Brian Armstrong
Co-founder and CEO of Coinbase
Where to hear and read about Brian Armstrong1 episode or test
Brian Armstrong is the co-founder and CEO of Coinbase, which he built from a nights-and-weekends Bitcoin wallet into a public company that sued the SEC and won. The thing most worth learning from him is how a single long-term mission, increasing economic freedom, lets him make unpopular calls that most CEOs avoid and absorb years of short-term pain.
We're willing to be misunderstood for a long time, and then you slowly start to have these breakthroughs.
Brian Armstrong: When Washington Tried to Kill Coinbase
Brian Armstrong: the story and the lessons
1How Brian Armstrong found his life's work
Armstrong's early ventures chased passive income, not a mission. After graduating in 2005 he aimed for about $100,000 a year passively, built the University Tutor matching site with a roommate, then refurbished rental houses in Houston.
Tutoring paid about $60 an hour against $7 or $8 at the library, which is what drew him in. He admits he was never passionate about tutoring or education.
A year alone in Argentina showed him what broken money does. He went to test himself, and watched inflation hollow out a country that was a top-10 economy around 1908 and is now about the 100th richest.
He read Ayn Rand and Milton Friedman there, and saw grand government buildings with cracks, ivy and graffiti, and people pessimistic about the future.
Seth Godin's The Dip narrowed his options to one field. He asked what he would do for 20 years even with little success, and only tech entrepreneurship made the list, so he sold the rentals, closed the side projects and moved to Silicon Valley.
Coinbase was founded within a few years and reached a $1 billion valuation within seven years of that decision.
Airbnb's payout problem showed him what global money should look like. Host payouts abroad lost 7% to 12% in fees through local oligopolies, and a $100 test transfer to Ecuador exposed how opaque the fees were.
He read the Bitcoin whitepaper in December 2010 while working at Airbnb and concluded the answer was a system that is fast, cheap, permissionless and decentralized.
He built the first prototypes after work, on his own laptop. He worked from 8:30 to midnight, five days a week and sometimes Sundays, keeping it separate from Airbnb time and property.
First came an Android Bitcoin wallet with a college friend at Google, then a cloud wallet that required reimplementing a Bitcoin node in Ruby.
2How Armstrong got Coinbase off the launchpad
A rushed co-founder pick fell apart within three months. He applied to Y Combinator with Ben Reeves of blockchain.info, whom he had just met over coffee, then had a hard conversation and went through YC solo.
He had searched about a year and a half for a partner because Paul Graham's essays favored complementary founding teams. His advice: "You should really cofound with people you've known for a long time."
Fred Ehrsam became the partner Coinbase needed. Ehrsam, a former Goldman Sachs FX trader, reached out after the seed round, worked unofficially, then became co-founder.
Within about three weeks he mapped the flow of funds and found Coinbase lost money on every trade from timing risk, fixed by changing the fee structure. Armstrong says Coinbase might not have survived without him.
Talking to users turned a wallet into a buy button. The first app only stored and sent Bitcoin and users did not come back; calling signups revealed they simply had no Bitcoin.
Bank (ACH) buying, which Coinbase was the first U.S. crypto company to offer, brought instant product-market fit. He describes the boulder rolling downhill instead of being pushed up.
A $30,000 legal opinion replaced a $5 to $10 million license. Banks hung up on him until YC partners pointed him to Silicon Valley Bank, whose compliance team flagged Coinbase as a possible money transmitter.
With only about $600K raised, he paid for a five-page opinion arguing it might not be one, which opened the account. He then wrote the ACH code himself, sending files to the bank by FTP.
One demand graph raised $25 million in about a week. Coinbase pre-bought Bitcoin daily, sending about $550,000 out against $600K raised, and the bank threatened to close the account.
With no deck, they showed investors daily buys and raised from Union Square Ventures and Ribbit. "The bank's going to close our account in two weeks if we don't get that money."
