Why Startups Fail: What the Data Actually Shows

Key Takeaways

#1 killer is PMF, not cash: According to CB Insights' analysis of 431 failed VC-backed startups, 43% failed due to poor product-market fit, not running out of money.
Cash is a symptom: "Ran out of capital" shows up in 70% of failures, but CB Insights explicitly calls it the final cause, not the root problem.
Premature scaling is the silent assassin: According to the Startup Genome Report, 74% of high-growth startups fail due to premature scaling, and those that scale properly grow 20x faster.
Funding doesn't save you: The 431 startups in CB Insights' dataset raised a combined $17.5B before dying. The median company raised $11M.
The clock starts at your last raise: Median time from last fundraise to shutdown is just 22 months, per CB Insights.

About 90% of startups fail eventually.

You've heard that stat so many times it stopped landing. Here's one that should: the 431 failed VC-backed startups in CB Insights' latest analysis raised a combined $17.5B in equity funding before dying. The median company raised $11M.

Money didn't save them. So what killed them?

The data gives a clear, uncomfortable answer. And it's not the one most founders want to hear.

The Real Reason Is Not What Founders Claim

When a startup dies, "ran out of cash" is the most common line in the post-mortem.

CB Insights analyzed post-mortems, founder interviews, and shutdown announcements from 431 VC-backed companies that shut down since 2023. "Ran out of capital" tops the list at 70%, but it's almost always the final cause of death, not the root problem.

Think of it like a car crash. The coroner says "blunt force trauma." That's technically accurate. It tells you nothing useful about avoiding the crash.

Capital running out is where these stories end. The more telling causes, poor product-market fit (43%), bad timing (29%), and unsustainable unit economics (19%), reveal why the capital dried up in the first place.

Startup failure reason is almost always upstream of the fundraising cliff. Running out of money is the mechanism of death. Building something nobody wants is why death was coming.

The Consistent Finding Across Every Dataset

CB Insights has now run this analysis twice, on very different sample sizes. The numbers barely moved.

The original stat comes from CB Insights, which analyzed 110+ startup post-mortems between 2014 and 2021. Founders wrote essays about why their companies died. "No market need" topped the list at 42%. In 2024, CB Insights updated the study with 4x more data. They analyzed 431 failed VC-backed companies that shut down since 2023. The headline barely changed: 43% failed due to poor product-market fit.

Same finding. Four times the data. A decade apart. That's not noise. That's signal.

A parallel academic analysis published in the Journal of Economic Growth and Entrepreneurship reached the same conclusion.

Researchers analyzed 353 startup post-mortem reports from the CB Insights platform using 65 failure factors. Results indicate that failure factors related to product/market misfit, lack of capital, great power of competition, law and regulation problems, and bad business model appear as the most important factors that lead startups to fail.

What the Full Breakdown Looks Like

Here is how startup failure reasons stack up across the most cited data sources. Percentages exceed 100% because most startups cite multiple causes:

Failure Reason	Share of Failures
Poor product-market fit	43%
Ran out of capital (symptom)	70%
Bad timing	29%
Unsustainable unit economics	19%
Wrong team	23%
Got outcompeted	19%
Pricing/cost issues	18%
Lack of focus	13%

Two-thirds of product-market fit failures were early-stage companies that never found a market. But 20 Series B+ companies also cited poor PMF as a primary cause.

This is the part that surprises people. PMF failure isn't just a seed-stage problem. Companies can raise a Series B and still be building something the market doesn't urgently need.

The Premature Scaling Trap Nobody Talks About Loudly Enough

Here is the failure mode that CB Insights data understates, because it hides inside every other category.

The Startup Genome Project found that 74% of high-growth internet startups fail due to premature scaling, and 93% of startups that scale prematurely never break the $100k revenue per month threshold.

Premature scaling means investing in customers, product features, headcount, or distribution before you've validated that the core model works. It feels like growth. It looks like execution. It is actually the fastest way to efficiently destroy capital.

Since February, the Startup Genome Project amassed a dataset of over 3,200 high-growth technology startups. Their research found that the primary cause of failure is premature scaling, an affliction that 70% of startups in their dataset possess.

And startups that scale properly grow about 20 times faster than startups that scale prematurely.

Let that sink in. The same underlying problem that kills companies, rushing growth before validating the model, also explains why the surviving ones pull away so fast.

Case Study: What $900M Couldn't Save

The most vivid recent example of how startup failure data plays out in real life is Olive, the healthcare automation company.

Olive was a healthcare AI startup focused on automating administrative tasks like insurance verification and prior authorization checks. Founded in 2012, it raised over $850 million and was once valued at $4 billion. Despite its rapid rise, financial struggles and declining funding led to its shutdown in 2023.

Topping the chart with nearly $1B raised in funding are healthcare AI automation startup Olive and digital freight brokerage Convoy. Both hit approximately $4B valuations during pandemic-era booms, and both shut down within two weeks of each other in October 2023 as their markets turned.

Olive's CEO Sean Lane cited the real cause: fast-paced growth and lack of focus as key factors that strained Olive's resources.

