Failed startups
57 startups that failed, and what each one teaches product teams.
Post-mortems are the cheapest education in product management. Every company on this page had smart founders, real customers, and — in most cases — hundreds of millions of dollars. They still failed, and almost always for reasons that were visible early to anyone watching the right numbers.
Each deep dive covers what the company built, the biggest reasons it failed, the product lesson worth keeping, and the early warning sign a PM could have spotted. The list runs from the dot-com crash to 2025, from Silicon Valley to Bengaluru, and covers failures of strategy, economics, technology, law, and integrity.
- startups dissected
- 57
- years of collapses
- 1999–2025
- countries
- 9
- recurring failure causes
- 12
Five lessons that repeat across the list
Read enough post-mortems and the same patterns keep surfacing, regardless of decade, country, or sector.
Lesson 1
Scale amplifies unit economics; it rarely fixes them
Webvan, Kozmo, MoviePass, and Bird all assumed that losses per order, ticket, or ride would shrink with volume. Fixed costs do get spread across more units — but when each unit loses money, more units just lose more money. Check contribution margin before you celebrate growth.
Lesson 2
Capital amplifies whatever is already true
Quibi, Katerra, Fast, and Color raised enormous sums before proving their core behaviour. The money didn't buy product–market fit; it bought headcount, expectations, and burn. Funding accelerates a model that works — and the collapse of one that doesn't.
Lesson 3
Trust is a product feature
FTX and Synapse lost money people believed was safe. Cruise lost its regulators' confidence after an incident; Yik Yak lost schools and parents. In finance, health, and safety-critical products, controls and transparency aren't overhead — they are part of what customers are paying for.
Lesson 4
Numbers you can't verify will eventually be verified for you
Theranos, IRL, GoMechanic, and Builder.ai each told a story their data couldn't support. The truth surfaced at the worst possible moment — a fundraise, an audit, an investigation. Honest metrics are slower to impress and far cheaper to defend.
Read: Theranos, IRL, GoMechanic, Builder.ai
Lesson 5
Being early is a way of being wrong
General Magic, Iridium, and Better Place were right about where the world was heading and wrong about when. If your product depends on infrastructure, standards, or partners you don't control, track them as closely as your own roadmap.
Read: General Magic, Iridium, Better Place
How startups die: 12 recurring causes
Most companies here failed for more than one reason, so counts add up to more than 57. Ordered by how often each cause appears.
Scaled too early
21 startupsNew cities, countries, factories, or product lines were added before the first one worked. Expansion multiplied complexity and burn while the core model was still unproven.
Broken unit economics
19 startupsEvery order, ride, or subscriber cost more to serve than it earned, so growth made losses bigger instead of smaller. Scale fixes fixed costs; it does not fix a negative margin on each unit.
e.g. Dunzo, Cazoo, Babylon Health
Ran out of runway
18 startupsThe business still needed outside money to reach profitability when capital markets closed — after the dot-com crash, or the 2022–2024 funding winter.
Technology didn't deliver
17 startupsThe core technology wasn't ready, didn't scale, or couldn't be made reliably — and the business plan depended on it working on a schedule.
e.g. Builder.ai, Humane, Lilium
Out-competed
14 startupsA better-funded rival, a platform owner, or an entrenched network effect won the market. Being first or being better-designed was not enough.
Governance failure
11 startupsBoards, controls, and checks didn't work: concentrated founder power, related-party deals, missing accounts, or nobody able to say where the money was.
e.g. BluSmart, Builder.ai, Byju's
No real market need
10 startupsThe product solved a problem customers didn't feel strongly enough to change behaviour or pay for — often because a cheaper, simpler alternative already did the job.
e.g. Humane, Hyperloop One, Zume
Fraud & misreported numbers
9 startupsRevenue, users, technology, or the safety of customer funds were misrepresented. When outside scrutiny arrived — a reporter, auditor, or new investor — the company collapsed with the story.
e.g. Builder.ai, Nikola, GoMechanic
Lost customer trust
9 startupsA safety incident, frozen funds, abuse, or broken promises destroyed the confidence users, regulators, or partners needed to keep using the product.