Survival mode lasted four or five years. Support tickets grew from 1,000 to 10,000 with no support team, and angry users found the office from a photo.
Ehrsam sometimes answered the door holding a golf club. Hackers attacked constantly and engineers quit after being paged three times a night; only by year four or five could Armstrong take a week off.
3Why Brian Armstrong sued the SEC
The SEC refused to give rules and then sued. Coinbase met the SEC about 30 times, was told to talk to its lawyer, and an enforcement action came the next day demanding it delist assets.
Armstrong's pitch had been simple: tell us the rules and we follow them. He believes Elizabeth Warren pushed Gary Gensler to curb crypto through enforcement instead of rulemaking.
He sued against nearly everyone's advice after months of deliberation. The decision took three or four months; he spoke with financial CEOs who had sued the SEC and won, and cited SpaceX and Palantir.
Coinbase sued first under the Administrative Procedures Act. He says the mission made it simple: a few years of pain was fine if the industry got built in the US.
The fight was costly, and it ended in a clean win. Legal costs ran $50 to $100 million and the stock took an estimated $10 to $20 billion hit, while many crypto startups folded.
Coinbase paid no fines and changed nothing; the SEC withdrew the case and judges called its conduct arbitrary and capricious. He admits earlier in the company's life it could not have afforded the fight.
Armstrong now pushes Congress so no regulator can repeat it. He backs the CLARITY Act, which passed the House, to settle which tokens are commodities under the CFTC and which are securities under the SEC.
He calls it a parochial US turf war, since the UK and Singapore have one regulator. He visits Washington about once a quarter, more at key moments, "so that a future Gary Gensler couldn't come in and try to kill the industry."
4Armstrong's Mission First turn and what it cost
A walkout over Black Lives Matter forced a culture decision. In 2020 about 300 employees walked out remotely after his town hall answer, and a hasty statement left him feeling he had compromised himself.
Town halls had drifted from products to social causes amid COVID isolation. They now use pre-submitted questions so one person cannot derail 3,000 others.
He made Coinbase apolitical at work and offered severance to dissenters. The Mission First post said no politics at work unless it concerns crypto and economic freedom; he expected up to half to leave, but only 5% took the package.
Staff begged him not to post it. After reading Jonathan Haidt's The Coddling of the American Mind, he concluded a vocal 1% made it feel like half the company opposed it.
Armstrong was ready to rebuild the company from scratch. Asked about losing half the company, he said they would have built it back, since he had been there at one, 10, 100 and 1,000 people.
He draws on Lee Kuan Yew's 'iron in the veins' speech and calls saying 'we're going this way' the core of leadership. "I could go back to being on my laptop again if I had to, right?"
Surviving a hit piece freed him from optimizing for optics. After the post, outlets implied Coinbase was racist and underpaid minorities, which he calls false, and he was told The New York Times set a team to dig up dirt.
Once the worst case landed and nothing terrible followed, the fear of perception faded. He thinks every leader should get such a piece once.
5Brian Armstrong's rules for running Coinbase
Every decision gets one owner and a short deadline. He wants clear DRIs and a single decision-maker, and calls himself the pace car for execution.
Clear owners get harder to keep as stakeholders multiply, and short time frames force teams to unblock and go.
He takes the blame so teams will take long-shot bets. A 20% chance at a 20X outcome is a bet he would take all day, though most companies refuse such odds.
His line to teams: "Go for it. If it fails, it's on me."
One yes from any funder greenlights an internal bet. Twice a year employees pitch, and product leaders, the CFO, the COO and a few engineers can each fund from their budgets; Armstrong voted no on USDC but another leader funded it.
USDC brought in about $800 million of revenue in 2025. He fears rejected ideas walk out the door, citing Wozniak leaving HP to found Apple. AI makes V1s cheap, so teams start at two or three people and earn an internal Series A at milestones.
