Clamping Down on Copper Theft Through AI-Enabled Surveillance

In March last year, Glendale, Arizona residents experienced a series of copper thefts that left entire areas without power. Police say thieves cut lighting systems at parks, schools and construction sites to steal copper wiring, creating damage that took the city months to repair.  

This type of incident is becoming more common as demand for copper continues to rise. Over the last two years, reported copper thefts have increased by as much as 30%, costing U.S. businesses around $1 billion a year, according to the U.S. Department of Energy. 

Much of this new demand is attributed to the creation of new and expansion of existing data centers to provide capacity for skyrocketing growth of artificial intelligence. More construction means greater demand for building materials, especially copper, which has caught the attention of thieves around the world.

Even as companies adapt alternative sources such as fiber optics and aluminium, copper remains essential for power distribution, grounding and network infrastructure for modern data center designs. Large-scale facilities can require massive volumes of copper, concentrated in small areas at active construction zones.

As copper demand rises, so too does its value, creating a modern-day gold rush for copper thieves at construction sites, shopping malls and a variety of other buildings.

Residential developments, commercial builds and infrastructure projects are experiencing the same pattern: stolen materials, damaged systems and unexpected delays caused by increasingly bold and aggressive copper theft.

Copper thieves are zeroing in on construction sites, which are built for efficiency rather than restriction. Materials are delivered in bulk and staged for installation, often sitting for days or weeks before they are fully integrated into secured systems.

Retail properties and malls have become especially vulnerable as thieves target rooftop HVAC systems to extract copper coils, refrigerant lines and wiring from equipment that is often out of sight and lightly monitored. College campuses and hospitals share these same exposures. 

Whether the setting is a construction site or a finished facility, the underlying vulnerability includes a combination of valuable materials, limited visibility and predictable periods of low activity.

Copper theft today follows a simple logic: go where access is easiest and detection is least likely.

The result is a temporary but critical gap between material delivery and full operational security  that thieves are increasingly exploiting.

Small Thefts Lead to Big Problems

What makes copper theft particularly disruptive is not only the value of the material itself, but the systems it supports. Even small amounts of theft can create outsized consequences. In the case of Glendale, Arizona, city officials spent months repairing damage that was inflicted within minutes.  

Removing a relatively small section of cable can damage surrounding infrastructure, forcing teams to replace more than what was taken. Electrical systems that have been tampered with must be reinspected, retested and recertified, adding time and cost well beyond the initial loss.

These disruptions rarely occur in isolation. Copper is embedded in systems that sit on the critical path of construction. When power distribution or grounding systems are affected, downstream work often stops. Commissioning schedules slip, sequencing is disrupted and recovery efforts introduce additional labor and coordination challenges. 

In data center projects, where timelines are tied to customer demand and capacity commitments, even minor delays can carry broader implications. Across construction more generally, the result is often the same: schedule pressure, cost escalation and operational uncertainty.

AI-Enabled Video Surveillance Tightens Security

Despite these risks, traditional approaches to construction site security remain largely reactive. Guards, fencing and standard camera systems can provide a level of deterrence, but they are difficult to scale across large or complex sites. Monitoring multiple camera feeds in real time is inherently challenging, and many incidents are only discovered after materials have been removed and damage has occurred.

This is where AI-enabled video surveillance is beginning to change the equation.

Rather than relying solely on human monitoring, AI-driven systems continuously analyze activity across camera networks, identifying patterns and flagging behavior that deviates from what is considered normal for a given environment. This shift from passive observation to active analysis is particularly valuable in construction settings, where activity levels fluctuate and sites are often unoccupied during nights and weekends.

By detecting anomalies such as unexpected movement, unauthorized access or unusual dwell times near stored materials, these systems can provide early warning of potential theft. In many cases, that means security teams can respond before materials are removed or damage is done. As these AI-enabled surveillance systems learn what is normal, their accuracy and effectiveness continue to improve.

Macnica, in partnership with icetana AI, is advancing this approach through self-learning AI video analytics designed specifically for complex, real-world environments. Unlike rule-based systems that rely on predefined thresholds, icetana AI learns the normal patterns of activity within a site and identifies deviations that may signal risk. This allows it to adapt to the unique conditions of each location, whether it’s a large construction project, shopping mall or school campus.

The value of this approach lies in its ability to scale. Instead of requiring personnel to monitor dozens or hundreds of video feeds, the system surfaces only the events that matter, enabling faster and more focused response. It also extends visibility into off-hours, when many theft incidents occur and when traditional monitoring is most limited.

Deployments in complex environments such as retail centers and public spaces have already demonstrated how this model can reduce crime, improve situational awareness and optimize security resources. Applied to construction sites, the same capabilities offer a way to close the gap between exposure and response.

As AI continues to drive rapid expansion in data center construction—and as broader construction activity continues to grow—the industry is entering a period where material security can no longer be treated as a secondary concern. Copper theft is not just a site-level issue. It is an operational risk that can disrupt timelines, increase costs and impact project delivery across sectors.

Addressing it requires a shift in mindset. Security must move from reactive to proactive, from manual oversight to intelligent monitoring.

AI-enabled video surveillance provides a practical path forward. By helping teams detect threats earlier, respond faster and maintain visibility across complex environments, it plays a critical role in protecting materials, preserving schedules and ensuring that the infrastructure being built today can be delivered as planned.

Please feel free to reach out to our team if you’re dealing with large-scale builds, data center expansion, or high-value materials!