Wrong timing
7 startupsThe idea was right but early or late: the enabling infrastructure didn't exist yet, the market moved during a long build, or a temporary boom was mistaken for the new normal.
Legal & regulatory
7 startupsThe business depended on a legal grey area, or on operating before permission was granted. When a court or regulator acted, the model stopped working.
No business model
4 startupsLots of users, but no credible way to make money from them. When growth stalled or funding dried up, there was no revenue to fall back on.
All 57 deep dives
Filter by why they failed, or by sector. Most recent failures first.
- Mobility & EV2019–2025India
BluSmart
All-electric ride-hailing service in Indian cities
BluSmart offered all-electric rides with no surge pricing and built a loyal following in Delhi NCR, Bengaluru, and Mumbai. Many of its cars were leased from Gensol Engineering, a listed company controlled by the same founders. When India's market regulator alleged in April 2025 that loans meant for those EVs had been diverted, BluSmart lost its fleet financing and stopped operating.
Governance failureRan out of runwayRead the deep dive →
- Enterprise & B2B2016–2025United Kingdom
Builder.ai
AI-assisted app development platform backed by Microsoft
Builder.ai said its AI-assisted platform made building a custom app as easy as ordering a pizza, raising about $450 million from investors including Microsoft. In 2025 a review sharply cut its reported revenue, a lender seized funds, and it entered insolvency.
Fraud & misreported numbersGovernance failureRead the deep dive →
- Commerce & Delivery2014–2025India
Dunzo
Indian hyperlocal delivery app that pioneered quick commerce
Dunzo began in Bengaluru as a WhatsApp-based errand service and became India's best-known hyperlocal delivery app. It pivoted to fast grocery delivery after a large Reliance investment, but Blinkit, Zepto, and Swiggy Instamart scaled faster. Cash ran out, salaries were delayed, and operations stopped in 2025.
Out-competedBroken unit economicsRead the deep dive →
- Media & Social2012–2025India
Hike
Indian messaging app known for stickers and local-language features
Hike became an Indian unicorn by offering stickers and features tailored to young Indian users. WhatsApp's network effect proved overwhelming, and Hike closed its messenger in 2021 to pivot to real-money gaming — a category India then banned outright in 2025.
Out-competedNo business modelRead the deep dive →
- Hardware & Devices2018–2025United States
Humane
AI-powered wearable pin designed to replace smartphone tasks
Former Apple designers launched the Humane AI Pin in 2024 as a screenless, voice-first assistant worn on clothing. Reviewers found it slow, unreliable, and prone to running hot. Sales were weak, and HP bought Humane's assets in 2025, shutting the devices down.
No real market needTechnology didn't deliverRead the deep dive →
- Mobility & EV2015–2025Germany
Lilium
German electric vertical take-off jet designed for air taxi services
Lilium developed an electric jet that could take off vertically, for regional air-taxi flights. Development and certification costs ran into billions; after Germany declined to back a loan, it filed for insolvency in 2024, and a rescue deal collapsed in early 2025.
Ran out of runwayTechnology didn't deliverRead the deep dive →
- Mobility & EV2014–2025United States
Nikola
Hydrogen and battery-electric truck maker
Nikola promised hydrogen-powered trucks and went public in 2020, briefly worth more than Ford. A short-seller report revealed that a promotional video of a truck in motion showed it rolling downhill. Its founder was convicted of fraud; the company struggled to deliver trucks and filed for bankruptcy in 2025.
Fraud & misreported numbersTechnology didn't deliverRead the deep dive →
- Energy & Climate2016–2025Sweden
Northvolt
Swedish battery maker meant to be Europe's answer to Asian cell giants
Northvolt raised more than $14 billion to build battery gigafactories in Sweden and planned more in Germany and Canada. Its production ramp-up struggled with low output and quality, BMW cancelled a major order, and with European EV demand slowing, it ran out of money and went bankrupt in 2025.
