What percentage of startups fail?

About 90% of startups fail eventually, with the steepest mortality in years 2–5. Roughly 20% fail within year 1, ~50% by year 5, and ~70% by year 10. The exact number varies by definition (acquihire vs shutdown vs zombie) but the order of magnitude is stable across studies — CB Insights, BLS, and Startup Genome all land in this range.

What is the #1 reason startups fail?

Building something nobody wants — also called 'no product-market fit' or 'no market need.' Every major post-mortem study (CB Insights' analysis of 110+ startup post-mortems is the most cited) puts this at the top, ahead of running out of money, team problems, or competition. The other reasons are usually downstream consequences of this one.

Do most startups fail because of money?

Running out of cash is the proximate cause of most shutdowns, but it is almost always a symptom. Startups run out of cash because their customer acquisition cost was higher than their lifetime value, because they hired against a revenue plan that never materialised, or because they were building the wrong product. Healthy demand fixes most cash problems; cash without demand fixes none.

How do AI startups fail differently?

AI startups fail in two distinctive ways. First, they confuse a wrapper around an LLM with a defensible product — model improvements collapse their value prop overnight. Second, they over-invest in model performance and under-invest in distribution, evaluation infrastructure, and the unsexy workflow integration that makes the product sticky. The 2024–25 graveyard is full of polished demos that never reached a second renewal.

Can a startup recover from failure signs?

Yes, if the founder is willing to do three things most refuse: cut the team and runway to extend, kill features and revenue streams that do not have customer pull, and re-validate from scratch with current customers — not the ones who churned. Most founders refuse one or more of these for emotional or sunk-cost reasons. The ones who do all three often turn around within 6 months.

Field manual

Notes from the launch room.

Essays, breakdowns, and field reports. Designed for operators who want clarity, not content.

AllLaunch StrategyBrand ArchitectureAI & ExecutionFounder PsychologyCase StudiesStrata NotesMarket Autopsies

Launch StrategyPOST-001

How to validate a startup idea without wasting six months.

Idea validation is binary: strangers paying, or not. The playbook, the five real signals, and the build-trap that kills 60% of first-time founders.

Live now→

Market AutopsiesPOST-002

Why startups fail: the 7 real reasons.

Startups don't fail because of bad ideas. They fail for seven boring, fixable reasons most founders refuse to look at.

Live now→

Playbook · JurisdictionPOST-003

Wyoming LLC for founders — the 2026 playbook.

Why founders skip Delaware and Estonia for Wyoming. Filing, banking, taxes, and the 5472 trap that kills 1 in 5 setups.

Live now→

Playbook · PortugalPOST-004

How to start a business in Portugal as a foreigner.

Visa routes (D2, StartUp, HQA), company formation, post-NHR tax reality, and the honest Lisbon vs Porto trade-off.

Live now→

Playbook · ThailandPOST-005

How to start a business in Thailand as a foreigner.

BOI vs Thai Limited, LTR Visa, Bangkok vs Koh Samui, and the foreign-ownership trap most guides skip.

Live now→

More field notes shipping regularly. Bring a real problem to the Launch Assistant — that's where the sharpest thinking lands.