Read the screenplay: FANNIEGATE — $7 trillion. 17 years. The biggest fraud in American capital markets.
RIP$50B+ LOST

Startup Graveyard

15 startups that aimed impossibly high, burned billions, and flamed out spectacularly.
The lessons are real. The money is gone.

15

Dead Startups

$50B+

Peak Value Destroyed

268

Days (Shortest: Pets.com IPO)

3

Founders in Prison

Scoring System

Each startup is scored on three dimensions out of 10: Hubris Level (how delusional was the leadership), Money Burned (how much capital was destroyed), and Lesson Value (how useful is the failure as a case study). Total score out of 30.

The Graveyard

Ranked by total severity score. These founders aimed high — and the lessons are worth studying.

RIP
#1

FTX

When your risk manager is your CEO's girlfriend

Peak: $32B20192022

What Went Wrong

Sam Bankman-Fried built the second-largest crypto exchange in the world, then allegedly used customer deposits to fund his hedge fund Alameda Research. The risk management team was his inner circle of housemates in the Bahamas. When CoinDesk exposed the balance sheet, $6 billion in withdrawals hit in 72 hours. The company collapsed in a week.

The Lesson

If a 30-year-old in cargo shorts is managing $16 billion and sleeping on a beanbag, ask harder questions. Regulatory arbitrage is not a long-term strategy.

Hubris10/10
Burn10/10
Lesson10/10
Total30/30Catastrophic
RIP
#2

WeWork

A community company that wasn't really a tech company

Peak: $47B20102023

What Went Wrong

Adam Neumann convinced SoftBank that renting desks was a tech company worth $47 billion. He trademarked the word 'We,' then bought the trademark back to himself for $5.9 million. He cashed out $700 million before the IPO imploded. The S-1 filing was so unhinged that it read like a philosophy dissertation written by someone who had just discovered kombucha.

The Lesson

Charisma is not a business model. If your CEO is buying wave pools and tequila companies with investor money, the unit economics probably don't work.

Hubris10/10
Burn10/10
Lesson9/10
Total29/30Catastrophic
RIP
#3

Theranos

Fake it till you make it... to prison

Peak: $9B20032018

What Went Wrong

Elizabeth Holmes claimed she could run 200+ blood tests from a single finger prick. The technology never worked. Not 'didn't work well' — it literally did not function. Theranos ran most tests on commercial machines from Siemens while telling investors and patients otherwise. The board included Henry Kissinger and George Shultz but zero medical scientists.

The Lesson

A board full of famous names is not the same as a board full of domain experts. When the product is healthcare, 'fake it till you make it' can kill people.

Hubris10/10
Burn7/10
Lesson10/10
Total27/30Catastrophic
RIP
#4

Quibi

Nobody asked for this

Peak: $1.75B20182020

What Went Wrong

Jeffrey Katzenberg raised $1.75 billion for a mobile-only streaming service with 10-minute episodes. It launched in April 2020 — the exact month everyone was stuck at home watching TV on the biggest screen they owned. You couldn't screenshot content. You couldn't cast to a TV. It shut down six months later. The content was fine. The premise was broken.

The Lesson

Raising $1.75 billion does not validate your thesis. Sometimes the market research is just... asking people if they'd use it. Quibi never did.

Hubris8/10
Burn9/10
Lesson8/10
Total25/30Devastating
RIP
#5

Zillow Offers

An algorithm can't judge a kitchen

Peak: $1B+ (division)20182021

What Went Wrong

Zillow built an algorithm to buy and flip houses at scale — iBuying. The algorithm consistently overpaid. It couldn't account for neighborhood vibes, a weird floor plan, or the fact that the kitchen hadn't been updated since 1974. Zillow ended up holding 7,000 homes it overpaid for and wrote down $569 million. They laid off 2,000 employees and exited the business entirely.

The Lesson

Real estate is local, emotional, and full of variables that algorithms struggle to price. If your model says a house is worth $400K and every human says $320K, trust the humans.

Hubris7/10
Burn8/10
Lesson9/10
Total24/30Devastating
RIP
#6

Juicero

A $400 juice bag squeezer that you could squeeze by hand

Peak: $120M20132017

What Went Wrong

Juicero built a $400 WiFi-connected juicer that squeezed proprietary juice packs. Then Bloomberg reporters discovered you could squeeze the packs by hand and get the same result. The machine had 400 custom parts, a scanner that checked pack expiration via QR code, and enough pressing force to lift two Teslas. All to squeeze a bag of juice.