Technology didn't deliverScaled too earlyRead the deep dive →
- Edtech2011–2024India
Byju's
Indian edtech giant built on a learning app and an aggressive sales force
Byju's rode the pandemic boom in online learning to a $22 billion valuation, then spent heavily on acquisitions, sponsorships, and a high-pressure sales machine. As demand normalized, delayed accounts, the exit of its auditor and board members, and disputes with lenders followed, and it entered insolvency proceedings in 2024.
Scaled too earlyGovernance failureRead the deep dive →
- Commerce & Delivery2018–2024United Kingdom
Cazoo
UK online used-car retailer that delivered cars to customers' doors
Cazoo bought, reconditioned, and sold used cars online with home delivery. It grew fast on heavy marketing, went public via SPAC at a $7 billion valuation, expanded across Europe, and made too little on each car. It retreated, restructured, and entered administration in 2024.
Broken unit economicsScaled too earlyRead the deep dive →
- Mobility & EV2013–2024United States
Cruise
GM-backed robotaxi company running driverless cars in San Francisco
Cruise ran driverless robotaxis in San Francisco. In October 2023 one of its cars dragged a pedestrian who had been struck by another vehicle. Regulators said Cruise didn't promptly share the full picture, suspended its permit, and GM ended its robotaxi funding in December 2024 after investing more than $10 billion.
Lost customer trustLegal & regulatoryRead the deep dive →
- Mobility & EV2016–2024United States
Fisker
EV maker founded by car designer Henrik Fisker — twice
Henrik Fisker's second EV company outsourced manufacturing of its Ocean SUV to Magna. Deliveries began in 2023, but software problems, recalls, and safety investigations hurt sales, a rescue partnership with Nissan collapsed, and it filed for bankruptcy in June 2024 — eleven years after his first EV company failed.
Technology didn't deliverRan out of runwayRead the deep dive →
- Media & Social2020–2024India
Koo
Indian multilingual microblogging platform positioned as a Twitter alternative
Koo offered microblogging in Indian languages and surged in 2021 amid a dispute between Twitter and the Indian government. Usage later declined, it never built meaningful revenue, and when funding dried up and acquisition talks failed, it shut down in July 2024.
No business modelRan out of runwayRead the deep dive →
- Fintech & Crypto2014–2024United States
Synapse
Banking-as-a-service middleware connecting fintech apps to partner banks
Synapse sat between consumer fintech apps and partner banks, keeping the ledgers that recorded whose money was whose. When it collapsed in 2024, those records didn't reconcile with bank balances, freezing funds for more than 100,000 people and revealing a shortfall of tens of millions of dollars.
Governance failureLost customer trustRead the deep dive →
- Health2013–2023United Kingdom
Babylon Health
Digital health company with an AI symptom checker and virtual GP appointments
Babylon offered video GP consultations and an AI symptom checker, including through an NHS partnership. It listed in New York via SPAC at $4.2 billion and expanded into US contracts where it bore the financial risk of patients' care. Losses mounted, and it collapsed in 2023.
Broken unit economicsScaled too earlyRead the deep dive →
- Mobility & EV2017–2023United States
Bird
Dockless electric scooter sharing that reached a $2.5B valuation
Bird put shared e-scooters on streets worldwide and became a unicorn in about a year. Early scooters wore out quickly, cities restricted permits, and dozens of competitors drove prices down. It went public via SPAC in 2021 and filed for bankruptcy in December 2023.
Broken unit economicsLegal & regulatoryRead the deep dive →
- Enterprise & B2B2015–2023United States
Convoy
Digital freight brokerage matching shippers with truckers
Convoy used software to match shippers' loads with truckers and cut empty miles. It was valued at $3.8 billion in 2022, but a sharp freight recession crushed rates and margins, investors pulled back, and it shut down in October 2023.
