, you’re going to use the most trusted one—the reserve currency—so that could be why. Europe, to their credit, did put... Sometimes Europe’s the leader on regulation, which is not necessarily what you want to lead on. But they did get out a regulatory framework before the U.S., to their credit. There is a Euro Coin—

John Collison:
But it’s very small.

Brian Armstrong:
It’s very small. It doesn’t have a lot of traction. I don’t know. I think the dollar is solving the need that people have. There’s actually one other thing that’s even smaller, which I think has interesting potential, called a flatcoin. I don’t know if you’ve seen these, but instead of tracking one‑to‑one backed by a dollar—which of course has some level of inflation between 2%–5% a year—a flatcoin tries to track CPI, which is the consumer price index, to maintain its purchasing power. So if a Big Mac costs—

John Collison:
Like TIPS?

Brian Armstrong:
Yeah, inflation‑protected—

John Collison:
Securities. Yeah.

Brian Armstrong:
Yeah. If you could buy a Big Mac for a dollar today, ideally with a stablecoin, 10 years from now you’d also be able to buy it for a dollar. There’s one that’s built by this company called Ampleforth, called SPOT, and it’s been tracking a 2019 dollar since 2019, and it’s now worth about $1.26. It went through crazy ups and downs; it’s actually been pretty flat. Economists will often debate, “Do you want to have 2%–3% inflation a year to incentivize people to spend their money?” But it’s nice to have one that maintains its purchasing power, for putting contracts together and pricing things that you want to be the same price in the future.

John Collison:
Yes, yes.

Brian Armstrong:
Whoa.

John Collison:
Whoa. Okay.

Brian Armstrong:
What was that? How much is the center center?

John Collison:
So the bullseye is 50, so 78.

Brian Armstrong:
78?

John Collison:
Okay.

Brian Armstrong:
Well—

John Collison:
18, 23, and four is 27.

Brian Armstrong:
Okay.

John Collison:
We did it. TradFi wins again.

What have you learned from Balaji?

Brian Armstrong:
I was about to say, for anyone who doesn’t know who Balaji is, you can probably—

John Collison:
They probably know who Balaji is. He’s Balaji.

Brian Armstrong:
He is one word.

John Collison:
It’s like Madonna. Exactly.

Brian Armstrong:
Yeah, one name. Balaji is a brilliant guy. He’s probably, I don’t know, in the top couple of the smartest people I’ve ever met in my life. He’s very out there. He’s truly an original thinker. He was briefly the Chief Technology Officer of Coinbase. He came in through an acquisition and did some amazing work. He taught me how to manage a totally different type of person, I would say. Balaji is kind of unmanageable. He’s what some people might call a free radical within an organization, where he really bounces around and he’s absorbing vast amounts of information—even things that aren’t his responsibility, maybe he’s not even supposed to be doing.

John Collison:
Was explicitly told not to do them.

Brian Armstrong:
Yeah, was explicitly told not to do them—does it anyway. Occasionally he would come back to me with incredible insights. I’ll give you one funny Balaji story. At one point he came back to me and said, “These are all the salespeople that are making more revenue than their salary, and these are all the people that are not.” The first thought I had was, “You’re not supposed to have access to anybody’s salary. How did you get that?” He said, “Don’t worry about it. I found it in some database—somewhere I wasn’t supposed to have access to.” The other thing I asked him was, “How did you connect all that up? Because last week I was asking the data team, ‘Can you connect up Salesforce to all of our salary data and let’s run some reports and have some more accountability?’” He said, “Oh, well, I couldn’t sleep this weekend, and I just knew something felt off, so I had to code it up and put it all together.” I said, “Okay, give me a little bit of time to go verify this.” He came in on Monday; he looked like he hadn’t slept the whole weekend. He’s kind of this mad genius. By the way, I went and ran it down with the actual data team, and it took them three weeks to replicate what he had done over the weekend, but it came back and he was 100% right. He continually did things like that, and he has incredibly high disagreeableness—which I learned from him as well—where he would go into a team and say, “Why is this not functioning well?” He would suffer no fools. He would not be afraid to go in there and turn half the people on a team—whether he had the permission to fire them or not. Sometimes he would just torture them to greatness or whatever.

