US Ban on Nvidia’s AI Chips Bypassed by Chinese Companies Through Global Data Centers

The US ban on the sale of Nvidia’s high-performance AI chips to China has not stopped Chinese companies from gaining access to these processors. According to a report by The Information, Chinese firms are still able to utilize Nvidia’s AI chips by renting Nvidia-powered servers at data centers run by tech giants like Google, Microsoft, and other global firms outside of China.

This workaround highlights a loophole in US export rules, with Google Cloud and Microsoft Azure offering to rent Nvidia’s AI chips to Chinese companies, including AI startups. Additionally, reports suggest that even Elon Musk’s X is in talks with Oracle to rent Nvidia’s highly sought-after H100 AI processors.

The report also mentions that smaller cloud providers specializing in offering access to Nvidia-powered servers worldwide are available to Chinese customers. Some of these providers are based in the US, while others are located in Europe and Asia.

Furthermore, recent reports indicate that Chinese developers can still access banned AI models, such as those used to build AI products like ChatGPT, through Microsoft’s Azure cloud service. Despite restrictions on certain technology in China, developers in the country can tap into these models via Microsoft’s service, which operates through a local joint venture with a Chinese company, 21Vianet.

It is clear that despite efforts to restrict access to advanced AI technology, Chinese companies are finding ways to leverage these powerful processors through partnerships with global tech companies and cloud providers. This raises questions about the effectiveness of export regulations and the challenges of controlling the spread of technology in an interconnected world.