Layered defense works and single products do not. Cable theft and vandalism are different problems, so treat them separately. For theft, make the cable slow to cut and the connector hard to walk away with. For vandalism, protect the parts people smash or pry at: use impact-rated screens and enclosures, keep connectors holstered and covered when idle, and lay out the site with good lighting and clear sightlines so nobody can work on a charger unseen. Then detect damage within seconds, route the alarm to somebody who will actually act, and restore the port the same day so a broken charger does not stay broken for a month. If you can only fund one layer first, fund delay on DC sites and detection on any site where a dead port costs more per day than the cable does.
Key takeaways
A stolen cable has two buyers, and they pay for different things. One wants the copper inside. The other wants a complete assembly with a molded connector on at least one end, which can be resold to a recycler or a second-hand buyer as a part. The copper in a five-metre AC cable is a small fraction of what the finished assembly sells for, which is why the trade does not slow down when scrap prices go flat. It also explains where the cut lands: taken at the pedestal, the cable keeps a usable connector.
DC fast charging cables carry far heavier conductors, so each cut is worth more. They are also thicker, stiffer, slower to get through, and on liquid-cooled designs there is a coolant hose inside as well. AC cables are the opposite: thinner walls, fast to cut, easy to coil and carry. DC is a value target, AC is a speed target. Protection priorities follow from that, not from a product catalogue.
When a site is hit twice, the second incident usually uses the same access route and the same weak point. Replacing the cable in exactly the same way, on the same pedestal, under the same lighting, is an invitation.
Cable harvested while the charger is idle, a session interrupted or a connector taken while a driver is charging, and plain vandalism such as a smashed screen or a bent connector each need a different countermeasure. A connector lock solves the second. A sleeve solves the first. Neither touches the third, which is a layout and materials problem.
A site that looks maintained, lit and alarmed is a worse target than an identical site that does not. Litter, a dark screen, a taped holster and standing water at the base all say the same thing to someone sizing up a target: nobody is watching. Repair the small things fast, because they are read as a signal about how the whole site is run.

Price the incident before you price the measure. Most operators undercount it, because the invoice only covers the part.
Two costs never reach an invoice: drivers who write the site off after finding it dead, and a site host or municipality that starts asking whether the operator can keep equipment working at all.
Annualize rather than reasoning per incident. Incidents per year times loss per incident, set against the annual cost of the measure plus the downtime it prevents. A cheap measure at a site hit four times a year beats an expensive one at a site hit once, and that comparison is invisible if you only look at one event.
Take a four-port public site and insert your own numbers: your revenue per port-hour, your measured utilization, a downtime assumption from half a day to three days. Run it three times with different downtime values. The arithmetic is what matters, and it transfers to any site you operate.
Fleets should cost a dead port as a schedule problem instead. Remaining ports absorb the displaced sessions, dwell stretches, and vehicles that miss their window end up charging at a worse time of day or at retail rates elsewhere. Repeat-attack sites break linear thinking entirely: the site hit repeatedly is the site most likely to lose its host contract, and that loss dwarfs the cable.
Give each layer a job and the spending decisions get easier.
Deterrence works on someone who has not done this before and does nothing against someone who has. Delay is the only layer that changes the result of a serious attempt. Detection does not stop the first incident, but it is what turns a month-long outage into a same-day swap.
Camera value is prosecution and an insurance claim. That value collapses when nobody watches the feed at 03:00 and the footage is unusable in the dark. Point the lens at the holster and the aisle rather than the horizon, and check at night whether a face or a plate is actually readable.
Light removes the cover a thief needs, and it is usually cheaper per unit of risk reduction than another camera. Light the holster and the walking route, not the whole lot.
A sign stating that cables are removed overnight is verifiable, and verifiable is what makes it credible. A generic CCTV sticker says nothing a thief has not seen a hundred times.
These pay only where a response exists. Ask what the response time actually is before committing to a monthly fee. A monitored alarm routed to a queue with a two-hour response is a log entry, not a control.
Delay first where the cable is expensive and downtime is costly. Detection second where daily port revenue is high. Deterrence throughout, because it is the cheapest layer on the list.
Two vendor positions worth knowing. CatStrap argues that cameras, steel cabinets and patrols alone do not stop repeat thefts, which matches the ranking above. ChargePoint describes its Protect feature as real-time detection rather than a passive deterrent, and places the alarm at the pedestal as well as on the owner phone.
Cut resistant does not mean uncuttable. It means the cable outlasts the tools and the patience of whoever is attacking it, so define the threat before you shop: bolt cutters and a minute, or a powered saw and five.
AC and DC economics diverge here. AC assemblies are thinner, cheaper and often field replaceable, so a sleeve is a judgement call. DC cables are heavy, expensive and rarely swapped by the operator, and on liquid-cooled designs the assembly is a sealed part, which is why protection pays back much faster there. The comparison of liquid-cooled and air-cooled DC charging cables explains how much assembly sits behind the jacket on each type.

