Major Chinese technology firms, including Alibaba and ByteDance, are reportedly circumventing U.S. restrictions on advanced AI chips by training their large language models (LLMs) in overseas data centers. This strategic move, primarily utilizing facilities in Singapore and other Southeast Asian nations, allows these companies to access the necessary computing power despite U.S. export bans designed to curb China’s technological advancement in artificial intelligence.
Key Takeaways
- Chinese tech giants are training AI models in offshore data centers to bypass U.S. restrictions on Nvidia chips.
- Companies like Alibaba and ByteDance are using facilities in Singapore and Southeast Asia.
- This strategy emerged after the U.S. banned the sale of Nvidia’s H20 chip to China.
- Chinese firms are also increasingly using domestic chips for AI "inference" tasks.
Offshore Training Strategy
Since April, when the U.S. government prohibited the sale of Nvidia’s H20 chip, specifically designed for the Chinese market, there has been a noticeable increase in Chinese companies opting for offshore AI model training. Sources indicate that firms are leasing capacity in data centers operated by non-Chinese entities in locations like Singapore and Malaysia. These data centers often house high-end Nvidia hardware, comparable to what U.S. tech giants use for their own AI development.
A Singapore-based data center operator confirmed the trend, stating, "It’s an obvious choice to come here. You need the best chips to train the most cutting-edge models and it’s all legally compliant."
Domestic Chip Development and Inference
While offshore training addresses the need for powerful chips for model development, Chinese companies are also turning to domestically produced chips for "inference." Inference is the process by which AI systems respond to user requests, and it constitutes a growing portion of overall AI workloads. Companies like Huawei are reportedly collaborating with domestic chipmakers, including DeepSeek, to advance China’s AI semiconductor and software capabilities.
DeepSeek, for instance, managed to build a substantial cluster of Nvidia chips before the export bans took effect and is now focusing on domestic development and partnerships with local chip manufacturers. This dual approach allows Chinese firms to continue their AI ambitions while navigating international trade restrictions.
Global Expansion and Data Constraints
Beyond AI training, Chinese tech companies are leveraging these Southeast Asian data centers to serve their international clientele and expand their presence in the global cloud computing market. Expansion into regions like the Middle East is also underway. However, a significant constraint remains: Chinese companies are prohibited from moving private data out of the country. This means that for AI models requiring customization based on specific local client data, the training process must still occur within China.
This situation highlights the ongoing technological race and the adaptive strategies employed by Chinese tech firms in the face of geopolitical challenges.
Sources
- Chinese tech giants may have found a way to bypass US blockade of Nvidia AI chips, The Times of India.
Founder Dinis Guarda
IntelligentHQ Your New Business Network.
IntelligentHQ is a Business network and an expert source for finance, capital markets and intelligence for thousands of global business professionals, startups, and companies.
We exist at the point of intersection between technology, social media, finance and innovation.
IntelligentHQ leverages innovation and scale of social digital technology, analytics, news, and distribution to create an unparalleled, full digital medium and social business networks spectrum.
IntelligentHQ is working hard, to become a trusted, and indispensable source of business news and analytics, within financial services and its associated supply chains and ecosystems
