Signal

How Jensen Huang built a multibillion-dollar business to let Chinese companies access Nvidia AI chips in SE Asian data centers, exploiting a gap in US sanctions

First reported by WSJ ·

The signal ●●●● Compiled by AI from WSJ, Techmeme, CNN, Fortune, BBC and 3 more
Why you might care

Chinese AI models are now competitive and cost-effective, potentially lowering AI service prices globally.

What happened

AI pioneer Kai-Fu Lee asserts that U.S. export restrictions on advanced chips have inadvertently spurred Chinese AI companies to develop highly competitive open-weight models with fewer resources. Lee argues that Chinese companies prioritize profitability over the pursuit of Artificial General Intelligence (AGI), contrasting with the U.S. approach of heavy investment in computing power and top researchers. He cites the example of DeepSeek, a Chinese startup that released a capable AI model in early 2025 with significantly less computing power, demonstrating that Chinese firms can operate AI at a fraction of the cost of their American counterparts. Lee refutes the idea that China's embrace of open-source technology is solely a government-driven effort to compete with the U.S., attributing it instead to intense market competition and the drive to serve investors. He estimates that Chinese open-weight models are now only six to 12 months behind leading American AI companies.

What it means

Kai-Fu Lee's analysis suggests that U.S. sanctions have fostered a new wave of Chinese AI innovation, leading to models that are remarkably cost-efficient and only slightly behind leading U.S. counterparts. This implies a market shift where competition will be driven not only by cutting-edge capabilities but also by affordability and resourcefulness. The success of Chinese firms in developing powerful models with less computational power challenges the prevailing narrative that immense spending is the sole determinant of AI advancement. It highlights the strategic advantage of Chinese companies in optimizing for profitability and operational efficiency, which could put pressure on global pricing for AI services.

The implications for the AI landscape are substantial, suggesting a more democratized access to advanced AI capabilities and potentially a faster pace of innovation globally. As Chinese companies leverage their cost-effective models, they could capture market share from U.S. competitors, especially in regions where cost is a significant factor. Investors and developers alike will need to re-evaluate strategies, considering the growing capabilities and financial acumen of Chinese AI firms. The next steps to watch will be how U.S. companies adapt to this more competitive and cost-conscious market, and whether their technological lead can be sustained against increasingly efficient rivals.

AI-written summary. May contain errors.

AI Funding