He is a very contrarian figure. About once a week someone would come into my office as CEO and say, “I can’t work with Balaji, and he’s causing so much collateral damage.” I’d say, “Yeah, but he’s also generating an enormous amount of value, and I need you to learn how to work with him. I realize it’s difficult, but it’s creating a very positive outcome across all these areas.” I knew that it wasn’t going to last forever because it was incredibly disruptive, but I learned from him how to come and be a turnaround CEO when needed, if I can call it that. In the past I was opting a little more toward trying to be liked instead of being clear about what we’re doing and where we’re going and what the bar is. He helped me be a better CEO and have a little more disagreeableness.

John Collison:
He encouraged you to let Brian be Brian—listen to yourself more. I liked the tweet a while back—what was it? The three scariest words in the English language are, “Balaji was right.” I think it was about some COVID prediction or something like that.

Brian Armstrong:
Yeah. He was frequently very early on his predictions. I think that—

John Collison:
Are there predictions he had that you thought were crazy at the time, or worldviews he had that you thought were crazy initially, and now you’ve come around?

Brian Armstrong:
Yeah. He was early warning me about the activist contingent within the company, and I was like, “This is not a big deal. People are—”

John Collison:
Just the apolitical‑company stuff, or what—

Brian Armstrong:
Yeah, which led to the mission‑first post. He was telling me about that years before, and I thought, “This can’t be a big deal.”

John Collison:
Okay. On that—in 2020 you released this blog post saying that Coinbase is a mission‑oriented company, and take all your other issues you might have and don’t come to them with us—and work somewhere else if you don’t like it. You had a buyout offer if people didn’t want to work at such a company, and a few percent of the company took the buyout. This was a big deal at the time, and everyone was talking about it. Part of what’s interesting is how much the industry has moved, where Coinbase was—I don’t want to say a pariah, but definitely notable at the time for being so stringent on this point. Recently, literally Google was putting out a blog post saying, “Leave your politics at home. We’re just focused on organizing the world’s information and we don’t want to hear from you about all your other stuff.” It feels like Silicon Valley has really shifted on this.

But one question I had is: you talk about being an apolitical company. I feel like apolitical is a convenient way to frame it, but it’s maybe more something like: there’s actually a very narrow set of views that you have to agree with to work at Coinbase. You have to believe in financial freedom, or you have to believe that crypto is a good thing for the world, and then everything else we’re just not going to engage with and we’d rather you don’t bring it up. Maybe that’s a technicality. Maybe that’s a nuance, but—

Brian Armstrong:
No, you’re 100% right. As I was writing that blog post, originally I was thinking of calling it an apolitical company, but I changed it to mission‑first. You’re absolutely right. We are unapologetically pro‑crypto; we’re about economic freedom—and that does have a policy angle to it, right?

In fact, we are trying to get massive change happening in D.C., with legislation passed. We are very unapologetic about engaging in politics on our mission. Everything outside of that—which sometimes has edge cases you have to go figure out—but everything outside of that, we try not to bring to work. It’s a mission‑first policy, as opposed to being explicitly apolitical across everything. That was an evolution, and we did take a lot of arrows for doing that, but I think it was a good move.

John Collison:
I presume you think—now looking back on it—you’re very glad that you did it.

Brian Armstrong:
Very glad. I was failing as a leader before that, because I was not creating clarity in the organization. We were having... Some people thought our mission included social‑justice activism on police brutality; some people thought... I clarified it and I said, “We’re all going this way. I’m sorry I failed to create clarity beforehand, but if you don’t want to work here at that company anymore now that I’m making it clear, here’s the exit package.” We had debates internally. People were terrified that I was about to do this; some people tried to convince me not to do it. Some people said 50% of the company will quit and all this—and it turned out to be 5%.