Ran out of runwayWrong timingRead the deep dive →
- Mobility & EV2016–2023India
GoMechanic
Indian car-servicing network built on partner garages
GoMechanic built a network of branded partner garages offering standardized car servicing across Indian cities. During due diligence for a planned SoftBank-led round, problems surfaced in its financial reporting; a co-founder admitted to 'grave errors of judgement', the round collapsed, and most staff were laid off.
Fraud & misreported numbersScaled too earlyRead the deep dive →
- Mobility & EV2014–2023United States
Hyperloop One
Startup developing high-speed pods in low-pressure tubes
Hyperloop One set out to build the hyperloop concept popularized by Elon Musk: pods moving at airliner speeds through low-pressure tubes. It built a test track and ran one passenger demo, but never won a commercial project, and it closed in December 2023.
Technology didn't deliverNo real market needRead the deep dive →
- Media & Social2017–2023United States
IRL
Social app for events and group chats, valued at $1.17B
IRL raised about $200 million and reached a $1.17 billion valuation by presenting itself as a viral, fast-growing social app. After its founder was suspended, a board investigation found about 95% of its users were automated or from bots, and the company shut down.
Fraud & misreported numbersGovernance failureRead the deep dive →
- Health2012–2023United States
Olive AI
Healthcare automation company selling AI 'digital workers' to hospitals
Olive raised over $850 million to automate hospital administration — billing, claims, and prior authorizations — with software 'digital workers'. Customers reported results that fell short of the pitch, and Olive expanded into many workflows at once. It wound down in 2023, two years after a $4 billion valuation.
Scaled too earlyTechnology didn't deliverRead the deep dive →
- Real Estate & Construction2010–2023United States
WeWork
Flexible office space provider valued at $47B as a 'tech company'
WeWork signed long leases on buildings and rented desks out on short terms, while presenting itself as a technology company. A failed IPO in 2019 exposed huge losses and governance problems. It later listed via SPAC, then filed for bankruptcy in 2023 when its lease costs became unsustainable.
Broken unit economicsGovernance failureRead the deep dive →
- Commerce & Delivery2015–2023United States
Zume
Robot-assisted pizza baked in delivery trucks, later a packaging company
Zume used robots to help make pizzas and trucks with ovens to bake them on the way to customers. After a $375 million SoftBank investment it chased a bigger platform vision, then pivoted to compostable packaging, but none of its businesses reached sustainable scale.
Scaled too earlyNo real market needRead the deep dive →
- Mobility & EV2016–2022United States
Argo AI
Self-driving technology startup backed by Ford and Volkswagen
Argo AI was developing self-driving systems with billions from Ford and Volkswagen. In 2022 Ford said profitable, fully driverless vehicles at scale were much further away than expected, and the two carmakers stopped funding Argo, which closed that October.
Technology didn't deliverRan out of runwayRead the deep dive →
- Fintech & Crypto2017–2022United States
Celsius Network
Crypto lending platform that promised high yields on deposits
Celsius took customers' crypto and paid high yields, marketing itself as safer than a bank. Behind the scenes it made risky, illiquid investments. When crypto prices crashed in 2022 it froze withdrawals and went bankrupt; its founder later pleaded guilty to fraud.
Fraud & misreported numbersLost customer trustRead the deep dive →
- Fintech & Crypto2019–2022United States
Fast
One-click checkout startup for online stores
Fast built one-click checkout for online merchants and raised about $124 million, including a round led by Stripe. It grew to hundreds of employees while reportedly earning around $600,000 in revenue in 2021, and shut down in April 2022 when new funding didn't come.
Scaled too earlyNo real market needRead the deep dive →
- Fintech & Crypto2019–2022Bahamas
FTX
Crypto exchange valued at $32 billion at its peak
FTX grew into one of the world's largest crypto exchanges. In November 2022, reporting on its sister trading firm Alameda Research triggered a run on deposits, revealing that customer funds had been used by Alameda. FTX filed for bankruptcy within days, and its founder was convicted of fraud.
