The Meta-Study
Why They Failed
A systematic analysis of 5,727 startups tracked by Failradar. What killed them, which sectors are deadliest, and where the rebuild opportunities hide.
1,003
Confirmed dead
3,992
Still active
732
Acquired
41
Sectors tracked
Survival Data
Failure rate by YC batch. Older cohorts have had more time to fail — the curve reveals when startups are most vulnerable.
Cause of Death
Top reasons startups fail, extracted from post-mortem analyses across the dataset.
Note: A startup can have multiple causes of death. Counts represent mentions across post-mortem analyses.
“Startups rarely die from a single cause. Most post-mortems reveal a cascade — competition compresses margins, which burns cash faster, which prevents the pivot that might have saved them.”
— Pattern across 1,003 post-mortems
Threat Ranking
All sectors ranked by mortality rate. Know the odds before you build.
| Rank | Sector | Dead | Total | Death Rate | Risk |
|---|---|---|---|---|---|
| 01 | Social & Communication | 37 | 71 | 52.1% | CRITICAL |
| 02 | E-Commerce | 36 | 78 | 46.2% | CRITICAL |
| 03 | Gaming | 13 | 29 | 44.8% | CRITICAL |
| 04 | Sales | 6 | 14 | 42.9% | CRITICAL |
| 05 | Media & Entertainment | 31 | 73 | 42.5% | CRITICAL |
| 06 | AR/VR | 8 | 20 | 40.0% | CRITICAL |
| 07 | Agriculture | 2 | 5 | 40.0% | CRITICAL |
| 08 | Transportation | 12 | 36 | 33.3% | HIGH |
| 09 | Customer Support | 2 | 6 | 33.3% | HIGH |
| 10 | Blockchain & Crypto | 14 | 43 | 32.6% | HIGH |
| 11 | Food & Beverage | 21 | 66 | 31.8% | HIGH |
| 12 | Manufacturing | 6 | 21 | 28.6% | HIGH |
| 13 | Productivity | 32 | 114 | 28.1% | HIGH |
| 14 | Cybersecurity | 12 | 43 | 27.9% | HIGH |
| 15 | Travel | 3 | 11 | 27.3% | HIGH |
| 16 | Marketplace | 60 | 221 | 27.2% | HIGH |
| 17 | Marketing | 10 | 37 | 27.0% | HIGH |
| 18 | Legal Tech | 5 | 20 | 25.0% | HIGH |
| 19 | Logistics | 17 | 70 | 24.3% | MODERATE |
| 20 | Fitness & Wellness | 2 | 9 | 22.2% | MODERATE |
| 21 | Cloud Infrastructure | 2 | 9 | 22.2% | MODERATE |
| 22 | Real Estate | 11 | 50 | 22.0% | MODERATE |
| 23 | Other | 130 | 604 | 21.5% | MODERATE |
| 24 | Hardware | 36 | 168 | 21.4% | MODERATE |
| 25 | Healthcare | 51 | 265 | 19.3% | MODERATE |
| 26 | Robotics | 10 | 52 | 19.2% | MODERATE |
| 27 | Data Analytics | 14 | 75 | 18.7% | MODERATE |
| 28 | Education | 33 | 182 | 18.1% | MODERATE |
| 29 | Climate Tech | 9 | 53 | 17.0% | MODERATE |
| 30 | Nonprofit | 2 | 12 | 16.7% | MODERATE |
| 31 | Developer Tools | 65 | 422 | 15.4% | MODERATE |
| 32 | Consumer | 8 | 60 | 13.3% | LOW |
| 33 | Biotech | 18 | 140 | 12.9% | LOW |
| 34 | Fintech | 87 | 692 | 12.6% | LOW |
| 35 | SaaS | 73 | 615 | 11.9% | LOW |
| 36 | Energy | 4 | 34 | 11.8% | LOW |
| 37 | HR & Recruiting | 3 | 29 | 10.3% | LOW |
| 38 | Government | 2 | 20 | 10.0% | LOW |
| 39 | Artificial Intelligence | 109 | 1158 | 9.4% | LOW |
| 40 | Aerospace & Defense | 2 | 22 | 9.1% | LOW |
| 41 | DevOps | 5 | 78 | 6.4% | LOW |
Note: Death rate = inactive companies / total companies in category. Older companies have had more time to fail. Data from 41 canonical categories across 5,727 YC startups + 1,006 supplemental entries.