Sometimes your fear gives you bad advice. I got more confidence to do it by talking to some of our employee groups, hearing from them directly, and watching some videos of people like Lee Kuan Yew and Ronald Reagan—in similar situations where... leadership‑inspiration moments. I’m curious, where does Stripe land on this topic, if you can talk about it?

John Collison:
Like I said, I think the norms—or the expectations—have very much shifted, in a productive way, I think, where people get that companies have a thing they’re trying to do in the world, and it’s hard enough to do that thing. It’s hard enough to get The GENIUS Act passed. It’s hard enough to make some impact in the world. The U.S. is a pretty big place, and there’s a broad variety of views on a whole bunch of social issues. The classic—if you were going to a dinner party in the 1950s—the advice would’ve been that you leave politics and religion aside as good work-dinner‑party conversation topics. This rediscovering of old wisdom—I think that it’s hard enough to do one thing.

It’s not that these aren’t very worthwhile endeavors. Both of us have pursuits we’re doing outside of work in various regards. But it’s really hard to get an entire company aligned around some cause, and so you want to spend that capital very carefully. Yours came at an interesting moment in time where there were a lot of expectations of companies to be broadly getting involved in a bunch of issues. It was notable—again—because that was such a popular idea at the time, but I think it’s become almost received wisdom at this point.

Brian Armstrong:
Yeah.

John Collison:
And yeah, I think we never went super far into the broad set of issues, but I think I broadly agree with you on companies mostly sticking to the knitting.

Brian Armstrong:
Yeah.

John Collison:
It strikes me that Coinbase, from a product point of view—when you’re in the early days of an industry, you kind of have to do everything. Microsoft—in the ’90s—was doing operating systems and applications, and they made some hardware—remember such as the Microsoft Natural Keyboard—and Flight Simulator—

Brian Armstrong:
Clippy.

John Collison:
—and, exactly, Encarta, everything. If you look at OpenAI in AI, they’re doing consumer, they’re doing B2B, they’re doing Sora, the video model—again, because AI is so fast‑changing and rapidly emerging. Similarly in crypto, you guys have the institutional product, you have the stablecoin with USDC, you have the brokerage, you have your L2 with Base, and you’re playing everywhere. That feels correct to me in a very fast‑moving space. It also seems like it would challenge focus, where everything is a good idea, because it could work in this fast‑moving industry. How do you focus, how do you resource things? How do you decide what to cut? Again, it’s easy with Stripe, because we’re trying to stick to the knitting. We’re trying to make the payments infrastructure good. But your world is a more open‑ended map.

Brian Armstrong:
It all depends how you count. I mean, you guys have a lot of products in the payment space.

John Collison:
I was simplifying Stripe for the purpose of the question.

Brian Armstrong:
Yeah. A lot of this comes down to the founders. In my case, sometimes I can’t help myself. I have so many ideas. The company usually is pushing back a little bit—“We need more focus.” I’m like, “Okay, you’re right. We should focus.” There is virtue in focus. There’s also virtue in having multiple revenue streams, because when one thing’s up, another thing’s down. There’s virtue in having some small teams working on a variety of things. USDC and Base started off as three–five‑person teams. USDC will probably add $800 million of revenue to our bottom line, and I would’ve never guessed that in the past year.

John Collison:
Would you actually not have guessed that? Was it not obvious at the beginning?

Brian Armstrong:
It was not. We have a system internally where twice a year employees can pitch what venture bets they want to do. We’ve tried to construct it so you don’t need a unanimous yes. In most companies you need your boss, your boss’s boss, all the way up to the CEO, to say yes to green‑light something—which means you get one no, you’re out. We try to create it a little more like internal venture capital. You can go to any of the product‑group leads, and there’s a handful of smart, hand‑picked engineers, and then myself and a few other people. If you get any one of us to say yes, out of their budget, to fund it, you’re green‑lit. It’s like pitching a bunch of internal venture capitalists.