The Lesson

Overengineering a solution to a problem that doesn't exist is the most Silicon Valley thing imaginable. Not everything needs to be 'disrupted.'

Hubris9/10
Burn5/10
Lesson9/10
Total23/30Devastating
RIP
#7

Better.com

CEO fires 900 people on a Zoom call, becomes a meme

Peak: $7.7B20162024

What Went Wrong

Vishal Garg built a legitimate digital mortgage company valued at $7.7 billion. Then in December 2021, he fired 900 employees on a single Zoom call — 9% of the company — three weeks before Christmas, reading from a script while employees cried on camera. The clip went viral. He accused laid-off employees of 'stealing' by being unproductive. He took a leave of absence, came back, and the company's SPAC merger valued it at $400 million. A 95% decline.

The Lesson

How you treat people in a downturn defines your company more than how you treat them in a boom. One viral moment can destroy a decade of brand building.

Hubris8/10
Burn6/10
Lesson9/10
Total23/30Devastating
RIP
#8

Webvan

Too early + too expensive = Game Over

Peak: $1.2B19962001

What Went Wrong

Webvan promised same-day grocery delivery in 1999 — a great idea that was 20 years too early. They built massive automated warehouses costing $35 million each, hired thousands of drivers, and expanded to 8 cities before proving the model worked in one. They burned through $830 million in venture capital in under two years. Instacart would later prove the concept — by not building any warehouses at all.

The Lesson

Being right about the future doesn't help if you're 20 years early and burning $1 million per day. Instacart proved the same thesis with 1% of the capital by using existing infrastructure.

Hubris6/10
Burn8/10
Lesson9/10
Total23/30Devastating
RIP
#9

MoviePass

Unlimited movies for $10/month... math?

Peak: $300M (parent Helios)20112019

What Went Wrong

MoviePass offered unlimited movie theater tickets for $9.95 per month when a single ticket cost $15. They were paying theaters full price for every ticket and losing $5+ per user per visit. Their plan was to get so big that theaters would cut deals. Theaters never cut deals. At one point the company was burning $40 million per month.

The Lesson

Subsidizing your way to market dominance only works if you eventually stop subsidizing. MoviePass had no path to profitability and everyone knew it except MoviePass.

Hubris7/10
Burn7/10
Lesson8/10
Total22/30Devastating
RIP
#10

Google+

Even Google can't force social

Peak: N/A (internal Google project)20112019

What Went Wrong

Google launched Google+ as a Facebook killer and tied it to every Google product — Gmail, YouTube, Search. They forced users to create Google+ profiles to comment on YouTube, which generated massive backlash. The platform had 'ghost town' energy from day one. Despite reportedly having 2 billion profiles, actual monthly active users were a fraction of that. A data breach in 2018 gave Google the excuse to finally shut it down.

The Lesson

You cannot engineer virality. Social networks are organic or they're nothing. Even the most powerful tech company on earth can't force people to hang out somewhere they don't want to be.

Hubris8/10
Burn6/10
Lesson8/10
Total22/30Devastating
RIP
#11

Jawbone

Couldn't survive the Apple Watch

Peak: $3.2B19992017

What Went Wrong

Jawbone was once the hottest wearable tech company in the world, making Bluetooth speakers and fitness trackers. The UP fitness band was a hit — until it wasn't. Build quality was terrible. Bands cracked, batteries died, syncing failed. Fitbit ate their lunch with a cheaper, more reliable product. Then Apple launched the Apple Watch and made the entire fitness band category obsolete. Jawbone burned through $930 million in funding and liquidated without returning a cent to investors.

The Lesson

Hardware is unforgiving. One generation of bad build quality and you lose your customers forever. And when Apple enters your category, you'd better have a moat deeper than 'we were here first.'

Hubris6/10
Burn8/10
Lesson8/10
Total22/30Devastating
RIP
#12

Pets.com

The sock puppet went harder than the business model

Peak: $300M19982000

What Went Wrong

Pets.com became the poster child of the dot-com bubble. They spent $11.8 million on a Super Bowl ad featuring a sock puppet mascot. They were selling 25-pound bags of dog food online and shipping them at a loss. Revenue for the entire year was $619,000 but they spent $11.8 million on marketing in a single quarter. The sock puppet had more brand equity than the company had revenue.

The Lesson

Brand awareness without unit economics is just expensive performance art. The sock puppet lives on as a cultural icon. The company lasted 268 days after its IPO.

