The story of the “KitKat heist,” in which a truckload of nearly 414,000 special-edition chocolate bars (about 12 tons) disappeared between Italy and Poland, sounds zany. The truck didn’t take a wrong turn; it effectively dropped off the face of the Earth.

And though Nestlé leaned into the mystery, launching a Stolen KitKat Tracker to help fans check batch codes, the shipping industry was left with an un-zany question. In an age of high-tech logistics, how does 12 tons of cargo vanish?

Spoiler alert: It doesn’t. The answer lies in the limitations of traditional supply chain tracking.

Last year, in the U.S. and Canada, supply chain theft reached roughly $725 million in losses, a 60 percent year-over-year increase. However, the number of incidents didn’t rise at nearly the same pace. In other words, criminals aren’t just stealing a little more. They’re stealing a lot smarter.

Can the supply chain smarten up? Yes, through Physical AI.

‍

Exploiting Blind Spots in the Supply Chain

‍

Online retailers are feeling similar effects, with criminals siphoning off goods meant for warehouses and other destinations. Often, bad actors take advantage of weak handoff processes. But almost all incidents share a common thread with the KitKat heist. Attackers are exploiting blind spots in the supply chain.

This is where Physical AI, powered by Wiliot’s battery-free IoT Pixels, changes the equation. In traditional supply chains, visibility is limited. Companies track shipments at a high level – containers, pallets, or trucks – but rarely at the level of the goods themselves.

Once a shipment leaves a facility, the system trusts it will arrive where it’s supposed to go. That’s a fragile assumption.

In the case of the KitKat heist, once the truck was intercepted, things went dark. There was no way to know if the cargo had been split, rerouted, or quietly introduced into secondary markets. Similarly, in frauds affecting large retailers, shipments appear fine in tracking systems until they’re quietly diverted, because there is no continuous, item-level verification.

Wiliot’s Physical AI platform injects smarts into the supply chain. Instead of tracking macro assets, it enables intelligence at a granular level. Unlike traditional tracking systems that rely on manual scans or battery-powered devices, Wiliot uses small, energy-harvesting IoT Pixels that broadcast signals without needing a battery. These tags can be embedded in packaging or attached to products, allowing them to be detected by nearby Bluetooth infrastructure, including smartphones, access points, or dedicated receivers.

The result is a living, dynamic map of where goods are and how they move. Every item becomes a data point, capable of constantly communicating its location, status, and journey through the supply chain.

More than tracking, it’s awareness. And the effects are profound. 

‍

Here are five ways Physical AI can enhance supply chain security and prevent future heists:

‍

1. Real-time detection of anomalies and dwell time

One of the most powerful aspects of Physical AI is its ability to understand what should be happening and flag when something deviates from that expectation. In a normal shipment, goods follow a predictable path: warehouse, transit, distribution center, store. Wiliot’s Physical AI platform learns these patterns and continuously compares real-world behavior against them. Plus, our IoT Pixels can sense light, temperature, and humidity changes, as well as dwell time, in the event that cargo idles in an unexpected environment.

These are precisely the types of events that create sudden, detectable deviations from expected behavior. Most cargo thefts involve a "cooling off" period, during which goods are stored in a secondary location before being fenced. So if a truck suddenly stops in an unexpected location (between Poland and Italy, perhaps), or if products “appear” in places they shouldn’t, the system can trigger actionable alerts.

‍

2. Breaking the anonymity of stolen goods

Once goods are stolen, they often re-enter the market through informal channels, secondary retailers, or even legitimate storefronts. Without a way to distinguish stolen items from legitimate ones, recovery becomes nearly impossible.

Physical AI introduces a form of persistent identity for products. Each tagged item carries a unique digital signature that can be detected long after it leaves its intended path. This means stolen goods don’t just disappear. They remain visible, emitting signals that can be picked up and analyzed. In effect, Physical AI removes the anonymity that theft depends on.

‍

3. Strengthening chain-of-custody

Many large-scale thefts also rely on deception. Criminals impersonate officials, forge documents, or exploit gaps in verification processes to gain control of shipments. Physical AI adds a digital layer of trust to these interactions.

Each handoff in the supply chain can be verified against expected patterns and authorized entities. If goods are transferred without proper validation – or if they appear in the possession of an unrecognized party – the system can flag the discrepancy. This is particularly important as fraud tactics evolve. 

Organized groups are increasingly using identity spoofing and digital manipulation to reroute shipments before actually stealing them. Physical AI offers the best defense.

‍

4. Extending visibility into stores

While high-profile heists grab attention, a significant portion of retail loss happens at the store level. Shoplifting, employee theft, and inventory shrinkage are persistent challenges. Physical AI extends the same item-level intelligence into retail environments. Products can be monitored as they move through the store, from stockroom to shelf to checkout. If items leave the store without being purchased, or if they are moved in unusual patterns, the system can detect and respond.

This both improves security and enhances operations. Retailers gain better insights into inventory levels, product placement, and customer behavior, creating a dual benefit of loss prevention and efficiency.

‍

5. A shift from reactive to proactive

Today, most supply chain systems are reactive. Losses are identified after the fact, often through audits or delayed reporting. Physical AI enables a proactive model. Issues are detected as they happen – or even before they fully unfold. Companies can intervene in real time, reroute shipments, alert authorities, or take corrective action before losses escalate. This is how supply chains – theft or not – should operate.

‍

From Broad Vulnerability to Essential Visibility

‍

The KitKat heist may stand out for its scale and novelty, but it represents a broader vulnerability in global commerce. From diverted retail shipments to fraudulent carrier schemes, the pattern is clear: Limited visibility creates nefarious opportunity.

And as the data shows, the stakes are rising for supply chains. Food and beverage theft alone jumped nearly 50 percent in 2025, demonstrating that even everyday goods are now targets.

Physical AI offers a new level of visibility, transparency, and control, turning products into intelligent participants in their own journey. Physical AI doesn’t just help recover stolen goods; it makes theft harder to execute, riskier to attempt, and easier to detect.

In a world where even chocolate can vanish without a trace, that kind of visibility is as valuable as it is essential.