The $1.3M theft that exposed AI's blind spot
Thieves stole $1.3 million worth of Nvidia and AMD GPUs from a San Francisco data center by posing as legitimate contractors, highlighting the vulnerability of AI infrastructure to cargo theft and the lack of security protocols around high-value hardware.
Background
- In November 2024, thieves stole a truck loaded with 2,400 Nvidia H100 GPUs (worth ~$1.3M) from a cargo lot in California. This isn't a one-off: organized cargo theft of AI hardware has become a multibillion-dollar problem, targeting the chips and servers that power the entire AI industry.
- Nvidia's H100 (and its successor, the H200/B200) are the de facto standard chip for training and running large AI models. They are in extreme shortage and often sold on secondary markets for far above MSRP, making them perfect high-value, hard-to-trace targets.
- "AI infrastructure" refers to the physical data centers, networking gear, power systems, and especially the GPU clusters that companies (both hyperscalers like Microsoft/Amazon and startups) must build to train or serve AI.
- The article argues that the physical supply chain — moving these chips from factories to data centers — has become a critical security weak point that the AI industry is only now starting to take seriously.