Key Takeaways
- Approximately 20% of startups fail in year one; 45% by year five; 65% by year ten
- Running out of cash causes 38% of startup failures, making it the most common operational cause
- No market need is cited in 42% of post-mortems as a primary failure reason
- Venture-backed startups survive at roughly 25% higher rates than bootstrapped companies at the five-year mark
- Restaurant and retail startups fail at significantly higher rates than software companies
Startup failure rate statistics 2026: what the data actually shows
Everyone's heard the 90% statistic. Startups fail. Most of them. Quickly.
The problem with that number is that it collapses everything into one figure (timing, industry, funding stage, founder experience, country) and the result is more useful for cocktail party conversation than for understanding what actually kills companies.
The disaggregated data is more useful. Survival rates vary enormously by what kind of company you're running, how much money you raised, where you're located, and what year you're measuring. If you're building a company and want to know your actual odds, the headline number gets you nowhere.
Section 1: Failure rates by year
The most consistent source on startup survival rates comes from the U.S. Bureau of Labor Statistics (BLS), which tracks business survival across all new employer establishments. Their longitudinal data, updated through 2024, shows a predictable pattern:
| Year in business | Survival rate | Cumulative failure rate |
|---|---|---|
| Year 1 | ~80% | ~20% |
| Year 2 | ~70% | ~30% |
| Year 3 | ~62% | ~38% |
| Year 5 | ~55% | ~45% |
| Year 7 | ~44% | ~56% |
| Year 10 | ~35% | ~65% |
Source: U.S. Bureau of Labor Statistics, Business Employment Dynamics, 2024
The first year is dangerous, but the real test comes between years two and five. That's when initial momentum fades, capital gets tight, and the market gives its genuine verdict. Year one mostly weeds out companies that were never viable. Years two through five weed out companies that were viable but couldn't execute.
The 90% failure figure is often attributed to tech startups specifically. That's closer to accurate when you include companies that never raised a formal round: informal ventures, lifestyle businesses that plateau, and founder-led service companies that dissolve quietly. The BLS data covers all new employer businesses, so the picture is broader.
For pure venture-backed tech startups, data from CB Insights and Startup Genome suggests a 75% to 90% failure or acquisition rate over a 10-year horizon. The precise figure depends on how you define failure: whether that includes acqui-hires, pivots that abandon the original thesis, or just companies that run out of money and shut down.
What the survival curve actually means by stage
The curve is not linear. The early years are disproportionately dangerous, then the risk of acute failure decreases, but the risk of slow death from stagnation increases.
A company in year seven that's still generating revenue but has stopped growing faces a different problem than a year-one company that hasn't found product-market fit. Both end up as failures in the data, but for different reasons.
Section 2: Why startups fail
CB Insights has published the most widely cited failure analysis, based on post-mortems from over 300 failed startups. Their findings, updated through 2024:
| Reason for failure | % of failures citing this as primary cause |
|---|---|
| No market need | 42% |
| Ran out of cash | 29% |
| Wrong team | 23% |
| Outcompeted | 19% |
| Pricing or cost issues | 18% |
| Poor product | 17% |
| Lacked a business model | 17% |
| Poor marketing | 14% |
| Ignored customer feedback | 14% |
| Product mistimed | 13% |
| Lost focus | 13% |
| Founder or investor conflict | 13% |
| Pivot failure | 10% |
| Burnout | 9% |
| Regulatory challenge | 8% |
Source: CB Insights, "The Top Reasons Startups Fail," 2024 update
These percentages add up to more than 100% because most failures involve multiple causes. The average failed startup cites 3.4 reasons.
No market need: what it actually means
The 42% figure for "no market need" is frequently misread. It doesn't mean founders built something nobody wanted. Most of these founders had real early customers. It means they built something that wasn't a large enough need, at the right price point, in a market big enough to support a standalone company.
A product that a hundred customers will pay $50/month for is a $60,000/year business. If you raised $2M building it, that's not a business. It's a mismatch between solution quality and market size.
Cash as symptom versus cause
Running out of cash (29%) deserves separate treatment because it is often a symptom rather than a root cause. A company that runs out of cash because it couldn't attract customers ran out of cash because of a product or marketing failure. A company that runs out of cash during a fundraising round it expected to close ran out of cash because of a timing or investor relationship failure.
