Startups Françaises Disparues : Sigfox, Ÿnsect, Luko, Take Eat Easy
Analyse des faillites de startups françaises : Sigfox (€290M IoT), Ÿnsect (€625M AgTech), Luko (insurtech), Take Eat Easy (post-mortem culte). Paris et Toulouse, cas réels et leçons concrètes.
- 22+Documented cases
- $6BCapital lost
- 80%Fail rate
Startup Ecosystem Overview
L'écosystème startup français est centré à Paris (Station F, Sentier) avec des pôles secondaires à Toulouse (deeptech, IoT), Lyon, et Sophia Antipolis. La French Tech a généré 30+ licornes (Mistral, Doctolib, Qonto). Bpifrance fournit un soutien public massif, mais le capital late-stage privé reste plus rare qu'au UK. Forces : ingénierie de pointe, recherche publique forte, AgTech/DeepTech reconnues. Faiblesses : exits limitées (peu d'IPO Euronext), procédures de redressement lourdes (Sigfox, Ÿnsect, Luko), et un marché domestique fragmenté EU.
Failures by Industry
Failure Reasons: France vs Global Average
Share of failures by root cause (%) — local pattern vs the 1004-startup global baseline.
- France
- Global average
Over-indexed
Unit Economics
France startups fail from this +4.3 pts more often than the global average (22.7% vs 18.4%).
Under-indexed
No Market Need / PMF
France startups fail from this -6.9 pts less often than the global average (18.2% vs 25.1%).
Methodology: Each startup's freeform failure reason is mapped to one of 9 canonical buckets (no-PMF, cash, unit economics, competition, fraud/governance, regulation, operations, team, pivot). Top 7 buckets by combined signal shown.
Cultural & Regulatory Factors
Cycles de Redressement Judiciaire
La France utilise massivement la procédure de redressement judiciaire (équivalent du Chapter 11 US). Sigfox, Ÿnsect, Luko, Habx et Cityscoot sont tous passés par cette procédure — souvent suivie d'une cession à prix bradé.
AgTech & DeepTech à Forte Intensité Capitalistique
Ÿnsect (€625M), Sigfox (€310M) montrent que les bets deeptech français consomment énormément de capital. Sans ancres industrielles ou IPO réussie, beaucoup atteignent l'insolvabilité avant la maturité commerciale.
Capital Late-Stage Limité
Les rounds Series C+ supérieurs à €100M restent rares en France hors top tier. Beaucoup de startups françaises vendent à des stratégiques internationaux (TUI achète Musement, Allianz Direct achète Luko) plutôt que de scaler indépendamment.
22 documented failures — the most-cited names from this market.
Capital raised before shutdown — France
USD millions raised by each documented failure.
Failed Startups (22)
Cazoo France (Market Exit)
Pan-European Retreat · UK used-car unicorn Cazoo entered France in 2021 with massive marketing then wit…
$2B
2018–2023
Sorare (Markdown)
NFT Crash & Volume Collapse · Paris fantasy-sports-NFT unicorn Sorare raised $680M from SoftBank at $4.3B in 2…
$735M
2018–2023
ManoMano (Massive Down Round)
Post-COVID Demand & Profitability Push · Paris DIY-marketplace unicorn ManoMano raised $725M then conducted multiple roun…
$725M
2013–2023
Ÿnsect
Capital-Intensive AgTech, Filed Safeguard · France's most-funded AgTech, insect-protein producer Ÿnsect, raised $625M then e…
$625M
2011–2024
Ledger (Crypto-Winter Down Round)
Crypto Winter & Markdown · Paris crypto-hardware unicorn Ledger reached $1.5B valuation in 2021, then condu…
$575M
2014–2024
Believe (Public Markdown)
Public-Market Failure & Take-Private · Paris-listed digital-music distributor Believe IPOed at €19.50 in 2021, then was…
$338M
2005–2024
Sigfox
Proprietary protocol, strategic architecture failure · Proprietary protocols are detrimental in infrastructure markets demanding intero…
$300M
2009–2022
Meero (Founder-Driven Crisis)
Mass Layoffs & Restructuring · Paris on-demand-photography unicorn Meero raised $293M at $1B valuation in 2019,…
$293M
2016–2023
Ornikar (Down Round & Layoffs)