What that means is, we get more of this risk‑taking culture. We try a lot of ideas. Some of them don’t work and we have to shut them down. That’s actually hard—how do you have the guts to cut? In a startup you’d run out of money and you can’t raise the next round. We have to mimic these rounds of investment. But once in a while they completely defy odds. I’ll tell you a secret: I actually voted no on the USDC project. I read it and said, “Ah, it’s just not really decentralized as much as I would want.” I had some reason in my head. Thank goodness somebody else on the team voted yes to fund it out of their budget, and I was totally wrong. I use that as an example a lot—where the best ideas don’t have to come from me; they have to come from everybody in the company, and we have to try more ambitious ideas.

John Collison:
How about Base?

Brian Armstrong:
Base is another example that was a venture bet. It started as a small thing where Jesse came to me and said he wanted to spin up a layer‑2. I didn’t have any specific idea about how to do it. The only thing we did was, we funded it and shielded it, because new ideas are fragile within big companies. I think one time you told me a great phrase: “You don’t want to give it the big‑company bear hug of death.” The only thing I did—to my credit—was I may have protected it, but I didn’t have any idea really what Jesse was working on for most of that. He iterated through three or four ideas, and Base was the fourth thing he came up with. It finally started to work and became the number one Layer‑2 solution on Ethereum, and now it’s the Base app that just came out.

As CEO, it’s not so much my job to come up with the next great idea. It’s to create the right environment where good ideas can happen and be nurtured, and the bad ideas eventually do get shut down. We have some discipline around that. The hardest time to work on new venture bets is when the core is under threat, or it’s a downturn. And the core is always under threat.

John Collison:
That’s why it’s the core. It’s like a successful business that other people want.

Brian Armstrong:
Right. This makes our decision‑making—I think it’s a very healthy decision‑making—but it’s a healthy tension in the org, where usually people are saying, “The core is underfunded; it’s under threat. How can we allocate more? I want to take resources from that thing, which is not generating revenue yet.” Not always, but sometimes I’m on the other side saying, “I want to fund the core. We also should have some percent—10% of our resources—going to these venture bets,” because five, ten years from now we need to have the next chapter coming out. You see sometimes the founders versus operators dynamic. If you have too much founder energy, they sometimes blow the place up. If you have too much operator energy, they’re just boring and they never do anything innovative ever again, but they’re very efficiently run.

John Collison:
It’s like a series of layered S‑curves.

Brian Armstrong:
Yeah. I’m in a lucky spot at Coinbase with Emilie Choi as the president, and she and I make a great team. We can both do it either way, but the natural—

John Collison:
It’s like a natural happy place.

Brian Armstrong:
Yeah. I try to be a little more founder—provide the risk tolerance, the venture bets. She makes it a well‑run company, pushes back when she’s like, “You know, Brian, if we just focused on the core in Europe, it might generate another $1 billion of revenue.” It’s a great combo.

John Collison:
Do you use prediction markets anywhere internally in how you run the business?

Brian Armstrong:
Not yet.

John Collison:
Shouldn’t you if you’re crypto‑pilled?

Brian Armstrong:
Yeah. We are integrating—working on prediction markets. We haven’t gotten full clarity. This new CFTC chair hasn’t been confirmed yet. The U.S. ones—U.S. citizens are not allowed to use the on‑chain ones yet. The ones that are approved to work in the U.S.—I believe each market has to be approved by the CFTC.

John Collison:
Oh, wow.

Brian Armstrong:
So if you had some fun internal pet project... We just don’t have the resources to go get everything. Hopefully it becomes lower friction at some point.

John Collison:
But then once they are approved, will I see a prediction market on projects or something like that? We can use them to—

Brian Armstrong:
Yeah, that’s a great idea.

John Collison:
What are other ways in which Coinbase is crypto‑pilled, AI‑pilled—works differently from a company founded ten or 20 years prior?

Brian Armstrong:
Like a lot of companies, we’re leaning as hard as we can into AI.

John Collison:
What does that mean concretely?

Brian Armstrong:
We’re doing a lot of the best practices. We made a big push to get every engineer on Cursor and Copilot.

