3 days ago
8
Nvidia is set to incur a $5.5bn charge due to US export restrictions on its H20 AI chip to China, as part of broader efforts to maintain a competitive edge in AI, while adapting its chip designs to comply with US regulations.
Nvidia has faced a $5.5bn charge following US export restrictions on AI chips. (Photo: Below the Sky/ Shutterstock)
Nvidia is set to face a $5.5bn charge following the US government’s decision to restrict exports of its H20 AI chip to China, a significant market for the company. The US government has targeted Nvidia’s AI chips as part of export controls to maintain its lead in the AI sector. In response, Nvidia has been developing chips that align closely with US regulations.
The US Commerce Department announced new licensing requirements affecting Nvidia’s H20, AMD’s MI308, and similar chips. “The Commerce Department is committed to acting on the President’s directive to safeguard our national and economic security,” said a department spokesperson.
Chinese tech giants increase orders for Nvidia chips despite US export controls
The H20 chip is Nvidia’s most advanced offering available for sale in China, playing a crucial role in engaging with China’s expanding AI industry. Chinese firms such as Tencent, Alibaba, and ByteDance have been increasing orders for these chips, driven by demand for affordable AI models from startups like DeepSeek.
Although the H20 chip is not the fastest for training AI models compared to those sold outside China, it competes effectively in inference tasks, which are becoming a substantial segment of the AI chip market. Nvidia CEO Jensen Huang recently highlighted the company’s strong position in this evolving market.
The US government cited concerns about the potential use of H20 chips in supercomputers as the reason for the sales restrictions to China. Despite its lower computing power, the H20’s high-speed connectivity to memory and other computing chips remains significant.
Earlier this week, Nvidia announced plans to invest up to $500bn in AI server construction in the US over the next four years. This investment involves partnerships with companies such as TSMC, Foxconn, Wistron, Amkor, and SPIL. Production of Blackwell AI chips will take place at TSMC’s Phoenix, Arizona facility, while supercomputer manufacturing plants will be established in Texas, with operations expected to commence within 12 to 15 months.
Nvidia has secured over one million square feet of manufacturing space to develop and test Blackwell chips in Arizona and AI supercomputers in Texas. The company noted the complexity of the supply chain for these products, requiring advanced technologies for manufacturing, packaging, assembly, and testing. Nvidia will collaborate with Amkor and SPIL in Arizona for packaging and testing processes.
Read more: Nvidia commits up to $500bn to build AI infrastructure in US
More Relevant
close
Sign up to the newsletter: In Brief
Your corporate email address *
I would also like to subscribe to:
Vist our Privacy Policy for more information about our services, how we may use, process and share your personal data, including information of your rights in respect of your personal data and how you can unsubscribe from future marketing communications. Our services are intended for corporate subscribers and you warrant that the email address submitted is your corporate email address.
