Cargo Thieves Have Stolen Over $150 Million in AI Hardware This Year

Cargo thieves have stolen more than $150 million in AI equipment this year, according to supply chain visibility provider Overhaul, whose AI Infrastructure Intel Report also records a 34% rise in thefts of data center equipment, chips, and computing components since 2024.

The number that should stop shippers cold is the unit economics. A single flagship AI server rack costs roughly $3 million, Overhaul found, meaning one ordinary truckload routinely carries eight figures of value on public roads before installation. Individual losses now regularly exceed $30 million — and thefts at that scale are occurring weekly, not occasionally.

That concentration of value per trailer explains why organized criminal networks have retooled around this freight. It also raises the stakes for carriers and forwarders handling AI infrastructure loads: a single compromised move can erase a year of margin on that account.

How are thieves actually taking the loads?

The answer, according to Overhaul's data, is that they have largely stopped using force. Deception was used in 92% of high-value technology shipment thefts the report examined. The toolkit includes:

• Fraudulent carrier identities
• Double-brokering schemes
• Spoofed communications
• Falsified paperwork

Each method works by slipping inside legitimate freight transactions rather than attacking them from outside. A load is booked, dispatched and tracked — and then simply delivered to the wrong hands, with paperwork that looks clean until it doesn't.

The shift from force to fraud changes what shippers must defend against. Physical security measures — seals, escorts, secure parking — do little against a thief who arrives with a valid-looking carrier identity and falsified documents. Verification of counterparties at every layer of the brokerage chain becomes the control that matters, and double-brokering is the specific structural weakness thieves are exploiting.

What does the CEO say shippers should do?

"AI hardware is some of the most concentrated value moving through the supply chain right now, and it's become the fastest-growing target for these criminal networks," said Barry Conlon, CEO and founder of Overhaul.

His caveat is notable for a company selling visibility services: tracking alone will not stop these thefts. "As thieves become increasingly deceptive, visibility into a shipment alone isn't enough to stop it from being taken," Conlon said. "Our findings give shippers the intelligence to know how crews take loads, where the equipment ends up, and what to put in place now so they can stop the next theft before it happens."

That admission cuts against the industry's standard pitch. If 92% of high-value tech thefts are fraud-based, then GPS dots on a map — the core product of most visibility platforms — confirm a theft in progress rather than prevent one. Conlon is effectively conceding the point while repositioning his firm's offering around intelligence: how crews operate, where stolen equipment ends up, and what controls to install before dispatch.

The report does not, it should be noted, detail where the stolen hardware goes — whether into gray-market resale abroad, secondary data center buildouts, or domestic channels. That gap matters for recovery efforts and for shippers weighing whether tracing capabilities are worth paying for; a claim of tracking destination markets deserves scrutiny until the underlying data is published.

What are the commercial consequences?

For shippers of AI infrastructure — chipmakers, server assemblers, data center builders racing to install capacity — the exposure per load is now measured in the tens of millions. A single $30 million-plus loss weekly across the sector implies annualized theft in the billions if the pace holds, though Overhaul's $150 million year-to-date figure suggests the largest losses are concentrated in a small number of high-value moves.

For carriers and forwarders, the fraud shift means the vetting burden grows. Onboarding a fraudulent carrier or accepting a double-brokered load can expose the intermediary to the full value of the cargo — a liability that, at $3 million per rack, can reach well into eight figures per trailer.

For insurers, weekly seven- and eight-figure losses on a single commodity class will force repricing of high-value technology cargo coverage, and likely tighter requirements around identity verification and routing controls before policies are written.

The trajectory is clear from the data: thefts are rising 34% year over year, the losses are getting larger, and the methods are getting more sophisticated. Shippers moving AI hardware should expect the targeting to intensify as the equipment's value density — millions of dollars per rack — continues to concentrate wealth on ordinary trucks moving down ordinary roads.