John Collison:
Was it you or Toby who mandated it?

Brian Armstrong:
I mandated it.

John Collison:
Yeah. You required people to have a call with you—

Brian Armstrong:
Yeah, that’s true. I did do that.

John Collison:
You required people to justify to you, the CEO, if they weren’t using AI—

Brian Armstrong:
That’s true. Originally they were coming back and saying, “All right, over the next quarter...” Or I can’t remember if it was two quarters—they’re like, “We’re going to get to 50% adoption.” I was like, “You’re telling me... Why can’t every engineer just onboard by the end of the week?” I went rogue and posted in the all‑in Slack channel—

John Collison:
Just a light dusting of founder mode.

Brian Armstrong:
“AI’s important. We need you to all learn it and at least onboard. You don’t have to use it every day yet until we do some training, but at least onboard by the end of the week. If not, I’m hosting a meeting on Saturday with everybody who hasn’t done it and I’d like to meet with you to understand why.” A few people were on vacation; there was a list of... Anyway, I jumped on this call on Saturday, and there were a couple people who had not done it. Some of them had a good reason, because they were just getting back from some trip or something, and some of them didn’t and they got fired. Some people really didn’t like that heavy‑handed approach, but I think it set some clarity that we need to lean into this and learn about it.

John Collison:
What’s your experience of AI coding been so far? It’s clear that it is very helpful to have AI helping you write code. It’s not clear how you run an AI‑coded codebase, and what the best way to do it is.

Brian Armstrong:
I agree. We’re still figuring that out. One thing we’ve started doing is: every month we host what we call an AI speed‑run, where one of the engineers volunteers that month to run a training for how they’re using it, and we try to cherrypick the teams that are doing it the best. We’re doing, I think, about 33% of code written by AI now. We have a goal to get to 50% by the end of the quarter. Let’s see if we get there. You probably can go too far with it. You don’t want people vibe‑coding systems moving money. We’ve encouraged people to code‑review it and have the appropriate checks in place with humans in the loop. Some of the front‑end pieces, you can iterate faster.

We want to make sure it’s used not just in the engineering teams; it really should be any team. Design is using it heavily. Product managers. I think FP&A could even be using this as, “Ingest all the data and tell me what you forecast the revenue to be.” We’re getting to a world where—even as CEO, by the way—I use it a lot. We use a decision‑making process called RAPIDS, and everyone writes their input. We have a row now for AI that writes its input in as one of the people that help make decisions. We’re testing the limits of when it can actually start to be the decision‑maker on some things and do better than humans.

John Collison:
Yeah. Last question. What advice do you have for us as we start to navigate the crypto world more, and as crypto becomes so much more relevant for payments?

Brian Armstrong:
I feel like I should be asking you the same question, because we’re starting to figure out—we’re entering payments. You guys are entering crypto. Hopefully we get to work more together. Crypto— it’s never as good as it seems. It’s never as bad as it seems. You’ve got to be in it for the long haul, through the ups and downs.

John Collison:
It’s very cyclical.

Brian Armstrong:
Yeah. I would say that crypto is about decentralized protocols, fundamentally. Those are the ones that have done the best. There are temptations at times—I won’t name some of the prior projects—but they’ve come together with small consortiums of companies. I think we really want to make more open standards, and there’s a trade‑off. If you truly have a decentralized protocol, it’s slower. Base, for instance—we started it, but it’s going through a progressive decentralization, and we went from stage zero to stage one decentralization. We’re getting to stage two soon. Everybody should—on that network—be able to exist on a level playing field with Coinbase. That’s where we’re getting to. In that ethos, it’s a permissionless system everybody can build on. That’s what’s going to actually disrupt traditional finance. It’s not going to be another proprietary system. There’s a spectrum between those two things. It’s not binary. That’s how to get really red‑pilled on crypto.

John Collison:
Embrace what makes crypto actually work.

Brian Armstrong:
Yep.

John Collison:
Yeah. All right. With that, thank you.

Brian Armstrong:
Thank you.