Armstrong works on whatever is the company's limiting factor. Borrowing Elon Musk's question, he goes after the gnarliest problem, from feuding teams to layoffs, even reviewing 300 resumes when needed.
Motivation comes from the mission, so grind work counts if it moves it. When confused he calls an expert, since someone with ten years on a problem can explain it in 30 seconds, and he follows his nose when he hears something weird the third time.
He pairs his product focus with a strong operator. Armstrong runs product while president and COO Emilie Choi runs operations.
An operator alone misses the next wave and a founder alone may blow the place up; he points to Zuckerberg and Sandberg, and Page and Brin with Schmidt.
Coinbase hired for spikes over polished resumes. For the first support hire he chose Olaf Carlson-Wee, a former lumberjack who wrote his thesis on Bitcoin, over a low-energy Google AdSense lead.
Carlson-Wee later became a billionaire running a crypto fund. Early hires were crypto zealots found at meetups and through cold LinkedIn messages, because Coinbase was not a hot company. Mission and values were only written down at a couple of hundred people.
6How Armstrong thinks about building in crypto
Bitcoin was a wedge, not a master plan. Coinbase grew into trading stocks, commodities and prediction markets, plus loans and a card, though he did not foresee stablecoins or prediction markets.
He says claiming a full plan from day one would be intellectually dishonest.
Ease of use grows the market more than purity does. He answered decentralization critics with the email analogy: one open protocol, many providers, then later added a self-custodial wallet.
Retail users who lost a password lost their Bitcoin, and institutions holding 80% to 90% of the world's money wanted enterprise-grade custody. He cites Steve Jobs's appliance computer creating a market bigger than hobbyists.
New industries get ignored, then fought, and crypto is in the fight. Banks once hung up on him; now five of the world's largest banks work with Coinbase on crypto integrations.
He compares the fight to Uber versus cabs and the Wright brothers finding buyers in Europe, and stresses an idea must be contrarian and right. He sees Satoshi as the inventor and himself as the engineer who commercializes it.
AI agents need wallets to finish their work. More than 50% of Coinbase's code is written by agents and about 60% of support inquiries are answered by them; Coinbase built tools giving any agent a stablecoin wallet.
Agents running overnight may need to buy compute, a paywalled paper or a domain, and cannot hold corporate cards.
He admits the Base App's social tokens fell short. Every post had its own coin, but he says the tokenomics lacked durability, so the app refocused on trading.
He still expects social tokens to work, perhaps tied to a creator's ad revenue.
7Why Armstrong says every company is a media company
Coinbase turned its earnings pitch into short vertical video. A marketer paired Armstrong's deck walkthrough with gameplay footage to lift retention, since 99% of people never read the shareholder letter.
He quotes David Ogilvy: "You can't save souls in an empty church." He treats each earnings call as a marketing moment despite finance staff urging him to stay on script.
He prefers going direct over routing news through journalists. Retail investors live on podcasts, X, blogs and Substack, and a reporter's filter carries bias.
The marketing team gets the same air cover as product, which produced Super Bowl QR codes and a karaoke stunt.
8How Brian Armstrong sustains himself and picks his next bet
Armstrong credits autism-spectrum traits for his focus. He can focus on interesting work about 12 hours a day, finds coding easy and a day of hard conversations taxing.
He doesn't let fear of being disliked stop him, though it still causes stress, and says unusual calls like suing a regulator seemed ordinary to him.
A strict routine keeps him from burning out. He watched YC peers burn out within three to four years and hit his own patch about every two years, responding by delegating and cutting direct reports.
His routine: sleep, heavy lifting, zone two cardio, brief meditation, the same clothes daily, and an hour screen-free before bed. Following Palmer Luckey, he starts with the highest-leverage task, which usually sucks.
He started NewLimit where no credible teams were working. AI, fusion and space had strong teams, so he chose longevity, putting in $100 million of his own money.
Dinners with biotech experts kept pointing to epigenetic reprogramming. NewLimit has restored function in human cells and expects its first drug candidate in clinical trials next year; Jacob Kimmel runs it day to day.
