Post-mortems
Why startups die
The successes get written about because they worked, which makes them a badly biased sample. These 12 failures are the more instructive half — what actually went wrong, what was genuinely visible at the time, and which specific diligence check would have surfaced it.
How to read this section
Every page separates three things that usually get blurred together: what happened, what was visible at the time, and what you should do differently. The middle one is the point. Hindsight makes every failure look obvious, and a signals list that could only have been written afterwards is entertainment rather than education.
We also keep fraud separate from failure. Most companies here failed without anyone lying, and treating those founders as though they had would be both unfair and useless. The fraud pages state what was charged and proven, with dates — and note where a sentence has since been reduced, pardoned or commuted, because several have.
Finally, each page links to the diligence guide that would most likely have caught the problem. The failures are the argument for the checklists.
4 companies
The numbers were not real
Companies where investors were deceived, and courts have said so. The useful question is not how bad the people were but what diligence would have surfaced.
Theranos
A blood-testing company valued in the billions on technology that did not work, whose founder was convicted of defrauding investors of hundreds of millions of dollars.
Read the post-mortem →Sold for $175m 2021 · founder sentenced 2025Frank
A student financial-aid startup sold to JPMorgan Chase for $175m on a customer base that was largely fabricated — the clearest recent case of diligence failing at the acquisition, not the investment.
Read the post-mortem →Founder convicted 2022 · pardoned 2025Nikola
An electric and hydrogen truck company whose founder was convicted of misleading investors about the company's progress — and who was subsequently pardoned.
Read the post-mortem →Founder convicted 2024 · sentence commuted 2025Ozy Media
A digital media company whose founder was convicted of defrauding investors and lenders by misrepresenting its finances, in a case that turned on an impersonated phone call.
Read the post-mortem →3 companies
Nobody wanted it enough
Products that worked, built by competent people, for a demand that was assumed rather than demonstrated.
Quibi
A short-form mobile video service that raised roughly $1.75bn, launched in April 2020 and closed about six months later — the largest and fastest consumer failure of its era.
Read the post-mortem →Shut down 2017 · over $120m raisedJuicero
A connected juice press that raised more than $120m from well-regarded investors, and was undone by a demonstration that the juice packs could be squeezed by hand.
Read the post-mortem →Bankrupt 2013 · over $850m raisedBetter Place
An electric vehicle battery-swapping network that raised more than $850m and filed for bankruptcy in Israel in 2013, having built infrastructure ahead of the demand it required.
Read the post-mortem →3 companies
The model never worked
Real customers, real revenue, and unit economics that got worse rather than better as the company grew.
Convoy
A digital freight marketplace, backed at scale by the best-known investors in technology, that shut down in November 2023 with its assets acquired by a competitor.
Read the post-mortem →Wound down 2023 · $4bn peak valuationOlive AI
A healthcare automation company valued at $4bn that wound down in 2023 because the technology did not deliver the savings it had promised customers.
Read the post-mortem →IPO withdrawn 2019 · Chapter 11 in 2023WeWork
A company once valued at $47bn that withdrew its public offering in 2019 after disclosure revealed the economics, and filed for Chapter 11 protection in November 2023.
Read the post-mortem →2 companies
It needed more than the market would give
Capital-intensive businesses whose funding requirement outlived the funding environment that started them.
Katerra
A construction technology company that spent more than $2bn in little over six years attempting to industrialise building, and filed for bankruptcy in June 2021.
Read the post-mortem →Bankruptcy March 2025 · a lost contract before itNorthvolt
Europe's flagship battery manufacturer, funded on an enormous scale to reduce the continent's dependence on Asian cells, which filed for bankruptcy in Sweden in March 2025.
Read the post-mortem →Startup failures: common questions
What is the most common reason startups fail?
Could investors really have seen these coming?
Why does this section cover fraud separately?
Are failure stories actually useful to an angel?
Put it to work
Failures are the argument for checklists
Almost every problem on these pages was findable with a question somebody did not ask. The diligence desk is that set of questions, organised by stage and by sector.