November 7, 2025 07:20 AM IST
First published on: Nov 7, 2025 at 06:29 AM IST
The AI race is heating up, with countries across the world drawing up ambitious plans and channelling huge resources into this space. In 2024, global private AI investments touched $252.3 billion, as per Stanford’s Artificial Intelligence Index report. The US widened its lead over other countries, with investments touching $109.1 billion. But, despite pouring in billions of dollars, it may well turn out that China, not the US, wins the AI race. That’s what Nvidia chief Jensen Huang has said in a recent interview. While fears of Chinese dominance have been voiced since the release of Chinese AI model DeepSeek, when the chief of the most valuable company in the world — Nvidia’s market capitalisation recently touched $5 trillion — whose chips form the backbone of the AI ecosystem, says so, it warrants a deeper examination.
According to Huang, who has earlier said that US AI models are not significantly ahead of the Chinese models, China is racing ahead due to lower energy costs and looser regulations. The massive data centres that are needed to handle AI processing require large amounts of power. Goldman Sachs, an investment firm, estimates that with the AI revolution gathering steam, data centre power demand will go up by 160 per cent. This surge in demand is already causing prices to soar in certain areas in the US. Reportedly, the Chinese are addressing this issue by offering energy subsidies that cut costs by half for large data centres, provided they use Chinese chips. This not only helps support their domestic chip industry, it also reduces reliance on Nvidia’s chips, whose supply is itself facing hurdles — the Trump administration has not allowed the firm to sell its most advanced chip, Blackwell, to China.
On the issue of regulation, Huang has argued that the new rules on AI by US states could result in “50 new regulations”. The fragmented regulatory landscape in the US means that firms will have to wrestle with varying requirements across jurisdictions. This could raise costs, impact innovation. As India moves towards building its own AI ecosystem — from chips to data centres and large language models —it must keep these issues in mind.
