10 Real Founder & Investor Stories

These are ten real companies and founders whose arcs are public and widely documented — the kind of thing anyone can confirm in a search. I've put them together because each one illustrates a pattern that also shows up, at scale, in an outcome dataset we built of roughly 6,900 accelerator companies across Y Combinator and Startmate.

1. Airbnb — the rejection emails, framed on a wall

In 2008 three founders were selling cereal boxes — "Obama O's," "Cap'n McCain's" — to keep the lights on, because their actual idea, letting strangers sleep on air mattresses in your flat, was one almost every investor they met found faintly ridiculous. Brian Chesky has since published the rejection emails. Seven investors passed. The polite ones said the market was too small.

Airbnb went public in December 2020 worth tens of billions. Seven serious, experienced people looked at the same company and priced it at roughly zero. They weren't stupid. They were pattern-matching against a world that didn't exist yet — and pattern-matching against the present is exactly how you miss the future.

You are allowed to be a company the room couldn't see yet.

2. Stripe — seven lines of code and a narrow wedge

When the Collison brothers started what became Stripe, the pitch was almost insultingly small: a few lines of code that let a developer take a payment. Not a bank. Not a platform. A snippet. They went through Y Combinator, shipped the narrow thing, and made it work so cleanly that developers chose it without being sold to. It's now one of the most valuable private companies on earth.

I think about Stripe whenever someone shows me a deck with a grand, sprawling vision and no wedge. Grand visions are cheap. A thing that works, that someone chooses on its own merits, is rare.

The narrow wedge that actually works beats the broad vision that almost does. Start where you can be undeniable.

3. Coinbase — the bet the base rate would have killed

In 2012, telling an investor you were building a company around Bitcoin was close to a punchline. The technology was fringe, the users were few, the regulatory picture was a fog. Brian Armstrong went through Y Combinator anyway. Coinbase went public in 2021 in a listing valued in the tens of billions. The early backers who leaned in — the ones willing to be non-consensus and right — earned a return that paid for a lot of misses.

A model trained on 2012's data would have scored Coinbase down, hard — because the whole point of the outlier is that it looks wrong against everything that has already worked. Scoring engines regress to the mean. Power-law winners live in the tail.

Use the machine to clear the noise. Spend your conviction on the bets that look wrong.

4. Twitch — the company that survived by becoming a different company

Twitch didn't start as Twitch. It started as Justin.tv — a single guy livestreaming his entire life from a camera on his head, in 2007. As a business, "broadcast one man's day" went roughly nowhere. But buried inside the traffic was a vertical that wouldn't stop growing: people watching other people play video games. They spun that vertical out, called it Twitch, and in 2014 Amazon bought it for a reported ~US$970 million.

The lesson isn't "pivot." The lesson is that the company that exited and the company that was funded were not the same company — and if you only study the exit, you learn nothing about the years of the wrong idea that came first.

The exit is one line. The story is the years above it. Judge the founder who's still iterating, not the pitch that's already clean.

5. Kyle Vogt — the same founder, twice

Kyle Vogt was one of the people behind that Justin.tv camera. When the livestreaming chapter ended, he didn't retire on the Twitch outcome. He started Cruise, a self-driving car company, went through Y Combinator again in 2014, and in 2016 General Motors acquired it — a reported billion-dollar-scale deal — barely two years in.

When we linked founders across our accelerator dataset by their verified identity — not their names, which false-merge constantly — this is the pattern that surfaced: a small set of people who show up on more than one company, having learned something the first time that doesn't fit in a pitch deck.

The second company is rarely a fluke. The first one was the tuition. Back the founder, not just the round.

6. Instacart — about twenty failures before the one that worked

Before Instacart, Apoorva Mehta reportedly tried around twenty other ideas — a social network for lawyers among them — and buried every one. He'd left a comfortable engineering job at Amazon to do it. By the time he landed on grocery delivery and applied to Y Combinator, he was famously late to the batch, and the story goes that he got in by building the product and delivering the partners a six-pack of beer through his own app. Instacart went public in 2023, valued in the billions.

Twenty misses isn't a red flag; on the evidence, it's the apprenticeship. The person who has failed at nineteen things and is still standing at the twentieth has a kind of information the first-timer simply doesn't have — about themselves, mostly.

The idea that works is usually not the first idea. It's the first idea after you'd earned it.

7. DoorDash — the boring company that ground it out

DoorDash was not a sexy pitch. Food delivery — a category littered with the corpses of companies that had tried it and burned. It started as "Palo Alto Delivery," a landing page and a phone number run by Stanford students. It went through Y Combinator in 2013. They ground it out for seven years and IPO'd in December 2020, valued in the tens of billions.

This is the counterweight to every "non-consensus genius" story. Sometimes the outlier isn't a dazzling insight — it's unglamorous operational conviction, applied to a hard problem, for longer than anyone else was willing to. We under-price endurance because it looks like nothing from the outside.

Not every fund-returner is a firework. Don't mistake unglamorous for unpromising.

8. Reddit — nineteen years from the first batch to the bell

Reddit was in the very first Y Combinator batch, summer 2005. The founders had pitched something else entirely — a mobile food-ordering idea — been told to build something different, and turned around a link-sharing site in a matter of weeks. It was acquired by Condé Nast in 2006 for a modest sum, wandered for years, nearly died more than once. It rang the opening bell as a public company in March 2024. Nineteen years after the batch.

The real clock on a company is measured in decades, and most of the value in Reddit's arc accrued in years when it looked like a footnote.

Time is the actual edge. Think in decades. The bell is a long way from the batch.

9. The names that keep showing up

Here's a thing you notice when you stop tracking companies one at a time and start tracking them by the thousand: the winners share investors. When we linked the investors in each round to the companies that later exited and rolled it up, the same handful of names surfaced again and again — the firms that were early in the big outcomes, the seed-stage angels whose signatures recur under the exits. It surprises no one that names like a16z and Sequoia turn up repeatedly. What's useful is that it stops being reputation and becomes a query.

I'll be honest about the limit: an investor's track record is survivorship too. You see the exits they backed; you don't automatically see the ones that died beside them. The count of wins is a real signal. The rate needs the losses linked in as well.

A cap table is a track record. Read it like one.

10. The story with no name

Every story above has a name, a logo, a happy ending you can look up. Here's the one that doesn't.

In the outcome graph we built this week, of the ~6,900 companies the best accelerators in the world accepted, about one in six with a known status is simply dead. No acquisition post. No IPO bell. No keynote about the journey. There are more than a thousand of them in the data, and I won't name a single one, because they've been through enough without a stranger turning their closure into a content hook.

But I want them counted, because they are the most honest thing in the whole dataset. They are the reason a scoring engine trained only on winners is a liar — it has never met them. They are the proof that acceptance was never a guarantee, and, read the other way, that a rejection was never a verdict.

The winners get the story. The dead ones get the truth. You need both to score anything honestly.

Frequently asked questions

Which famous startups were rejected before they succeeded?

Airbnb is the best-documented example — Brian Chesky has published emails from seven investors who passed in 2008, before the company's tens-of-billions IPO in December 2020. The broader pattern holds across the data: an accelerator or investor's 'no' prices fit, not the founder's ultimate quality.

Do serial founders really do better?

The pattern shows up clearly when you link companies by a founder's verified identity rather than their name. Kyle Vogt, for instance, helped build the livestreaming company that became Twitch and then founded Cruise — two separate acquisitions. The second company is rarely a fluke; the first one is the tuition.