SCORE's 2024 Small Business Failure data adds texture: cash flow problems contribute to 38% of small business failures, and 82% of small businesses that fail cite cash flow as a contributing factor. The distinction between "cause" and "contributing factor" matters. Cash is almost always present at death, but it's rarely the original disease.
Section 3: Failure rates by industry
Startup survival rates vary significantly across industries. BLS Business Employment Dynamics data, combined with industry-specific surveys, gives the following five-year survival rates:
| Industry | 5-year survival rate | 5-year failure rate |
|---|---|---|
| Healthcare and social assistance | 60% | 40% |
| Finance and insurance | 58% | 42% |
| Real estate | 57% | 43% |
| Professional and technical services | 55% | 45% |
| Manufacturing | 52% | 48% |
| Retail trade | 47% | 53% |
| Accommodation and food service | 39% | 61% |
Source: U.S. Bureau of Labor Statistics, Business Employment Dynamics, 2024; National Restaurant Association, 2024
Tech startups: the real picture
Tech startups don't appear cleanly in BLS category data because they span multiple NAICS codes. CB Insights and Startup Genome data suggests venture-backed software companies have a five-year survival rate of about 40%, accounting for shutdowns, acqui-hires, and pivots that effectively end the original business.
Software startups that reach product-market fit survive at much higher rates than those that don't. The failure distribution is bimodal: either you find the fit and the curve improves dramatically, or you don't and you're almost certainly shutting down within 18 to 24 months of your initial capital.
Restaurants: the actual numbers
The restaurant failure rate is commonly quoted at 60% in year one. The actual first-year failure rate is closer to 17% (National Restaurant Association, 2024). The more accurate picture is that restaurant three-year failure rates reach around 60%, which is where the statistic gets distorted.
Restaurants do fail faster than most other business types. High fixed costs, thin margins, intense local competition, and labor dependency make the model unforgiving. But the year-one death rate is not 60%.
Section 4: Failure rates by funding stage
Funding stage is one of the clearest predictors of survival. Companies with external capital survive at significantly higher rates than bootstrapped companies, though the reasons are more complex than simply "more money equals more time."
Survival rates by funding type
| Funding type | 3-year survival rate | 5-year survival rate |
|---|---|---|
| No external funding (bootstrapped) | 58% | 34% |
| Friends and family only | 61% | 38% |
| Angel-funded | 67% | 44% |
| Seed VC-backed | 71% | 47% |
| Series A+ backed | 78% | 56% |
Sources: Crunchbase, 2024 Startup Funding Analysis; Kauffman Foundation, State of Startup Funding, 2024
The survival advantage of venture-backed startups isn't purely about cash runway. Venture-backed companies also have investor networks, board oversight, hiring assistance, and performance pressure that can function as forcing functions for finding product-market fit faster.
The failure mode for venture-backed companies is also different. A bootstrapped company that isn't working can wind down quietly over 18 months. A company that raised $10M and missed its metrics is often shut down or sold within 6 months once investors lose conviction.
The Series A mortality cliff
One of the more consistent patterns in the funding data is what CB Insights calls the "Series A crunch." Of seed-stage companies, only about 40% successfully raise a Series A. The rest either shut down, stay self-sustaining, or continue without additional institutional backing.
Of companies that do raise Series A, roughly 60% go on to raise Series B, and about 48% of Series B companies eventually reach a meaningful exit or sustainable scale.
The practical implication: if you've raised seed funding, the 18 months after your seed round are a make-or-break period. The companies that raise Series A are the ones that have demonstrated specific, measurable signals: usually some combination of revenue growth rate, retention, and market size evidence that gives institutional investors conviction.
Section 5: Bootstrapped vs. funded survival rates
The bootstrapped-versus-funded comparison needs more nuance than the survival rate table provides.
Bootstrapped companies: slower death, different failure modes
Bootstrapped companies fail at higher rates over five years, but they also have more flexibility to pivot, reduce scope, and survive in a small, sustainable form. A bootstrapped business generating $200,000 in annual revenue with one founder might look like a "failure" relative to its original ambitions but still represents ongoing work and income.
Funded startups face a harder binary. If you raised $3M at a $12M valuation, a $500,000 outcome is effectively a failure from the investor's perspective, even if it represents positive cash flow for the founder.