Fraud & misreported numbersGovernance failureRead the deep dive →
- Fintech & Crypto2018–2022South Korea / Singapore
Terraform Labs
Creator of the TerraUSD algorithmic stablecoin and the Luna token
TerraUSD was a stablecoin meant to hold $1 through an algorithm tying it to the Luna token, rather than through cash reserves. An affiliated platform paid close to 20% yields on it. In May 2022 large withdrawals broke the peg, the mechanism spiralled, and both tokens collapsed.
Technology didn't deliverFraud & misreported numbersRead the deep dive →
- Real Estate & Construction2015–2021United States
Katerra
SoftBank-backed startup that tried to vertically integrate construction
Katerra set out to fix construction's productivity problem by owning the whole process — design, materials, factories, and on-site building. Backed by around $2 billion from SoftBank, it expanded faster than it could execute and filed for bankruptcy in 2021.
Scaled too earlyBroken unit economicsRead the deep dive →
- Media & Social2011–2020United States
MoviePass
Movie-ticket subscription that dropped to $9.95 a month for nearly unlimited films
In 2017 MoviePass cut its price to $9.95 a month for up to one film a day, while paying cinemas close to full price for each ticket. Subscribers surged into the millions — and so did losses, because the more people used the service, the more money it lost.
Broken unit economicsRan out of runwayRead the deep dive →
- Media & Social2018–2020United States
Quibi
Mobile-only streaming service for premium shows in 10-minute 'quick bites'
Quibi raised $1.75 billion before launch to produce star-studded shows in episodes under ten minutes, watchable only on phones. It launched as the pandemic kept people at home, was hard to share, and competed against free short video. It shut down six months after launch.
No real market needWrong timingRead the deep dive →
- Mobility & EV1999–2020United States
Segway
Self-balancing personal transporter hyped to reinvent how cities move
Before anyone had seen it, the Segway was described as a world-changing invention. When it launched at around $5,000, it turned out to be too expensive for most consumers, unwelcome on many sidewalks, and unnecessary for most trips people could walk, cycle, or drive.
No real market needLegal & regulatoryRead the deep dive →
- Mobility & EV2014–2019China
ofo
Dockless bike sharing that put millions of yellow bikes on streets worldwide
ofo let users unlock yellow bikes with a phone and leave them anywhere. Backed by billions, it raced rival Mobike to flood cities in China and abroad. Bikes broke, were stolen, or piled up, fares stayed near zero, and the company collapsed owing deposits to millions of users.
Scaled too earlyBroken unit economicsRead the deep dive →
- Health2012–2019United States
uBiome
Consumer and clinical gut-microbiome testing startup
uBiome began by selling at-home microbiome kits to curious consumers, then pivoted to clinical tests billed to health insurers. Prosecutors allege it billed for unnecessary tests and re-tested old samples to inflate revenue; its co-founders were indicted in 2021.
Fraud & misreported numbersGovernance failureRead the deep dive →
- Health2003–2018United States
Theranos
Promised hundreds of blood tests from a single finger-prick
Theranos claimed its compact machines could run a wide range of blood tests on a few drops of blood. They couldn't; the company ran many tests on modified commercial equipment, and results were unreliable. Investigative reporting exposed it, regulators acted, and its leaders were convicted of fraud.
Fraud & misreported numbersTechnology didn't deliverRead the deep dive →
- Hardware & Devices1999–2017United States
Jawbone
Maker of Bluetooth headsets, Jambox speakers, and UP fitness bands
Jawbone built stylish consumer electronics and moved into fitness wearables with its UP band. Reliability problems, intense competition from Fitbit and Apple, and an expensive legal fight drained the company despite very large funding.
Technology didn't deliverOut-competedRead the deep dive →
- Hardware & Devices2013–2017United States
Juicero
Wi-Fi-connected juicer that pressed proprietary juice packs
Juicero sold a connected machine that pressed proprietary packs of chopped produce into juice. Launched at $699 and cut to $399, it was undermined when reporters showed the packs could be squeezed by hand into a glass almost as well.