Sector Autopsy
Deep dives into the deadliest sectors. Named failure patterns, primary threats, and the insight that explains why.
Social & Communication
37 dead / 71 tracked
Failure Pattern
The Attention Trap
Primary Threat
Network effects favor incumbents
“Second best is often worth zero in the attention economy. Users consolidate into one platform per need.”
E-Commerce
36 dead / 78 tracked
Failure Pattern
The Logistics Cliff
Primary Threat
Operational scalability at thin margins
“The physical world has friction that code does not. Every warehouse, truck, and return label is a cost your pure-software competitor never pays.”
Gaming
13 dead / 29 tracked
Failure Pattern
The Hit Machine
Primary Threat
Power-law distribution of success
“One viral title funds a studio. Zero viral titles kills it. Gaming is a hits business disguised as a technology business.”
Media & Entertainment
31 dead / 73 tracked
Failure Pattern
The Content Treadmill
Primary Threat
Constant production required
“Content is a treadmill. Stop running and the audience disappears. The unit economics of creation rarely survive the attention span of consumption.”
AR/VR
8 dead / 20 tracked
Failure Pattern
The Timing Paradox
Primary Threat
Technology ahead of market demand
“You can be right about the future and still be too early to survive it. The graveyard is full of companies that were correct about where the world was heading.”
Hardware
36 dead / 168 tracked
Failure Pattern
The CAPEX Cliff
Primary Threat
High capital burns before PMF
“Software startups iterate weekly. Hardware startups iterate quarterly. By the time you discover the market wants something different, your runway is ash.”
Revival Vectors
Three emerging patterns where failed startup models can be rebuilt with modern technology stacks.
“Look for Series B failures from 2015-2018 that died of unit economics. Rebuild with serverless and AI stacks to flip margins from -40% to +20%.”
Human-in-the-Loop Replacement
Cause of Death
Marketplaces and service businesses died on operational scalability. Human coordination costs grew linearly while revenue needed to grow exponentially.
53% of marketplace failures cite operational bottlenecks
Rebuild Vector
Replace human coordination with agentic AI. LLMs handle matching, negotiation, and quality control that previously required hiring.
Serverless Economics Pivot
Cause of Death
Monolithic infrastructure startups failed on unit economics. Fixed server costs meant losing money on every customer until hitting massive scale.
62% of infrastructure failures cite unit economics
Rebuild Vector
Scale-to-zero architecture eliminates the fixed-cost death spiral. You pay per-request, not per-server. Margins flip from -40% to +20%.
Hyper-Personalized Vertical
Cause of Death
EdTech and content startups bottlenecked on static content creation. One-size-fits-all products couldn't retain users who wanted personalized experiences.
48% of education failures cite engagement and retention problems
Rebuild Vector
Generative AI enables adaptive, personalized content at marginal cost near zero. Every user gets a unique experience without a content team.
Field Notes
12 rules distilled from the data. Not theory — patterns observed across thousands of outcomes.
Validate demand through founder interviews before writing code. Over 36% of failures cite no market need.
Build lean MVPs. Minimize overhead until product-market fit is confirmed with paying users, not just signups.
Stay hyper-focused on a single problem for a clear customer segment. Pivot sprawl is a top-5 killer.
Nail unit economics early. Ensure LTV exceeds CAC before scaling spend. Negative margins don't flip at scale.
Develop go-to-market alongside the product. Distribution is not something you bolt on after launch.
Build a complementary founding team covering product, sales, and finance. Solo technical founders underperform.
Track runway daily. Cut costs at the first warning sign, not after the bank account hits zero.
Pivot based on data and customer behavior, not gut feelings or board pressure.
Create clear differentiation versus existing alternatives. "Better UX" is not a moat.
Test market timing as critically as product-market fit. Right idea, wrong year is still a dead company.
Incorporate compliance from day one in regulated sectors. Retrofitting is 10x more expensive.
Use AI to solve operational bottlenecks, not to replace your core value proposition.
Don't just read the data
Run your idea against the dead.
The AI validator cross-references your startup concept against 1,003 real failure outcomes.