Pre-Seed · April 2026
Autonomous Tunnel Monitoring
The same flight every month. AI that gets smarter with each pass.
Defect growth tracked to the millimeter.
Operators subscribe to infrastructure intelligence, not to a drone.
The Problem
"Is this crack getting worse?"
"Two inspectors may inspect the same bridge on different occasions and have slightly differing opinions on a defect, sometimes resulting in what appears to be a miraculous self-healing."
Mind Foundry, industry analysis
Every inspection is different. Different team, different path, different conditions.
Human inspectors are subjective and fatigued. Early-stage defects missed until emergency.
No temporal baseline. No comparison. No growth curve. Nothing to audit.
Emergency repairs cost 10-100x what early detection would have.
The Scale
"Critical infrastructure is
perhaps not as safe
as the public thinks."
UK Institution of Civil Engineers
State of the Nation · 2025
Germany's infrastructure renewal commitment through 2035. Europe's infrastructure is past design life.
cost per day · manual inspection
emergency vs. preventive
Our Solution
TRL 7
Same trajectory every flight. No pilot. Beyond radio range.
AI defect detection. Auto-catch problems.
Temporal comparison. "This crack grew 2.3mm since Feb."
Co-developed with
MLOps platform for the edge-AI pipeline.
SNCF subsidiary. Analytics & digital twins.
Demo
Temelín nuclear power plant · Autonomous · GNSS-denied · No pilot · Fully repeatable
The Product
Monthly flight · same format · comparable across years.
Severity-ranked · Temporal · Same-day
Why Now
Metro and rail operators designated as critical entities by July 2026. Risk assessments due within 9 months.
Germany has committed €500B to transport infrastructure renewal through 2035.
The Netherlands reports a €35B gap between current maintenance budgets and what aging tunnels and bridges actually need.
3D LiDAR: <€500 (10× cheaper than 2020).
Edge AI: 7.5× faster inference than previous-gen accelerators.
The hardware stack to do this autonomously is finally shippable.
Market
Czech Republic + France
~400 km metro tunnels · ~700 km rail tunnels
€826K-2.3M SAM in beachhead
SOM ~7-13% (€55K-290K realistic capture Y1-2)
DACH, Nordics
SAM: €140–280M
19.5% global CAGR
Y5 target: 50 DaaS clients, €5M ARR
Business Model
Continuous autonomous monitoring on a monthly subscription. Zero client CAPEX. Priced per tunnel installation.
Recurring, predictable, high retention.
Single engagements. Early revenue channel and lead generation funnel into DaaS.
Low commitment for first-time customers.
Software-like gross margins on a hardware-backed service.
One drone amortizes over many flights; the data compounds in value each time it flies.
Competition
Collection = repeatable path · Detection = AI defect ID · Tracking = temporal comparison.
* Replays pilot-flown path, no accuracy specs, pilot required. ** LiDAR geometry change detection (±5mm), no defect classification. † AiDIT AI inspection for ship hulls, not tunnels.
Traction
status · pre-revenue · first paid pilot targeted Q4 2026
MLOps platform powering the edge-AI pipeline.
SNCF subsidiary. Analytics & digital twins.
Team
CEO & CTO · Full-time
PhD @ CTU Prague. 10 years autonomous navigation, GNSS-denied flight systems, 500+ flight hours. DARPA SubT, Dronument. Core drone tech.
CPO · Full-time with funding
PhD @ Inria France. 2 years robotics sales and UX/UI dev @ Fly4Future. Leading France: Inria, accelerators, clients. Core sales skills.
+ 4 key employees joining with funding: Marketing & field engineer, AI researcher, HW engineer, robotics R&D engineer.
The regulation is here.
The infrastructure is aging.
The window will not stay open.
pavel@eclypsrobotics.com