We’re building the bond for drones and AI that work on critical infrastructure. Before a bonded machine is switched on, its insurer can state what it’s physically able to do. After any incident, they can prove exactly what it did and whose decision caused it.
Drones and AI are starting to inspect power lines, pipelines, bridges and substations. To insure one, an insurer needs two answers: what can this machine do, and what did it do? Today the answers are logs, dashboards and the operator’s word. None of them settle a claim.
Trust our logs.
Here’s the proof.
Not telemetry, not a dashboard, not a compliance report. A bond is what an insurer buys, and every layer we’re building is a part of it.
Today a machine’s limits live in settings that a bug, an update or a person can quietly change.
We’re building a certified envelope of speed, force and airspace that’s sealed into the machine and registered publicly.
A hack, a bad model or a wrong command can still reach the motors.
We’re building a gate that sits between any autonomy and any motor, so an out-of-bounds command physically can’t get through.
A machine can drift for months between inspections without anyone knowing.
We’re building checks that run on schedule and without warning, so the machine proves its health for the whole policy period.
Most records live on the machine itself, with gaps nobody can see.
We’re building a signed record that’s copied off the machine as it happens and shows any gap or tampering.
After an incident, the operator, the manufacturer and the software vendor all blame each other.
We’re building a record of who was responsible at the instant of each decision, so that question has one answer.
Experts spend months reconstructing what happened before anyone gets paid.
We’re building the rules that turn the record into a liability split, a settled claim and next year’s premium.
Every regulator asks for its own paperwork, assembled after the fact.
We’re building evidence packages generated from the same record, in the formats aviation, automotive and AI regulators already ask for.
Today’s security checks answer one question: is this the software we expected? That’s necessary, and nowhere near enough to underwrite a machine that can hurt someone. We build on those checks and add what an insurer needs: limits that can be priced, a record that settles claims and one answer to who was responsible.
Price the risk before the policy starts, and settle claims from the record.
Lower premiums for machines that prove their limits every day.
One gate between any autonomy and any motor, from any supplier.
Evidence in the formats you already ask for, from the same record.
We’re developing an open standard for how insurers and regulators read a bonded machine, with a conformance program and a mark. A machine that carries the mark can be underwritten the same way, whoever built it.
Every world-class hub started somewhere. Silicon Valley was orchards once. It took a few people who decided not to wait for permission.
Why not us.
Why not here.
Why not now.
Power grids / Pipelines / Bridges & rail / Telecom towers / Water / Anywhere a machine works where it could hurt someone
Machines that move are about to need insurance the way cars and planes do. If you want to build what makes that possible, we should talk.
People who build the gate between the autonomy and the motors.
Introduce yourself →Engineers who know what a drone should do when something goes wrong.
Introduce yourself →Builders of records that survive lost links, lost machines and bad actors.
Introduce yourself →People who turn a machine’s record into a price.
Introduce yourself →Specialists who know what regulators ask for, and why.
Introduce yourself →Partnerships, product and operations people who want to bring the bond to market.
Introduce yourself →Insurers · Operators · Manufacturers
For insurers, reinsurers, infrastructure operators and machine makers who want autonomy they can underwrite.