If you work in tech, build custom rigs, or try to upgrade hardware lately, you already know the painful reality: staying up-to-date has never been more expensive. Between higher than ever inflation on computer parts and the endless, ravenous appetite of massive AI data centers sucking up available hard drives and RAM, all these parts that used to be affordable have skyrocketed. Parts that once cost around $60 can easily run you $300 to $400 now – roughly a 400% price spike that hurts everyday builders and small businesses alike.
That is precisely why this recent headline hit so close to home!
According to a Department of Justice press release, federal authorities arrested 38-year-old Greg Lui, CEO of Earthmade Computer, for orchestrating an elaborate multi-year scheme to smuggle over $300 million worth of export-restricted technology into China.
Lui allegedly used a web of false paperwork, stolen identities, and international shell companies. To mask the true destination of high-end servers, investigators say he partnered with freight forwarders in Southeast Asian hubs like Malaysia and Singapore.
The trail of evidence uncovered by the FBI and DOJ includes:
At the center of this smuggling ring were NVIDIA A100 and H100 GPUs. While the FBI noted these aren’t NVIDIA’s absolute bleeding-edge models, they are high-performance powerhouses.
These GPUs are specifically designed to ingest massive streams of data and train large language models (LLMs). Washington maintains strict export controls because supplying foreign competitors with enough of these chips risks major national security implications, potentially fast-tracking military applications or leaps in AI capabilities.
It’s important to note that while the NVIDIA A100 and H100 lines are slightly older generation data center GPUs, they remain absolute workhorses for training large language models and running local AI architectures.
Because these enterprise cards have become more accessible on the secondary and commercial markets over time, their relatively balanced cost-to-performance ratio makes them a primary target for illicit stockpiling. When bad actors manage to divert hundreds of millions of dollars worth of these systems overseas, it hands foreign entities—like China’s expanding tech sector—a massive, immediate surge of computing power to fast-track their AI training pipelines.
When you’re shelling out four times what you used to pay for basic components just to keep a workstation running, finding out that domestic resources are being siphoned off into illegal overseas stockpiles is more than just frustrating—it feels like a betrayal of the local tech community.
While mega-corporations and illicit brokers chase massive payouts by feeding foreign AI data centers, the everyday consumer, independent developer, and small-business builder are left footing the bill through inflated hardware costs. Hardware components are already a scarce, precious commodity right now; watching millions of dollars in critical infrastructure get smuggled out under false pretenses shows just how distorted the global tech supply chain has become.
What are your thoughts on the current state of hardware pricing and export security? Drop a comment below!