Failed International Expansion & Layoffs · Paris driving-school unicorn Ornikar reached $750M valuation in 2021, then condu…
$170M
2013–2023
Navya
Broken unit economics, hardware trap · Hardware-as-a-Service models need positive unit economics within 18 months of de…
$120.0M
2014–2023
Blade (Shadow)
Unsustainable unit economics, high infrastructure costs · Cloud gaming requires first-party content or alternative monetization strategies…
$110.0M
2015–2021
BeReal
Feature Copied by Instagram · BeReal's "authentic photo" feature was copied by Instagram within months. Social…
$90M
2019–2024
EasyMile
Pilot purgatory, unsustainable burn, no scale · Hardware startups must avoid pilot purgatory and achieve sustainable unit econom…
$90M
2014–2024
Luko
Acquisition at Discount, Then Insolvency · Paris home-insurance unicorn-aspirant Luko was acquired by Allianz Direct subsid…
$73M
2018–2024
Cityscoot (Insolvency)
Mobility Unit Economics · Paris electric-scooter-sharing pioneer Cityscoot raised €60M from RATP and Allia…
$70M
2014–2024
Cubyn
Unsustainable unit economics, high capital intensity · Logistics startups with high capital intensity and low margins often struggle wi…
$70M
2014–2024
Habx (Acquired in Distress)
French Property Market Collapse · Paris PropTech Habx raised €40M to digitalize new-build property pre-sales then …
$45M
2016–2024
Withings (Nokia Acquisition Crisis)
Nokia Write-Down & Failed Integration · Paris connected-health pioneer Withings was acquired by Nokia for €170M in 2016,…
$30M
2008–2017
Take Eat Easy
Outcompeted by Deliveroo & UberEats · Belgian-French food-delivery pioneer ran out of cash before reaching scale. The …
$16M
2013–2016
Sunrise
Bad business model (hard to monetize) · Even popular apps can struggle if they operate in a market segment that is inher…
$8.2M
2012–2016
Tilkee (Acquired Distress)
Failed to Scale Internationally · Lyon B2B SaaS Tilkee raised €8M from Bpifrance then was acquired in distress in …
$8M
2013–2023
Vivalatina
Poor marketing and competition · Even with a unique product, intense competition in e-commerce requires a strong …
Unknown
2014–2019
Lessons for France Founders
- ✓Anticipez le redressement judiciaire comme un outil, pas une honte — Luko et Sigfox y ont survécu en partie
- ✓Pour la deeptech, sécurisez des ancres industrielles tôt (TotalEnergies, Air Liquide, L'Oréal) — le capital privé seul ne suffit pas
- ✓Construisez international dès le seed — le marché français seul rarement suffit pour des sorties venture
- ✓Pour l'insurtech, ne sous-estimez jamais la rentabilité — Luko et Coya prouvent que le full-stack EU brûle vite
Frequently Asked Questions
Quel est le taux d'échec des startups en France ?
Environ 80% des startups françaises échouent dans les 10 ans, légèrement au-dessus de la moyenne européenne. L'écosystème bénéficie du soutien Bpifrance et d'un fort capital amorçage, mais souffre d'un déficit de capital de croissance privé.
Quelle est la plus grande faillite de startup française ?
Par capital levé, Ÿnsect (€625M en redressement judiciaire en 2024) et Sigfox (€310M, repris par UnaBiz à prix bradé en 2022) sont les plus importantes. Luko (€80M) est le cas insurtech emblématique. Take Eat Easy reste le post-mortem le plus lu d'Europe.
Pourquoi tant de deeptech françaises échouent-elles ?
Les bets deeptech (Ÿnsect, Sigfox, Lilium côté allemand) demandent €500M-€1B+ pour atteindre la maturité commerciale. Le capital Series C+ français privé ne suit pas, et les marchés publics (Euronext) ne valorisent pas suffisamment le risque deeptech.
L'écosystème French Tech se porte-t-il bien en 2025-2026 ?
Oui pour l'IA (Mistral, H, Photoroom) et la fintech profitable (Qonto, Doctolib). Non pour les bets capital-intensifs (insurtech, AgTech industrielle, IoT) où les défaillances continuent. Le segment SaaS B2B se porte mieux que le consumer.