Hubris6/10
Burn6/10
Lesson9/10
Total21/30Severe
RIP
#13

Solyndra

$535M government loan, bankrupt in two years

Peak: $535M (DOE loan guarantee)20052011

What Went Wrong

Solyndra manufactured cylindrical solar panels using a CIGS thin-film process that was innovative but expensive. They received a $535 million government loan guarantee under the Obama administration's clean energy initiative. Then Chinese manufacturers flooded the market with cheap silicon panels, cratering prices by 75%. Solyndra's cost structure was built for a world where silicon panels stayed expensive. They didn't.

The Lesson

Government backing doesn't insulate you from market forces. If your entire business plan depends on competitors staying expensive, you don't have a business plan — you have a prayer.

Hubris6/10
Burn7/10
Lesson8/10
Total21/30Severe
RIP
#14

Segway

Was supposed to change cities. Ended up in mall cop memes.

Peak: $500M (pre-launch hype)19992020

What Went Wrong

Before launch, Steve Jobs said it was 'as big a deal as the PC.' John Doerr predicted it would reach $1 billion in revenue faster than any company in history. Then it launched and it was... a $5,000 scooter. Cities didn't redesign infrastructure. Consumers didn't replace cars. The Segway became a symbol of overhyped technology — used primarily by mall security guards and tourist groups.

The Lesson

Pre-launch hype is not market validation. The product worked fine. The assumption that cities would reorganize around it was delusional.

Hubris8/10
Burn5/10
Lesson7/10
Total20/30Severe
RIP
#15

Vine

Had it all, lost it, and gave TikTok the blueprint

Peak: $30M (Twitter acquisition)20122017

What Went Wrong

Twitter bought Vine for $30 million before it even launched. It became the defining social platform of 2013-2015 — six-second loops that birthed an entire generation of internet culture. Then Twitter refused to pay creators, wouldn't build monetization tools, and ignored every feature request. Vine's top creators left for YouTube and Instagram. Twitter killed the app in 2016. TikTok launched in 2017 and built the exact product Vine should have become.

The Lesson

Creators are your product. If you won't pay them, someone else will. Twitter had a five-year head start on short-form video and squandered it completely.

Hubris4/10
Burn3/10
Lesson10/10
Total17/30Severe

Glen's Take

I've read the postmortem on every single one of these companies. The pattern is almost always the same: a charismatic founder with a big vision, investors who fell in love with the story instead of the spreadsheet, and a board that didn't ask hard enough questions until it was too late.

The thing is — ambition isn't the problem. Every successful company looked delusional at some point. Amazon lost money for seven years. Tesla was 48 hours from bankruptcy in 2008. SpaceX had three failed launches before the fourth worked. The difference between a visionary and a cautionary tale is usually execution discipline and intellectual honesty.

WeWork wasn't wrong that people want better offices. FTX wasn't wrong that crypto needed a good exchange. Webvan wasn't wrong that people would order groceries online. The ideas were often right. The execution — the burn rate, the governance, the ethics, the timing — that's where they died.

Failure is the tuition of entrepreneurship. But you don't have to pay it yourself. Study these tombstones carefully. The $50 billion they burned collectively is your free education.

Get Glen's Musings

Occasional thoughts on AI, Claude, investing, and building things. Free. No spam.

Unsubscribe anytime. I respect your inbox more than Congress respects property rights.

More From Glen

Essential Reading

FAQ

What is the biggest startup failure of all time?

By peak valuation, WeWork holds the record at $47 billion. The company's failed IPO in 2019 and eventual bankruptcy in 2023 represented one of the largest destructions of private market value in history. FTX ($32 billion) and Theranos ($9 billion) round out the top three.

What was wrong with Theranos?

Theranos claimed it could run over 200 blood tests from a single finger prick. The technology never actually worked. The company ran most tests on commercial machines while telling investors otherwise. Founder Elizabeth Holmes was convicted of fraud in 2022 and sentenced to over 11 years in prison.

Why did FTX collapse?

FTX collapsed in November 2022 after reports that Alameda Research's balance sheet was largely composed of FTX's own FTT token. This triggered $6 billion in withdrawals over 72 hours. Sam Bankman-Fried was convicted of fraud and sentenced to 25 years in prison.

What startup burned the most money?

WeWork burned through over $14 billion in total funding. Webvan burned $830 million in under two years. Jawbone went through $930 million. But in terms of speed, MoviePass was burning $40 million per month at its peak — roughly $1.3 million per day.

Know a founder who needs to see this?

Keep Exploring