While marketed as AI-driven automation, Olive relied heavily on manual intervention. Many processes required human oversight, contradicting its pitch of autonomous AI.

Olive hit nearly every failure mode on the list at once: overpromised PMF, premature scaling, lack of focus, unsustainable unit economics. Funding just delayed the reckoning.

Why Timing Is the Most Underrated Killer

Bad timing drove 29% of failures in CB Insights' post-2023 cohort, hitting climate, food, and blockchain startups hardest as markets failed to materialize after peak funding cycles.

The alt-protein wave is a clean example.

Food and agriculture accounts for 54 companies, or 13% of the shutdown cohort. Roughly a third of these are alt-protein or cultivated meat startups like Believer Meats ($390M raised) and Motif FoodWorks ($344M), pointing to a broader unwinding of the alternative protein wave.

These weren't bad companies building bad products. They were building ahead of consumer adoption curves and market infrastructure that never arrived on the timeline investors priced in. Timing isn't about luck. It's about correctly reading how fast a market is actually moving versus how fast you need it to move to survive.

The Warning Signs Show Up Before the End

Here is the part with real practical value for anyone evaluating startups.

While founders cite a range of causes in shutdown post-mortems, CB Insights' predictive signals show measurable deterioration in company health in the months leading up to shutdown. Among companies with full 12-month data, 72% saw their health score decline in the year before death, with scores dropping by 15% on average.

Failure isn't sudden. The signals accumulate. The companies that survive are the ones where founders or investors catch the pattern and act before the runway is gone.

The median time from last fundraise to death is 22 months. Over half of the companies in CB Insights' dataset died within 2 years of their last raise. However, nearly a quarter of startups had been "walking dead" for over 3 years since their last raise before officially going under.

There's a useful due diligence heuristic here. If a startup hasn't raised since 2022 or earlier, it is either genuinely self-sustaining or a zombie. There's not much in between.

Currently, there are nearly 50,000 VC-backed startups that haven't raised funding since the start of 2023.

What Serial Founders Know That First-Timers Don't

The failure rate isn't uniform across founders.

First-time founders face particularly steep odds, with only an 18% success rate. Those who have previously failed fare slightly better at 20%, while entrepreneurs who have already built a successful company enjoy the highest odds at 30%.

The reason isn't IQ or hustle. Serial founders are faster at recognizing PMF signals, earlier at cutting losing bets, and more disciplined about scaling timing. They've seen the movie. They know which chapter the disaster happens in.

For investors, this is one of the most reliable signals available. Founder track record and market experience separate the cohort more cleanly than almost any other variable.

How to Use This Data in Practice

The failure data points at three non-negotiable checkboxes for any startup worth backing:

Validate before building at scale. The 43% PMF failure rate means nearly half of all failed startups never confirmed the market wanted what they were selling. Talking to 20 customers before writing code is not optional.
Read the timing. Bad timing killed 29% of the post-2023 cohort. Ask: is the market moving fast enough that a company can reach sustainability before the funding window closes? This isn't pessimism, it's arithmetic.
Watch for premature scaling signals. Founders rush to expand teams, pursue multiple markets, or overbuild their product before laying essential operational foundations.

The 20x growth gap between properly and prematurely scaled companies is the difference between a winner and an expensive lesson.

This is exactly what a systematic evaluation framework helps with. Tools like Unicorn Screener are built to screen startups across these exact dimensions: product-market evidence, team quality, timing signals, and unit economics. Before committing capital, running a structured score is the fastest way to spot which failure mode a company is most exposed to. You can also browse the live leaderboard to see how current startups stack up across these criteria.

The red flags investors most often miss map almost perfectly onto the CB Insights failure taxonomy. PMF gaps, wrong team composition, and burn rates that assume growth that hasn't arrived yet. The data doesn't change year over year because human psychology doesn't change: founders fall in love with their idea, investors fall in love with the narrative, and markets stay stubbornly indifferent.

What This Means for You

Screen for PMF evidence first. Not pitch decks, not TAM slides. Real paying customers, real retention data, real pull from the market.
Treat "ran out of cash" as a diagnosis, not a cause. When a startup dies citing cash, ask what burned it. The answer is almost always one of the four root causes in the CB Insights data.
Check the fundraise timeline. A last-raise date from 2022 is a major flag in 2026. Either explain why they're self-sustaining, or move on.
Score your next deal systematically. Try Unicorn Screener for a research-backed evaluation across the dimensions the data actually says matter.

No model is perfect.

According to Professor Shikhar Ghosh at Harvard Business School, about 75% of venture-backed startups fail to return their investors' capital. However, his study is based on a sample of startups that received venture capital funding, which represents only a fraction of all startups.

All failure data has survivorship and reporting bias baked in. Use it as a framework, not a formula.

The startups that beat the odds aren't the ones with the most money. They're the ones that figured out the market wanted what they were building, before the runway ran out.

Want to screen startups like a top-tier VC? Score any startup for free with our research-backed evaluation model.