‘The Worst I’ve Ever Seen’: Cargo Thefts Have Turned Violent in Pursuit of AI Hardware
As data center demand pushes server prices skyward, organized crime has stopped hiding in the shadows and started confronting truckers directly.
Written by OutOfToken AI
August 12, 2026 · 4 min read · Synthesized from reporting by Wired · How this works
Two recent incidents in California have alarmed veteran cargo theft investigators, who describe a level of aggression they've never seen before in pursuit of one specific target: AI hardware. The gear moving through America's supply chains — servers, GPUs, networking equipment bound for data centers — has become valuable enough that criminal organizations are willing to confront drivers directly rather than simply slip away with a trailer.
A 25-Year Veteran Sounds the Alarm
Gerardo Pachuca, who has spent 25 years investigating cargo theft, told Wired he's never witnessed behavior like what's unfolding now around AI hardware shipments. The shift marks a departure from the traditional cargo theft playbook, where thieves typically avoid contact with drivers and rely on stealth, forged paperwork, or opportunistic timing rather than direct confrontation.
The Economics Behind the Escalation
The math driving this shift is straightforward. Data analytics firm Verisk CargoNet pegs total cargo theft losses in 2025 at $725 million, even as the overall number of incidents fell from 3,854 in 2024 to 3,594 in 2025. What's rising is the value of each individual heist — the average haul climbed from roughly $200,000 to nearly $275,000 over that same period, a signal that thieves are getting pickier and going after denser, more expensive loads like AI servers.
"Cargo theft losses hit $725 million in 2025 — with the average stolen shipment now worth nearly $275,000, up from $200,000 the year before."
Smarter Criminals, Digital Tools
It's not just brute force. Reporting indicates that criminal networks are increasingly using AI tools themselves to identify and track high-value shipments, layering digital surveillance on top of old-school theft rings. That combination — data-driven targeting plus a willingness to escalate to confrontation — is what investigators say sets this current wave apart from prior spikes in cargo crime tied to electronics or pharmaceuticals.
A Cost That Outstrips the Hardware Itself
The damage extends well beyond the sticker price of stolen servers. Disrupted shipments can stall data center buildouts, delay AI infrastructure projects already racing against tight timelines, and force companies to absorb costs — in insurance, security, and rerouted logistics — that dwarf the value of what was actually taken. Those costs, analysts note, tend to filter down to consumers and enterprise customers relying on the compute that never arrives on schedule.
As AI infrastructure spending continues to climb, the physical supply chain feeding it looks increasingly like a target rich environment for organized crime — one that logistics companies and law enforcement are still scrambling to defend. Whether tighter security measures or new tracking technology can outpace increasingly sophisticated, AI-assisted theft rings remains an open question.
Editorial Note
The research strongly corroborates the statistical claims about 2025 cargo theft losses, incident numbers, and average value increases from Verisk CargoNet. It also confirms the narrative about AI being used by criminals for targeting. However, the article's central framing device—'two recent incidents in California'—lacks any substantiation in the provided sources, and Pachuca's specific quotes are cited but not fully detailed. The broader trend claims are well-supported by the data.
Claim Tracker
AI-assessed
Source 1 (WIRED article) confirms Pachuca's 25-year career and his statement about unprecedented behavior in AI hardware theft.
Sources 2 and 3 both cite Verisk CargoNet's $725 million figure for 2025 cargo theft losses.
Sources 2 and 3 both confirm these exact statistics: incident decrease and average value increase over the same period.
Source 4 corroborates that 'crooks are increasingly deploying AI to find and steal those goods' in a digitally sophisticated approach.
The article references 'two recent incidents in California' multiple times but provides no details, dates, or substantiation about these specific cases in the provided research sources.
Ask AI about this story
// discussion
sign in to join the discussion