No real market needBroken unit economicsRead the deep dive →
- Media & Social2013–2017United States
Yik Yak
Anonymous, location-based message board popular on college campuses
Yik Yak let people post anonymously to everyone within a few miles, and it spread rapidly across US campuses. Anonymity also enabled harassment and threats, schools pushed back, and when the company removed anonymity to fix it, users left.
Lost customer trustNo business modelRead the deep dive →
- Hardware & Devices2012–2016United States
Pebble
Kickstarter-funded smartwatch with long battery life and an e-paper screen
Pebble helped create the modern smartwatch category with a simple, affordable watch and record-breaking Kickstarter campaigns. When Apple and others entered and the market shifted toward fitness, thin hardware margins and limited capital left it unable to keep up.
Out-competedRan out of runwayRead the deep dive →
- Commerce & Delivery2014–2016India
PepperTap
On-demand grocery delivery from neighbourhood stores in Indian cities
PepperTap delivered groceries from local stores within hours and expanded rapidly across Indian cities. Heavy discounts drove orders, but small baskets, low grocery margins, and delivery costs meant many orders lost money, and it closed in 2016.
Broken unit economicsScaled too earlyRead the deep dive →
- Commerce & Delivery2010–2015United States
Fab.com
Flash-sale design store that reached a reported $1B valuation
Fab pivoted from a struggling social network to daily design flash sales and grew explosively. It then poured money into European expansion, acquisitions, and owned inventory, changed strategy repeatedly, and collapsed under its own costs.
Scaled too earlyBroken unit economicsRead the deep dive →
- Media & Social2002–2015United States
Friendster
Early social network that pioneered online friend connections
Friendster attracted millions of users within months of its 2003 launch, then failed to handle them: as the network grew, pages became painfully slow. While engineers struggled, Myspace and later Facebook offered faster, freer experiences, and US users left.
Technology didn't deliverOut-competedRead the deep dive →
- Commerce & Delivery2012–2015United States
Homejoy
On-demand home cleaning booked through an app
Homejoy offered home cleaning booked online, often with steep first-clean discounts. It acquired lots of customers but struggled to keep them at full price, many cleaners and customers could arrange jobs directly, and lawsuits over worker classification became the final blow.
Broken unit economicsLegal & regulatoryRead the deep dive →
- Hardware & Devices2009–2015United States
Quirky
Community-powered invention platform turning user ideas into products
Quirky let its online community submit and vote on invention ideas, then designed, manufactured, and sold the winners, paying inventors royalties. It launched hundreds of products and had some hits, but each one carried design, tooling, and inventory costs that most never earned back.
Broken unit economicsScaled too earlyRead the deep dive →
- Media & Social2010–2015United States
Rdio
Well-designed music streaming service from the founders of Skype
Rdio was praised for its design and social features and launched in the US before Spotify. But it couldn't grow its paid base against better-funded rivals, while licensing costs swallowed most of its revenue, and it ran out of money in 2015.
Out-competedBroken unit economicsRead the deep dive →
- Mobility & EV2011–2015United States
Sidecar
Peer-to-peer ridesharing pioneer that predated the UberX model
Sidecar launched peer-to-peer ridesharing in San Francisco in 2012 and pioneered ideas such as up-front pricing and shared rides. Uber and Lyft adopted similar models with vastly more capital, outspent it on incentives, and Sidecar closed at the end of 2015.
Out-competedRan out of runwayRead the deep dive →
- Media & Social2010–2014United States
Aereo
Streamed broadcast TV through thousands of tiny rented antennas
Aereo let subscribers watch and record over-the-air TV online by assigning each user a tiny individual antenna, an architecture designed so it wouldn't have to pay broadcasters. Broadcasters sued, and in 2014 the Supreme Court ruled that Aereo was publicly performing their content.
Legal & regulatoryRead the deep dive →
- Mobility & EV2007–2013Israel
Better Place
Electric-car network built around swapping batteries in minutes
Better Place planned to make electric cars practical with networks of charging spots and robotic stations that swapped a depleted battery for a full one. It spent heavily on infrastructure in Israel and Denmark, but only one car model was compatible, and only around 1,400 were ever deployed.
