Nvidia CEO warns China's AI infrastructure advantage could reshape the global AI race

Nvidia CEO Sounds Alarm on China's Rapid Advancements in AI

Jensen Huang, Nvidia's chief executive, has issued a stark warning about China AI capabilities and the significant infrastructure advantages that could reshape the global competition for AI dominance. Speaking to John Hamre, president of the Center for Strategic and International Studies, in late November, Huang drew attention to the speed of data center construction that separates the two superpowers. "If you want to build a data center here in the United States from breaking ground to standing up a AI supercomputer is probably about three years," Huang noted, adding that in contrast, "they can build a hospital in a weekend" 1

1

2

2

. This dramatic difference in infrastructure development capabilities represents a critical factor as nations compete to build the AI supercomputing resources needed to power next-generation artificial intelligence.

China Commands Significant Energy Infrastructure for AI Expansion

Beyond construction speed, Huang emphasized China's commanding lead in energy capacity, a resource essential for powering massive data centers. China has "twice as much energy as we have as a nation, and our economy is larger than theirs. Makes no sense to me," the Nvidia CEO observed 2

2

. More concerning for US competitiveness, China's growing energy supply continues to climb "straight up" while American capacity remains relatively flat 3

3

. This AI infrastructure advantage positions China to rapidly scale its artificial intelligence operations without the energy constraints that could hamper US development. Tech leaders and investors have echoed these concerns, with investor Kevin O'Leary stating that energy security, not funding, represents the critical driver of AI data center expansion 3

3

.

China Captures 30% of Global AI Usage Through Open-Source Models

The infrastructure advantages translate into measurable market gains. A report compiled by OpenRouter and venture capital firm Andreessen Horowitz reveals that China now commands nearly 30% of global AI usage through its open-source Large Language Models (LLMs). This figure emerges from analysis of 100 trillion tokens processed by AI systems 1

1

. Just one year ago, Chinese open-source models represented barely 1% of token usage, making this growth trajectory particularly striking. Within the open-source category specifically, Chinese models now average about 13% of weekly token usage, nearly matching the 13.7% from the rest of the world combined 1

1

. The AI race has intensified as DeepSeek, Alibaba Qwen, and Moonshot AI emerge as major players, diversifying China's AI ecosystem beyond any single model.

US Maintains Lead in AI Chip Technology But Faces Challenges

Despite these concerns, Huang maintained that Nvidia remains "generations ahead" of China on AI chip technology and semiconductor manufacturing processes 2

2

. However, he cautioned against complacency, warning that "anybody who thinks China can't manufacture is missing a big idea" 2

2

. The Nvidia CEO has previously stated China is "nanoseconds behind America" in the AI race . Meanwhile, fragmented U.S. regulations pose additional obstacles. Alphabet CEO Sundar Pichai recently urged national AI rules to avoid a confusing patchwork of state laws that could slow American progress 3

3

.

Massive Investment Pours Into US Data Center Construction

American tech companies are responding with substantial investment. Nvidia and other tech giants plan to pour over $100 billion into data center buildout in the US over the next year alone 2

2

. Raul Martynek, CEO of DataBank, estimates the average data center costs $10 million to $15 million per megawatt, with typical facilities requiring 40 MW. Industry experts project 5 to 7 gigawatts of capacity will come online in the coming year, translating to $50 billion on the low end and $105 billion on the high end 2

2

. Huang expressed optimism about President Donald Trump's push to reshore manufacturing jobs and spur AI investments 2

2

. Yet the three-year timeline for US data centers versus China's accelerated construction schedules highlights the scale of the challenge ahead in this intensifying global competition for AI dominance.