He wants regulatory sandboxes for new technology. He would designate about 10 plots of federal land as special economic zones for nuclear, biotech, crypto or drones.
Products that work there could then seek a federal license. He also backs ResearchHub to speed science funding.
Brian Armstrong: When Washington Tried to Kill Coinbase
Armstrong shows how a founder with a decades-long mission builds a company in a brand-new industry: surviving near-death moments early, then fighting a hostile regulator and an internal culture split instead of backing down. The decision that makes it worth hearing is suing the SEC against nearly everyone's advice, which cost billions in market value in the short run and ended in a win with no fines and no changes to the business.
Armstrong's early ventures chased passive income, not a mission. After graduating in 2005 he aimed for about $100,000 a year passively, built the University Tutor matching site with a roommate, then refurbished rental houses in Houston.
Tutoring paid about $60 an hour against $7 or $8 at the library, which is what drew him in. He admits he was never passionate about tutoring or education.
A year alone in Argentina showed him what broken money does. He went to test himself, and watched inflation hollow out a country that was a top-10 economy around 1908 and is now about the 100th richest.
He read Ayn Rand and Milton Friedman there, and saw grand government buildings with cracks, ivy and graffiti, and people pessimistic about the future.
Seth Godin's The Dip narrowed his options to one field. He asked what he would do for 20 years even with little success, and only tech entrepreneurship made the list, so he sold the rentals, closed the side projects and moved to Silicon Valley.
Coinbase was founded within a few years and reached a $1 billion valuation within seven years of that decision.
Airbnb's payout problem showed him what global money should look like. Host payouts abroad lost 7% to 12% in fees through local oligopolies, and a $100 test transfer to Ecuador exposed how opaque the fees were.
He read the Bitcoin whitepaper in December 2010 while working at Airbnb and concluded the answer was a system that is fast, cheap, permissionless and decentralized.
He built the first prototypes after work, on his own laptop. He worked from 8:30 to midnight, five days a week and sometimes Sundays, keeping it separate from Airbnb time and property.
First came an Android Bitcoin wallet with a college friend at Google, then a cloud wallet that required reimplementing a Bitcoin node in Ruby.
A rushed co-founder pick fell apart within three months. He applied to Y Combinator with Ben Reeves of blockchain.info, whom he had just met over coffee, then had a hard conversation and went through YC solo.
He had searched about a year and a half for a partner because Paul Graham's essays favored complementary founding teams. His advice: "You should really cofound with people you've known for a long time."
Fred Ehrsam became the partner Coinbase needed. Ehrsam, a former Goldman Sachs FX trader, reached out after the seed round, worked unofficially, then became co-founder.
Within about three weeks he mapped the flow of funds and found Coinbase lost money on every trade from timing risk, fixed by changing the fee structure. Armstrong says Coinbase might not have survived without him.
Talking to users turned a wallet into a buy button. The first app only stored and sent Bitcoin and users did not come back; calling signups revealed they simply had no Bitcoin.
Bank (ACH) buying, which Coinbase was the first U.S. crypto company to offer, brought instant product-market fit. He describes the boulder rolling downhill instead of being pushed up.
A $30,000 legal opinion replaced a $5 to $10 million license. Banks hung up on him until YC partners pointed him to Silicon Valley Bank, whose compliance team flagged Coinbase as a possible money transmitter.
With only about $600K raised, he paid for a five-page opinion arguing it might not be one, which opened the account. He then wrote the ACH code himself, sending files to the bank by FTP.
One demand graph raised $25 million in about a week. Coinbase pre-bought Bitcoin daily, sending about $550,000 out against $600K raised, and the bank threatened to close the account.
With no deck, they showed investors daily buys and raised from Union Square Ventures and Ribbit. "The bank's going to close our account in two weeks if we don't get that money."