Scaled too earlyNo real market needRead the deep dive →
- Media & Social2010–2012United States
Color
Proximity-based photo-sharing app that raised $41M before launch
Color raised $41 million before launching an app that automatically shared photos with strangers nearby. On day one, most users opened it alone and saw an empty screen. The money bought a large team and huge expectations, but not a reason for anyone to keep using it.
No real market needScaled too earlyRead the deep dive →
- Energy & Climate2005–2011United States
Solyndra
Cylindrical thin-film solar panels backed by a US federal loan guarantee
Solyndra's distinctive cylindrical solar panels avoided expensive polysilicon, which made sense when silicon prices spiked. When polysilicon prices collapsed and cheap conventional panels flooded the market, its cost advantage vanished — just after it had built a large new factory.
Out-competedWrong timingRead the deep dive →
- Hardware & Devices1990–2002United States
General Magic
Apple spin-off building a handheld 'personal communicator' in the early 1990s
General Magic designed a handheld device with a touch interface, messaging, and downloadable software in the early 1990s. The wireless networks, chips, and batteries needed to make it useful didn't exist yet, and the devices built on its software sold poorly.
Wrong timingTechnology didn't deliverRead the deep dive →
- Commerce & Delivery1998–2001United States
Kozmo.com
Free one-hour delivery of videos, snacks, and essentials in US cities
Kozmo delivered DVDs, snacks, and convenience items within an hour, with no delivery fee and no minimum order, using couriers on bikes and scooters. Customers loved it; the cost of carrying a single candy bar across a city made sure it could never pay for itself.
Broken unit economicsScaled too earlyRead the deep dive →
- Media & Social1999–2001United States
Napster
Peer-to-peer music sharing that reached tens of millions of users
Napster let anyone download almost any song for free by sharing files directly between users' computers. It became one of the fastest-growing internet products ever, but it was built on copyright infringement, and the record industry sued it out of existence.
Legal & regulatoryNo business modelRead the deep dive →
- Commerce & Delivery1996–2001United States
Webvan
Online grocery delivery, built city by city with its own warehouses
Webvan promised groceries delivered within a 30-minute window and spent hundreds of millions building automated warehouses before it had proven customers would pay enough to cover them. When the dot-com market closed in 2000, it had no path to profit and no new capital.
Scaled too earlyBroken unit economicsRead the deep dive →
- Commerce & Delivery1998–2000United Kingdom
Boo.com
Global online fashion store with a cutting-edge, graphics-heavy website
Boo.com tried to launch a premium online fashion retailer in many countries and languages at once, with a graphics-heavy website that most shoppers' dial-up connections couldn't handle. Delays and overheads ate its funding before sales could arrive.
Scaled too earlyTechnology didn't deliverRead the deep dive →
- Commerce & Delivery1998–2000United States
Pets.com
Pet supplies sold online, fronted by a famous sock-puppet mascot
Pets.com spent heavily on brand advertising to sell heavy, low-margin pet food online, often below cost and with subsidized shipping. Its mascot became famous; its unit economics never worked, and it shut down less than a year after going public.
Broken unit economicsScaled too earlyRead the deep dive →
- Hardware & Devices1991–1999United States
Iridium
Global satellite phone network conceived and backed by Motorola
Iridium launched 66 low-orbit satellites so anyone could make a call from anywhere on Earth. By the time service opened in 1998, cellular networks covered the places most customers actually went, and its bulky phones and high call charges attracted only a sliver of the subscribers it needed.
Wrong timingBroken unit economicsRead the deep dive →
How this list was built
We included companies that raised significant capital or reached wide public recognition, and whose collapse — bankruptcy, liquidation, shutdown, or a sale far below peak value — is documented in public reporting, court records, or company statements. Funding and valuation figures are approximate and based on public reporting. Where legal cases are involved, convictions are described as convictions and charges as allegations. Spotted an error? Email hello@productmanagement.space.