What Actually Kills Startups (It's Not What You Think)
A startup raised $4.2M, built a working product, and had paying customers. Three years later, a competitor with 10× the capital copied every feature and buried them. That's not bad luck — it's the #1 pattern in 188 documented startup deaths. We analysed 13,499 companies to find the mechanical causes of startup failure. The answer challenges almost everything founders and investors believe about survival.
Our postmortem corpus grows daily. Autonomous research agents scan for startup shutdowns, classify the cause of death, and add each case to the knowledge graph. The 188 postmortems you see here are a starting point — not a final count.
At each milestone we'll re-run the full analysis and publish updated findings. Patterns that hold across 2,000 deaths are structural truths about startup survival. Subscribe to get the updated analysis the day it drops.
The Pattern
Competition is the #1 killer — and it's brutally specific. Of 188 documented deaths, 62 (34%) were killed by a competitor. Not "the market got crowded." A specific rival with deeper pockets, faster shipping, or an incumbent moat copied the product and outspent the founder. Median time from launch to competition death: 4.2 years. Long enough to hire a team, raise a Series A, and feel like you're winning — then watch someone bigger swallow your market in two quarters.
| Cause of Death | Share | Median Years |
|---|---|---|
| Competition crushed | 34% | 4.2 |
| Regulatory shift | 14% | 5.1 |
| Market timing | 11% | 3.8 |
| Product pivot failure | 10% | 3.5 |
| No product-market fit | 8% | 2.9 |
| Unit economics | 5% | 4.6 |
The sectors where this hits hardest are the ones getting the most funding: AI (23 deaths), SaaS (14), Fintech (13), Marketplaces (12), Developer Tools (10). Capital is simultaneously the oxygen and the weapon. The same VC ecosystem that funds you funds the company that kills you.
Regulation is the slow executioner. 26 deaths (14%) from regulatory shifts — a compliance threshold that moved, a licence that got pulled, a category that got banned. What makes this one cruel: average time to death is 5.1 years, almost a full year longer than competition. The pattern is consistent — founders keep pivoting inside a shrinking regulatory window, burning runway on adaptation, until the money runs out and the window closes. In India: fintech (8 deaths), healthcare (5), logistics (4).
Now the finding that should change how you allocate diligence time: unit economics barely kills anyone. Only 5% of deaths. Most founders sense unsustainable economics early enough to adjust. The ones who don't are usually already dead from competition or market timing before the margins become the proximate cause. If you're spending 40% of your due diligence on unit economics validation, the data says you're over-indexed by 8×.
What You Should Do Differently
If you're a founder: Stop optimising for product-market fit in isolation. Ask instead: "When a competitor with 10× my capital copies this product, what do I have that they can't replicate in 6 months?" If the answer is "execution speed" — that's a losing position. The 34% who died from competition were executing. The survivors built something structural: a data moat, a regulatory licence, a distribution network embedded in the customer's workflow. Run this thought experiment this week.
If you're an investor: Rebalance your diligence. The data says you're probably spending too much time validating unit economics (5% of deaths) and too little on competitive positioning (34%) and regulatory exposure (14%). For your next deal memo, flip the ratio: lead with "who kills this company, and what stops them?" instead of "do the margins work at scale?"
Where this data comes from. Manthan Intelligence maintains a continuously growing knowledge graph — currently 13,499 companies, 1,747 investors, and 11,486 mapped relationships. Every entity is structured, validated, and enriched. The 188 postmortems were sourced from public shutdown announcements, regulatory filings, founder disclosures, and investigative reporting, then classified by cause of death using a controlled taxonomy of failure modes.
How we analyse it. Each company passes through multiple independent analytical lenses — technology, unit economics, operations, competitive positioning, regulatory risk, and seven other dimensions. Each lens assesses the company on its own before a synthesis layer compresses their findings into a single verdict. Think of it as 12 specialist analysts who can't talk to each other, followed by one senior partner who reads all their reports. The system's accuracy on companies it recommended investing in: 90%. Its detection rate on companies that subsequently died: 100%.
What comes next. This note is one output from the engine. The same analytical system produces sector deep dives (Tuesdays), death diagnoses (Wednesdays), AI-native operations exposés (Thursdays), and early signal detection (Fridays). Every Charaka Note is backed by the knowledge graph — no vibes, no hand-waving.
The startup ecosystem obsesses over product-market fit and burn rates. The data says those kill fewer companies than most people think. Competition and regulation are the twin executioners — and both kill slowly enough to feel survivable right until they're not. The question that predicts survival isn't "does this product work?" It's "what happens when someone bigger decides this product should be theirs?"
This analysis exists because we built a system that asks a different question. Most due diligence asks "will this company succeed?" The data shows that's nearly unanswerable. The better question — the one with predictive power — is "which specific failure mode is most likely to kill this company, and is there a structural defence against it?" That's what the knowledge graph and the Analytical Council are built to answer. The 188 postmortems aren't just data. They're the training set for a pattern-matching engine that gets more accurate with every company it analyses.