COPPER THEFT · CRITICAL INFRASTRUCTURE
Copper theft is a behavior problem, not a metal problem. The dwell, the scope, the stage-and-return — every theft has a recon pattern hours or nights before the wire comes out. Edge Orbital builds the pre-incident detection layer that catches the recon, not just the loss.
The Department of Energy estimates copper theft costs American businesses on the order of $1 billion per year. The number that matters more — the one that puts a co-op operations manager in front of an emergency capital request — is the damage multiplier: a $100 piece of wire pulled from a substation can produce a $1M repair bill when the resulting fire takes a transformer with it. Industry estimates put the multiplier at 10–28× the value of the stolen copper.
And these are only the named losses. Most copper-theft events never reach a press release. They show up as line items in a co-op’s monthly operating report, as customer outage tickets, as an avoided-replacement-cost the operations manager learns to absorb. The defenders are operating on insurance-renewal cycles and capital plans; the thieves are operating on a 24-hour commodity-price feed.
The thing every long-tenured lineman, dispatcher, and site security manager knows from field experience — and that the post-incident incident reports almost always confirm — is that copper theft is rarely opportunistic. The economics demand otherwise. To extract enough wire to be worth the risk, a crew needs to know where to enter, where the wire runs, what the surveillance posture is, and how long they have. They case the site. They walk the fence. They drive past at 11pm, again at 2am, again the following Tuesday. They stage a vehicle. They come back.
Every one of those steps is observable. The dwell is observable. The vehicle stage-and-return pattern is observable. The irregular movement against the fence line at hours when no one else is moving against the fence line is observable. The problem isn’t that the recon is invisible — it’s that nobody is watching at the layer where it shows up. Reactive monitoring sees the cut. Pre-incident detection sees the casing.
Why isn’t this being done already? The same reason the rest of the unguarded grid is unguarded: the procurement-consulting-fee economics of the incumbent monitoring vendors don’t reach below the Fortune-500 procurement cycle. The math that works for a co-op operations manager — a low-five-figure annual line item that prevents a single million-dollar loss — doesn’t fit a six-figure enterprise sales motion. The vendors that have the right capability charge for the wrong customer. The vendors that fit the customer don’t have the capability.
This is the buyer tier the incumbents can’t profitably serve, and it is — by line-mile, by feeder count, and by total exposed copper — the majority of the U.S. distribution grid. Rural electric cooperatives. Municipal utilities. Independent power producers. Battery energy storage operators. Solar farms. Small refineries, midstream terminals, regional water utilities, agricultural processing sites. The wire is everywhere; the protection is in a few enterprise-procurement strongholds.
Pre-incident detection on the recon patterns that precede every copper-theft event. Three architectural properties anchor the system:
1 · Behavioral signatures, not bounding boxes. The detection layer is built to recognize casing, scoping, dwelling, returning — the activity sequence that precedes a cut, not the cut itself. The model separates the dog walker from the surveyor; it separates the line-crew night call from the third visit by the same vehicle in 48 hours. Sensor fusion (visual + radio + spatial-temporal) makes the discriminator robust to the conditions that defeat a perimeter camera alone — fog, low light, weather, the back of the site that’s outside the camera’s pan range.
2 · Each site stands alone. The node runs full perception locally. No central-cloud dependency, no SCADA integration cost, no broadband-uplink assumption. A node continues to do its job when backhaul is contested, when a regional outage takes the internet down, when the conditions that make the site valuable are exactly the conditions when monitoring matters most.
3 · Operator-budget pricing, ops-manager-routed alerts. The unit economics target the line item that closes inside a co-op operating budget. Alerts route to the ops manager and the line crew — text, dispatch, app — not to a SOC contract or a 24/7 monitoring desk. The decision to roll a truck stays with the people who know the site, the conditions, and the crew. Capability before claim: we are pre-pilot and explicit about it.
The same behavioral-detection layer that catches the copper-theft recon catches the multi-night reconnaissance that precedes a coordinated physical attack on transformers and breakers. The pattern shape is the same. The 2022 Moore County incident in North Carolina — which knocked 45,000 customers offline for days — was preceded by a reconnaissance pattern of exactly this kind. Dual-use, by design in the American Dynamism sense: civilian critical-infrastructure protection and adjacent threat-recognition served by the same underlying perception stack, without re-engineering.
If the pattern fits, the page below has the contact path. We are pre-pilot and explicit about that — what we are building toward is the deployment that proves the architecture against the customer profile.
Edge Orbital is building pre-incident detection for copper theft — designed for the operating-budget tier the incumbent monitoring vendors don’t reach, supporting the national interest by closing the loop between casing and loss.
Read the full critical-infrastructure thesis or talk to us about a pilot.