US export enforcement is intensifying around high-end GPUs, as shown by the smuggling arrest and new retail non-export declarations. The constrained supply and pricing of 32GB VRAM accelerators also affect local AI deployment economics.
A modder addressed melting power connectors on RTX 5090 cards by building a custom distributor. The episode points to tight design margins on high-end GPUs used for continuous local inference, with implications for safety and reliability in self-host...
Chinese chip company Biren is preparing the BR20X GPU, a new accelerator aimed at Nvidia's position in China's AI market. With no public specs yet, the move signals mounting competitive pressure on domestic hardware, with implications for on-premise ...
US retailer Micro Center reportedly now requires additional paperwork for RTX 5090 purchases, including a no-export declaration. The move signals that high-end consumer GPUs are being treated as strategic AI hardware, adding friction for self-hosted ...
The US Department of Justice has arrested Earthmade Computer CEO Greg Lui, accused of smuggling Nvidia A100 and H100 server chips worth over $300 million into China using false paperwork and transit through Malaysia and Singapore. The chips, while no...
A user collected eBay prices for GPUs with 32GB VRAM under $1600 and turned them into a chart with Claude's help. This signals a second-hand market that is increasingly central for those wanting local inference without data-center-level spending.
The Indian case highlights an often overlooked issue: declared warranty coverage does not always match actual support. For those using consumer GPUs in on-prem workloads, the Zotac episode adds a TCO variable tied to after-sales reliability, which ca...