Most of the loss is decided at the drawing stage, months before anyone buys a lock.
End the session, open the contactor and get the live conductor out of the puddle before anyone starts investigating. A unit that retries a session after a control pilot fault is a hazard, not an inconvenience. If a member of the public reaches a severed end that is still energized, the failure has moved from property damage to a serious safety incident.
The control pilot and proximity pilot circuits carry low voltage, so severing an AC cable looks like a lost pilot, a lost proximity, or both at once. DC connectors have their own low-voltage pins and the same logic applies. An independent cut-detect loop or an earth continuity check gives a cleaner signal with fewer false positives than inferring from the pilot alone.
Protection layers described as over-current, under-voltage and similar are electrical protections, not tamper detection. That distinction matters when you are comparing platforms: evaisun’s DC EV Charger, for example, lists 8-layer protections covering over-current, under-voltage and related faults, which protects the charger and the vehicle, not the cable. Ask specifically which input the charger uses to conclude that a cable is severed, and what it does next.
A normal unplug opens the proximity pin in a known sequence. A cut takes the pilot and the earth with it. Debounce windows, startup self-tests and a clean distinction between a worn connector and a severed cable are what keep the alarm from being muted by week three, which is the usual way these projects die.
The audible and visual alarm at the pedestal acts at 02:00 and can push a would-be thief away mid-attack, which is the approach ChargePoint takes with Protect, pairing on-site alarms with instant SMS and email alerts to the station owner. The message gets a technician dispatched. Neither replaces the other.
Status notifications entering a faulted state, vendor-specific error codes and diagnostics uploads. The mapping between charger fault codes and the alarm rules in your charge point management system is where alerting projects quietly fail, so test it with a real cut on a bench unit before the site depends on it.
The event log, the last session record and the camera clock are the evidence behind a police report and an insurance claim. Export them with timestamps and store them off the unit.
A cellular dead spot, a backhaul outage, or a charger whose alarm exists only in the cloud. A severed cable on an offline unit reports nothing at all, which is why the backup path deserves the same questions as the primary one.

Site type decides which layer carries the weight, because it decides who is present and when. The table above condenses the five common cases. Three points are worth spelling out, because they are the ones operators get wrong.
Single pedestal, no window overlooking it, no staff, all night. That combination makes socketed the better default and makes detection close to mandatory, because nobody will notice a dead port until a driver complains. If your charge point management system alarms go to an inbox nobody reads before 09:00, the detection layer is decorative.
A fence is the strongest single measure available anywhere, but the overnight risk window and the charging window are the same hours, so the equipment is exposed precisely when the yard is empty. Route the alarm to the depot manager phone rather than a shared network inbox, and expect the fleet link to be the reason nobody acts. If you are still mapping the charging schedule, what an EV fleet is and how it is charged covers the duty-cycle basics that make an assembly swap window realistic.
Where there is no realistic response within hours, spend on delay and on visible upkeep, and treat detection as a monitoring cost you may not be able to convert into action. Buying an expensive alerting stack for a site nobody can reach is the most common misallocation in this category.
Workplace and multifamily sites sit at the opposite end. Access control does more than any hardware bolted to the charger, and at multifamily sites the more frequent problem is petty unplugging between residents rather than outright theft, which puts connector locks ahead of sleeves.
Split every requirement by where it can actually be implemented. Cable construction, tamper sensing, fail-safe logic, holster interlock, OCPP event mapping and field-replaceable cable assemblies are factory decisions. Sleeves, locks, cameras, lighting and backhaul are site decisions that can follow later.
Anything needing the enclosure opened or the firmware changed belongs in the purchase order, not in an upgrade plan. Retrofit is real for cable-side hardware and largely fictional for internal sensing.
Retrofit compatibility carries a commercial cost as well as a technical one. An armored sleeve that does not fit the holster or the retractor turns a five-minute cable swap into a two-hour site visit, which is exactly the downtime the measure was supposed to reduce.
Distributors can turn this into a service line: stock spare assemblies, publish a swap procedure, and quote a restoration time in the bid. A competitor ordering a cable from overseas cannot match that number, and it is a stronger differentiator than another discount point.
Settle certification and warranty before you resell the accessory. A retrofit solenoid lock or sensor changes the unit configuration, and the certificate on the wall belongs to the configuration that was tested. ChargePoint’s licensing of its cut-resistant cable technology to third-party manufacturers shows the retrofit market is maturing, not that every accessory is automatically accepted by every platform.
Order matters after an incident. Get it wrong and the evidence is gone before the report is filed.
Socketed or tethered at the next site. Which layer you fund first at sites already installed. Which anti-theft requirements go into the next purchase order as written lines rather than wishes. Those three cover most of the exposure, and none of them require a new budget cycle if you make them at specification stage.
If tamper behavior, cable construction or OCPP alarm mapping will be part of that order, settle it with the engineering team before the configuration is frozen, not after delivery. evaisun’s AC and DC charging platforms hold UL and CE certification and are delivered to partners in more than 30 countries, so the conversation can start from a real configuration instead of a brochure claim. The charging solutions overview is where to check the platform ranges before you write the specification.
Only partly. A cut-detect input, the fail-safe logic and the OCPP event mapping are set at the factory, so an installed unit without them will not gain them from a software update. What can be added later is cable-side: an armored sleeve, a connector or holster lock, better lighting, cameras, and a charge point management system rule that treats an unexpected fault as a possible theft.
It depends on how the policy classifies the loss, so confirm that before the first incident rather than after it. Cable theft may fall under theft, burglary or vandalism wording, and each route can carry a different deductible and evidence requirement. Photographs of both cut ends, the exported event log and the cable part number are what insurers usually ask for, along with a police report number.
US
English中国大陆
简体中文German
Deutsch
Russian
Русский
French
FrançaisSpanish
español
Portuguese
PortuguêsJapanese
日本語
Italian
Italia
Korean
한국인Subscribe the newsletter to get updated to news and promotions