Survival mode lasted four or five years. Support tickets grew from 1,000 to 10,000 with no support team, and angry users found the office from a photo.
Ehrsam sometimes answered the door holding a golf club. Hackers attacked constantly and engineers quit after being paged three times a night; only by year four or five could Armstrong take a week off.
The SEC refused to give rules and then sued. Coinbase met the SEC about 30 times, was told to talk to its lawyer, and an enforcement action came the next day demanding it delist assets.
Armstrong's pitch had been simple: tell us the rules and we follow them. He believes Elizabeth Warren pushed Gary Gensler to curb crypto through enforcement instead of rulemaking.
He sued against nearly everyone's advice after months of deliberation. The decision took three or four months; he spoke with financial CEOs who had sued the SEC and won, and cited SpaceX and Palantir.
Coinbase sued first under the Administrative Procedures Act. He says the mission made it simple: a few years of pain was fine if the industry got built in the US.
The fight was costly, and it ended in a clean win. Legal costs ran $50 to $100 million and the stock took an estimated $10 to $20 billion hit, while many crypto startups folded.
Coinbase paid no fines and changed nothing; the SEC withdrew the case and judges called its conduct arbitrary and capricious. He admits earlier in the company's life it could not have afforded the fight.
Armstrong now pushes Congress so no regulator can repeat it. He backs the CLARITY Act, which passed the House, to settle which tokens are commodities under the CFTC and which are securities under the SEC.
He calls it a parochial US turf war, since the UK and Singapore have one regulator. He visits Washington about once a quarter, more at key moments, "so that a future Gary Gensler couldn't come in and try to kill the industry."
A walkout over Black Lives Matter forced a culture decision. In 2020 about 300 employees walked out remotely after his town hall answer, and a hasty statement left him feeling he had compromised himself.
Town halls had drifted from products to social causes amid COVID isolation. They now use pre-submitted questions so one person cannot derail 3,000 others.
He made Coinbase apolitical at work and offered severance to dissenters. The Mission First post said no politics at work unless it concerns crypto and economic freedom; he expected up to half to leave, but only 5% took the package.
Staff begged him not to post it. After reading Jonathan Haidt's The Coddling of the American Mind, he concluded a vocal 1% made it feel like half the company opposed it.
Armstrong was ready to rebuild the company from scratch. Asked about losing half the company, he said they would have built it back, since he had been there at one, 10, 100 and 1,000 people.
He draws on Lee Kuan Yew's 'iron in the veins' speech and calls saying 'we're going this way' the core of leadership. "I could go back to being on my laptop again if I had to, right?"
Surviving a hit piece freed him from optimizing for optics. After the post, outlets implied Coinbase was racist and underpaid minorities, which he calls false, and he was told The New York Times set a team to dig up dirt.
Once the worst case landed and nothing terrible followed, the fear of perception faded. He thinks every leader should get such a piece once.
Every decision gets one owner and a short deadline. He wants clear DRIs and a single decision-maker, and calls himself the pace car for execution.
Clear owners get harder to keep as stakeholders multiply, and short time frames force teams to unblock and go.
He takes the blame so teams will take long-shot bets. A 20% chance at a 20X outcome is a bet he would take all day, though most companies refuse such odds.
His line to teams: "Go for it. If it fails, it's on me."
One yes from any funder greenlights an internal bet. Twice a year employees pitch, and product leaders, the CFO, the COO and a few engineers can each fund from their budgets; Armstrong voted no on USDC but another leader funded it.
USDC brought in about $800 million of revenue in 2025. He fears rejected ideas walk out the door, citing Wozniak leaving HP to found Apple. AI makes V1s cheap, so teams start at two or three people and earn an internal Series A at milestones.
Armstrong works on whatever is the company's limiting factor. Borrowing Elon Musk's question, he goes after the gnarliest problem, from feuding teams to layoffs, even reviewing 300 resumes when needed.