What the data shows about bootstrapped success
Indie.vc and Lighter Capital surveys of bootstrapped founders show:
- 64% of founders who bootstrapped past year three were still operating their original business five years later
- Bootstrapped founders report higher overall satisfaction (68%) compared to VC-backed founders (51%)
- Bootstrapped companies in professional services and software have the highest five-year survival rates, around 55-60%
The funding question is really a question about what you're optimizing for. Venture funding accelerates growth but also accelerates the binary outcome. Bootstrapping preserves optionality but limits speed.
Section 6: Operational cost mismanagement as a failure factor
Cost mismanagement sits underneath many failure labels. Companies that "ran out of cash" often ran out of cash because their cost structure grew ahead of their revenue. Companies cited for "pricing issues" often priced wrong because they didn't understand their actual cost structure.
SCORE's 2024 data is specific: 82% of failed small businesses cite cash flow problems as a contributing factor, and of those, 29% point specifically to unplanned operational costs as the primary trigger.
The costs that most often catch founders off guard:
| Cost category | % of founders who underestimated it | Average underestimation |
|---|---|---|
| Payroll and benefits (fully loaded) | 71% | 28% below actual |
| Legal and compliance | 68% | 42% below actual |
| Marketing and customer acquisition | 64% | 35% below actual |
| Technology infrastructure | 59% | 31% below actual |
| Recruiting and hiring | 57% | 38% below actual |
Sources: SCORE, 2024; SBA Small Business Statistics, 2025; Fundera, 2024 Startup Cost Survey
Payroll surprises are the most common. Most founders budget for gross salary but miss employer payroll taxes (7.65% on top of wages), benefits (typically 20-30% of gross wages), and recruiting costs (15-20% of first-year salary per hire). A $130,000 software engineer typically costs $165,000 to $185,000 per year fully loaded.
For a detailed breakdown of startup operational costs by category and stage, see the startup operations cost breakdown research.
Section 7: Geographic variations in startup failure rates
Geography matters, though not because some cities have more talented founders. Ecosystem density, access to capital, customer concentration, and talent pool availability all vary by location and all affect survival rates.
U.S. regional survival rates
Startup Genome's 2024 Global Startup Ecosystem Report scores regional ecosystems on startup output, connectedness, and funding access. Survival data roughly tracks those scores.
| Metro area | Estimated 5-year survival rate | Ecosystem rank (Startup Genome) |
|---|---|---|
| San Francisco Bay Area | 52% | 1 |
| New York City | 49% | 2 |
| Boston | 51% | 5 |
| Los Angeles | 46% | 7 |
| Austin | 44% | 12 |
| Chicago | 42% | 14 |
| Mid-tier metros (aggregate) | 38-40% | N/A |
| Rural and micropolitan areas | 31-35% | N/A |
Source: Startup Genome, 2024; Kauffman Foundation Regional Startup Activity Index, 2024
The Bay Area and Boston advantage is largely a capital and talent access story. Founders in those markets are closer to investors, have access to deeper engineering talent pools, and can often close enterprise deals faster because their target customers are nearby.
The gap between top-tier and mid-tier metros has narrowed somewhat since 2020, partly because remote work reduced the cost penalty of operating outside major startup hubs. But the access-to-capital gap remains real.
International failure rates
The U.S. data is the most consistently measured, but comparable figures exist for other major startup ecosystems.
- UK: 60% of startups fail within three years (Companies House, 2024)
- Canada: 54% of startups fail within five years (Industry Canada, 2024)
- India: 90% of startups fail within five years, with 71% shutting down in their first year (Tracxn, 2024), though India's definition of "startup" is broader and includes many informal ventures
- Germany: 44% of startups fail within three years (KfW Research, 2024)
- Australia: 60% of small businesses fail within three years (Australian Bureau of Statistics, 2024)
The India statistic is frequently cited without context. The 90% figure comes from a broader base of registered businesses, many of which are sole-trader service businesses that were never startups in the venture-track sense.
Section 8: What separates survivors from failures
Looking at post-mortem data from failed companies alongside case studies from survivors, a few consistent patterns emerge.
Revenue timing
Companies that generate their first dollar of revenue within 12 months of founding survive at significantly higher rates than those that don't. This isn't because early revenue is the goal in itself (many successful companies have longer pre-revenue periods). It's because the act of selling forces a market feedback loop that pre-revenue companies can avoid indefinitely.
A company that has sold something knows whether it has a product and a customer. A company that hasn't sold anything is still guessing.
Team completeness at founding
Startups with a complete founding team (typically someone who can build the product and someone who can sell it) survive at roughly 30% higher rates than solo founders and single-skill founding teams (Kauffman Foundation, 2024).
