Lessons from $5b founders (part I)

Scarcity is a teacher that abundance can never be.

Pagar.me raised $300K total. With two months of runway left and an investor who refused to fund or release them, Dubugras and Franceschi had to grow their way to survival. The discipline they learned under that constraint (unit economics, ruthless prioritization, speed) became the operating system they brought to Brex.

Your unfair advantage is the thing you keep trying to escape.

Dubugras and Franceschi arrived in Silicon Valley determined to leave fintech behind. They tried VR. It went nowhere. The moment they returned to payments (the domain where they had three years of scar tissue and hard-won expertise), Brex found instant product-market fit. The thing you’re running from may be the thing that wins.

The co-founder relationship is the company.

Dubugras and Franceschi met arguing on Twitter at 15. They have built two companies together across two continents over 13 years. The relationship comes before the company, always.

Design your role around your strengths, not the org chart.

Dubugras openly says he dislikes managing. So Brex split the CEO role: he took external (fundraising, customers, PR), Franceschi took internal (product, engineering, operations). Both became better.

The decision that gets you criticized is often the one that saves you.

Dropping 60,000 customers in 2022 generated terrible press. But the SMB segment was destroying margins. The pivot to enterprise stabilized the business and led to 80-90% revenue growth by 2024.

Builders don’t retire.

Dubugras stepped back from Brex at 28. Before the Capital One deal even closed, he was already hiring for a new stealth company. The fire required to start a company does not get extinguished with an exit event.

The best startup ideas come from listening, not pitching.

Tang’s app prototype failed at the macaroon shop. But because he stayed to listen, Chloe showed him the booklet of canceled orders—and that single insight became DoorDash. The company was born from a conversation, not a pitch deck.

Launch before you’re ready. A landing page is enough.

PaloAltoDelivery.com was a static page with PDF menus and a personal cell phone number. It took two hours to build. It validated the entire concept within hours. Most founders over-build before testing demand.

Go where the competition isn’t.

GrubHub and UberEats fought over Manhattan. DoorDash went to the suburbs and college towns - places where delivery had never existed. By the time competitors noticed, DoorDash had locked up those markets.

Sometimes the margin is created through unscalable actions (for unreasonable periods of time).

The founders personally delivered food for months. They answered calls during class. They hired pizza drivers on the spot. Every DoorDash employee still spends their first week as a Dasher. The unscalable phase taught them everything the data could not.

Delivery is not software. Every order is a real-world problem.

You cannot clone a delivery the way you clone code. Tang’s transition from CPO to head of DoorDash Labs reflects this insight: the future of logistics requires solving hard physical-world problems (robotics, autonomy, route optimization) not just building better apps.

Side projects are the curriculum. Most will fail. Some will change everything.

Before DoorDash, Tang built a news reader, a calendar app, a messaging app, a social news aggregator, and wrote a book. All of them failed or faded. But each one taught him something he used at DoorDash. The portfolio of failures is the real education.

The biggest opportunities hide inside the problems nobody wants to talk about.

Data labeling was the unsexy foundation of every AI model. Wang built the infrastructure while everyone else chased the models themselves. Within five years, OpenAI, Meta, and Microsoft all depended on his platform.

Enterprise sales is won at the booth, not on the stage.

Scale’s first customers came from Wang walking booth to booth at CVPR with a laptop, demoing to anyone who would watch. The $29 billion company started with cold pitches, not warm introductions.

The room you choose to be in defines the company you become.

Wang chose Washington over Sand Hill Road as his second audience. By briefing Congress and the Pentagon, he turned Scale from a vendor into a national security asset - a positioning no competitor could replicate.

Build the platform, not the product.

Scale never built its own AI models. It built the infrastructure that made everyone else’s models possible. When the industry shifted from self-driving cars to LLMs, Scale’s platform shifted with it.

Naivete is an asset, not a liability.

Wang dropped out at 19 with no industry experience. He has said repeatedly that not knowing what was ‘supposed to be hard’ allowed him to attempt things incumbents wouldn’t. Fresh eyes see what expertise blinds.

When you can’t buy the asset, write about it.

Buterin couldn’t afford Bitcoin. So he earned it by writing for $3.50 per article. That writing became Bitcoin Magazine, which became his education, his network, and his platform. The indirect path (deep intellectual engagement rather than speculation) gave him the insight to build Ethereum.

Write the paper before you build the product.

The Ethereum white paper came before the team, the funding, and the code. A clear, rigorous articulation of the idea attracted everything else. Buterin did not pitch investors. He published an idea and let the idea do the recruiting.

Build the standard, not the application.

Ethereum’s power comes from what others build on top of it. The ERC-20 token standard enabled ICOs, DeFi, NFTs, DAOs - entire industries Buterin never built. He created the programmable substrate and let the ecosystem emerge. Platforms outlast products.

Survive the crisis that defines you.

The DAO hack, the 2018 crash, the Ethereum killer competition—each could have ended the project. Instead, each forced a decision that made Ethereum stronger. The hard fork showed the community could act. The crash cleared out speculation. The killers motivated The Merge.

Decentralize yourself before you decentralize the system.

Buterin’s greatest challenge is reducing his own influence over a system designed to have no central authority. He has deliberately delegated, stepped back, and supported independent teams. The hardest part of building a decentralized system is letting go of it.

Validate demand before you build anything.

Villig recruited fifty drivers and a waiting list of riders before writing a single line of code. He proved demand existed on the streets of Tallinn, then built the product. Most founders build first and hope for demand later. Villig inverted the sequence.

When the data says go where nobody else is going, trust the data.

Bolt’s quantitative model pointed to African cities when every instinct said to chase London and Berlin. Villig followed the data. Africa became Bolt’s second-largest market and the strategic wedge that differentiated it from Uber permanently.

Frugality is a weapon more than it is a constraint.

Uber burned $6.3 million per day. Bolt ran near break-even. Lower overhead meant lower commissions, which meant happier drivers and cheaper rides. The company that learned to fight with nothing could not be outspent by the company that learned to fight with everything.

Lower commissions win marketplaces.

Bolt takes 15 percent commission per ride. Uber takes 25 percent. The ten-point gap means Bolt drivers earn more and Bolt riders pay less. In a commodity market where service quality is roughly equivalent, economics determine the winner.

You can build a global company from anywhere.

Villig never lived in Silicon Valley. He built an $8.4 billion company from Tallinn, a city of 450,000 people. Estonia’s digital infrastructure, Skype’s precedent, and Villig’s willingness to expand globally from day one proved that geography is not destiny.

Micro obsession before macro expansion

Bolt spent its first five years doing one thing: ride-hailing. Only after achieving dominance in that category did it add scooters, food delivery, and car-sharing. The near-bankruptcy from premature expansion taught Villig that focus is a survival strategy.

Chief capitalist to out-of-distribution individuals.