Motivation comes from the mission, so grind work counts if it moves it. When confused he calls an expert, since someone with ten years on a problem can explain it in 30 seconds, and he follows his nose when he hears something weird the third time.
He pairs his product focus with a strong operator. Armstrong runs product while president and COO Emilie Choi runs operations.
An operator alone misses the next wave and a founder alone may blow the place up; he points to Zuckerberg and Sandberg, and Page and Brin with Schmidt.
Coinbase hired for spikes over polished resumes. For the first support hire he chose Olaf Carlson-Wee, a former lumberjack who wrote his thesis on Bitcoin, over a low-energy Google AdSense lead.
Carlson-Wee later became a billionaire running a crypto fund. Early hires were crypto zealots found at meetups and through cold LinkedIn messages, because Coinbase was not a hot company. Mission and values were only written down at a couple of hundred people.
Bitcoin was a wedge, not a master plan. Coinbase grew into trading stocks, commodities and prediction markets, plus loans and a card, though he did not foresee stablecoins or prediction markets.
He says claiming a full plan from day one would be intellectually dishonest.
Ease of use grows the market more than purity does. He answered decentralization critics with the email analogy: one open protocol, many providers, then later added a self-custodial wallet.
Retail users who lost a password lost their Bitcoin, and institutions holding 80% to 90% of the world's money wanted enterprise-grade custody. He cites Steve Jobs's appliance computer creating a market bigger than hobbyists.
New industries get ignored, then fought, and crypto is in the fight. Banks once hung up on him; now five of the world's largest banks work with Coinbase on crypto integrations.
He compares the fight to Uber versus cabs and the Wright brothers finding buyers in Europe, and stresses an idea must be contrarian and right. He sees Satoshi as the inventor and himself as the engineer who commercializes it.
AI agents need wallets to finish their work. More than 50% of Coinbase's code is written by agents and about 60% of support inquiries are answered by them; Coinbase built tools giving any agent a stablecoin wallet.
Agents running overnight may need to buy compute, a paywalled paper or a domain, and cannot hold corporate cards.
He admits the Base App's social tokens fell short. Every post had its own coin, but he says the tokenomics lacked durability, so the app refocused on trading.
He still expects social tokens to work, perhaps tied to a creator's ad revenue.
Coinbase turned its earnings pitch into short vertical video. A marketer paired Armstrong's deck walkthrough with gameplay footage to lift retention, since 99% of people never read the shareholder letter.
He quotes David Ogilvy: "You can't save souls in an empty church." He treats each earnings call as a marketing moment despite finance staff urging him to stay on script.
He prefers going direct over routing news through journalists. Retail investors live on podcasts, X, blogs and Substack, and a reporter's filter carries bias.
The marketing team gets the same air cover as product, which produced Super Bowl QR codes and a karaoke stunt.
Armstrong credits autism-spectrum traits for his focus. He can focus on interesting work about 12 hours a day, finds coding easy and a day of hard conversations taxing.
He doesn't let fear of being disliked stop him, though it still causes stress, and says unusual calls like suing a regulator seemed ordinary to him.
A strict routine keeps him from burning out. He watched YC peers burn out within three to four years and hit his own patch about every two years, responding by delegating and cutting direct reports.
His routine: sleep, heavy lifting, zone two cardio, brief meditation, the same clothes daily, and an hour screen-free before bed. Following Palmer Luckey, he starts with the highest-leverage task, which usually sucks.
He started NewLimit where no credible teams were working. AI, fusion and space had strong teams, so he chose longevity, putting in $100 million of his own money.
Dinners with biotech experts kept pointing to epigenetic reprogramming. NewLimit has restored function in human cells and expects its first drug candidate in clinical trials next year; Jacob Kimmel runs it day to day.
He wants regulatory sandboxes for new technology. He would designate about 10 plots of federal land as special economic zones for nuclear, biotech, crypto or drones.
Products that work there could then seek a federal license. He also backs ResearchHub to speed science funding.