That aligns with the CB Insights "wrong team" cause (23% of failures). What researchers mean by wrong team isn't that the founders were incompetent. It's usually that the team was missing a skill the company couldn't compensate for, most often sales or operations.
Operational discipline before scale
A pattern in post-mortems that doesn't get enough attention: companies that scale headcount before achieving repeatable unit economics fail at very high rates. The temptation to hire aggressively after a funding round, before the business model is proven, is a consistent precursor to burn rate crises 18 to 24 months later.
The relationship between operational cost mismanagement and failure is analyzed in depth in the startup hiring costs research, which covers what it actually costs to build an early team.
Section 9: What founders can do with this data
Reduce execution risk, not just ideation risk
Most founder attention in the early stage goes toward product decisions. The failure data suggests operational decisions (hiring timing, cost structure, cash management) deserve comparable attention. A company with a mediocre product and tight operational discipline can survive long enough to improve. A company with a great product and sloppy operations often can't.
Know your industry's base rate
Building a restaurant with a 61% three-year failure rate is a different risk calculation than building a B2B software company with a 40% three-year failure rate. Neither is a reason not to do it, but knowing the base rate helps you make decisions about capital structure, how fast to scale, and how long to give the business before pivoting.
Delegation as a survival mechanism
One underappreciated pattern in the survival data: founders who build operational support early (including administrative and coordination support) report better focus metrics and lower burnout rates, both of which correlate with survival. The founder burnout statistics research shows that founders who delegate effectively report 35% lower burnout rates than those who don't.
Operational drag accumulates faster than most founders expect. The companies that survive long enough to find product-market fit are often the ones whose founders are most protected from that drag: the scheduling, email management, research, and coordination work that fills hours without moving the business.
A virtual assistant is one of the most cost-effective ways to remove that drag. For founders running 60-hour weeks across everything, offloading 10 to 15 hours of administrative work a week can meaningfully change the quality of decisions in the hours that remain.
Conclusion
The startup failure rate is high. About half of all new businesses don't make it to year five. For venture-backed tech startups that fail to find product-market fit, the rate is worse.
But the headline number hides the structure. Failure concentrates in specific patterns (no market need, premature scaling, cost mismanagement, team gaps) and that structure is at least partially actionable. Founders who study what kills companies and build against those patterns are doing something more concrete than hoping their company is the exception.
The data points in a consistent direction: find customers early, control your cost structure, build a complete team, and protect your ability to think clearly by not spending your highest-leverage hours on your lowest-leverage tasks.
Sources and methodology
- U.S. Bureau of Labor Statistics. Business Employment Dynamics: Survival of Private Sector Establishments. Washington, D.C.: BLS, 2024.
- CB Insights. The Top Reasons Startups Fail. CB Insights Research, 2024.
- Startup Genome. Global Startup Ecosystem Report 2024. San Francisco: Startup Genome LLC, 2024.
- Kauffman Foundation. State of Startup Activity in the United States. Kansas City: Kauffman Foundation, 2024.
- SCORE. Small Business Failure Rates 2024: Trends and Causes. SCORE Mentors, 2024.
- U.S. Small Business Administration. Small Business Facts. Washington, D.C.: SBA, 2025.
- National Restaurant Association. Restaurant Industry 2024 Outlook. Washington, D.C.: NRA, 2024.
- Crunchbase. Startup Funding and Survival Analysis, 2024. Crunchbase Pro Data, 2024.
- Fundera (NerdWallet). Startup Costs Survey and Small Business Statistics. 2024.
- Startup Genome / Crunchbase. Series A Crunch Analysis: Seed to A Conversion Rates. 2024.
- Companies House (UK). UK Company Dissolution Statistics 2024. Companies House, 2024.
- Industry Canada. Key Small Business Statistics 2024. Ottawa: Government of Canada, 2024.
- Tracxn. India Startup Ecosystem Report 2024. Tracxn Technologies, 2024.
- KfW Research. KfW Startup Monitor 2024. Frankfurt: KfW, 2024.
- Australian Bureau of Statistics. Counts of Australian Businesses, Including Entries and Exits. ABS, 2024.
- SaaS Capital. 2025 Spending Benchmarks for Private B2B SaaS Companies. SaaS Capital LLC, 2025.
Research compiled May 2026. Statistics reflect the most current available data from cited sources. Some figures represent composite averages from multiple survey populations.