Skip to main content
AI in ASIA
US China AI race
News

Huang's dire warning on US-China tech war

NVIDIA CEO Jensen Huang warns China could overtake the US in AI development as government backing gives Chinese firms decisive advantages.

Intelligence Desk4 min read

AI Snapshot

The TL;DR: what matters, fast.

NVIDIA CEO warns China could overtake US in AI development despite current US compute advantage

China's government subsidies and strategic backing give domestic AI companies operational advantages

US deployed 25 zettaFLOPs vs China's less than 1 zettaFLOP in 2025, but gap may narrow rapidly

Advertisement

Advertisement

NVIDIA's Chief Sounds Alarm as China Gains Ground in AI Race

NVIDIA CEO Jensen Huang has delivered his starkest warning yet about America's position in the global AI competition, declaring that China could overtake the US in artificial intelligence development. Speaking at the Financial Times' Future of AI Summit, Huang argued that government backing gives Chinese tech companies a decisive advantage whilst American innovation faces regulatory roadblocks.

The comments come as fresh data reveals the dramatic scale of the technological arms race. Despite US export controls, China has mobilised significant resources toward AI development, integrating artificial intelligence into its national industrial strategy.

The Numbers Behind the Warning

Huang's concerns aren't merely speculative. Recent industry analysis shows both nations are racing to deploy massive computing resources, though current capabilities remain heavily skewed toward the US and its allies.

"China is going to win the AI race. As I have long said, China is nanoseconds behind America in AI. It's vital that America wins by racing ahead and winning developers worldwide." - Jensen Huang, CEO, NVIDIA

The NVIDIA chief has repeatedly emphasised that this isn't just about hardware superiority. It's about building the global developer community that will define AI's future applications.

By The Numbers

  • US and allies deployed 25 zettaFLOPs of compute in 2025, compared to China's less than 1 zettaFLOP
  • NVIDIA sold 4.5 million chips in 2025, contributing roughly 20 zettaFLOPs of computing power
  • China plans to integrate AI into 90% of manufacturing by 2030 through its national strategy
  • Chinese EV production hit over 15 million units in 2025, up 33% year-on-year
  • China's tech exports surged 90% year-on-year, demonstrating rapid scaling capabilities

China's Strategic Advantages Create Competitive Pressure

The fundamental challenge lies in structural differences between the two superpowers' approaches. China's government provides substantial energy subsidies, making power-intensive AI operations significantly cheaper for domestic companies. Huang famously noted that "power is free" in this context, giving Chinese firms a crucial operational advantage.

Meanwhile, American companies face increasing regulatory scrutiny at state and federal levels. Huang suggests this "cynicism" is hampering Western progress, arguing that more optimism is needed to compete effectively with China's comprehensive AI strategy.

The geopolitical tensions have created a complex web of restrictions and counter-measures. US policymakers continue pressuring NVIDIA to limit sales of advanced semiconductors to Chinese companies, whilst China leverages its near-monopoly on rare earth elements as a strategic counter.

Factor United States China
Government Support Limited, regulation-focused Comprehensive subsidies and planning
Energy Costs Market-driven pricing Heavily subsidised for tech companies
Developer Access Global but restricted by export controls Domestic focus with growing capabilities
Manufacturing Integration Gradual private sector adoption 90% target by 2030 via state planning

Export Controls May Be Backfiring

Perhaps most concerning for US policymakers is evidence that export restrictions might be accelerating rather than slowing Chinese AI development. Huang pointed out in May that the US crackdown on chip exports has been a "failure" in some ways, spurring Chinese tech firms to fast-track their own AI capabilities.

"The impact [of US export controls on advanced chips] has not been especially significant. China had already planned to reduce its reliance on American [technology]." - Mark Kennedy, Technology Policy Expert

This dynamic is playing out across Asia's broader tech landscape. The AI chip restrictions have prompted price increases as companies scramble for alternative supplies, whilst Chinese firms like Tencent launch competitive AI models to reduce foreign dependence.

The White House has confirmed that the Trump administration won't allow NVIDIA to sell its most advanced Blackwell chips to China, illustrating the ongoing tension between geopolitical strategy and technological collaboration. Yet this approach may be creating precisely the outcome it sought to prevent.

The Developer Dilemma

Huang's most pointed criticism centres on America potentially losing access to global AI talent. At NVIDIA's GTC event, he argued that isolation could prove counterproductive:

The following challenges highlight why maintaining global developer relationships matters:

  • Half the world's AI developers are based in China, representing a massive talent pool
  • Cutting ties could accelerate Chinese development of indigenous AI stacks
  • American technology adoption depends on global developer communities
  • Innovation thrives on diverse perspectives and collaborative development
  • Export restrictions may fragment the global AI ecosystem permanently

The stakes extend beyond mere market share. As AI becomes integral to everything from manufacturing to defence, the nation that leads in AI development will shape global technological standards for decades.

Is China really ahead in AI development?

Current computing capacity suggests the US maintains a significant lead, with 25 times more deployed compute power. However, China's rapid progress in specific applications and manufacturing integration shows the gap is narrowing faster than many expected.

How do export controls affect the global AI race?

Export restrictions have had mixed results, limiting China's access to cutting-edge hardware whilst spurring domestic innovation. Some experts argue these measures may be accelerating rather than slowing Chinese AI development through forced self-reliance.

What role do energy subsidies play in AI competition?

China's energy subsidies create substantial cost advantages for AI operations, which are notoriously power-intensive. This allows Chinese companies to reinvest savings into research and development, potentially creating a compounding competitive advantage over time.

Can the US compete without Chinese developers?

Huang argues that losing access to China's developer community would seriously hamper American AI leadership. With roughly half the world's AI talent based in China, isolation could fragment innovation and slow global technology advancement.

What's at stake in the US-China AI race?

The victor in AI competition will likely set global standards for decades, influencing everything from industrial automation to consumer technology. This extends far beyond economic benefits to fundamental questions of technological sovereignty and geopolitical influence.

The AIinASIA View: Huang's warnings deserve serious attention, but the picture is more nuanced than headlines suggest. Whilst China has mobilised impressive resources and government support, the US retains substantial advantages in computing infrastructure and global developer networks. The real risk lies not in any single factor but in policy approaches that fragment the global AI ecosystem. Rather than racing toward isolation, both superpowers would benefit more from competition that drives innovation whilst maintaining the collaborative networks that have historically accelerated technological progress. The challenge is balancing legitimate security concerns with the openness that fuels breakthrough developments.

The AI race between the US and China represents one of the defining technological competitions of our time. As both nations pour resources into artificial intelligence development, the outcomes will reshape not just their own economies but the entire global technology landscape. The broader implications for Asia's tech industry and China's domestic AI ecosystem continue evolving rapidly.

What's your view on Huang's warnings about America's position in the AI race? Drop your take in the comments below.

YOUR TAKE

We cover the story. You tell us what it means on the ground.

What did you think?

Written by

Share your thoughts

Join 3 readers in the discussion below

This is a developing story

We're tracking this across Asia-Pacific and may update with new developments, follow-ups and regional context.

Advertisement

Advertisement

This article is part of the China's AI Regulatory Model learning path.

Continue the path →

Latest Comments (3)

Tony Leung@tonyleung
AI
26 November 2025

Huang's point about China's energy subsidies making AI chip power "free" really sticks out. In HK, managing operational costs, especially power for data centers, is a constant battle with regulatory overheads. Wonder how much that "free power" translates into a competitive edge for their AI development long-term.

Rohan Kumar
Rohan Kumar@rohank
AI
16 November 2025

@rohank: Jensen Huang is right about the competition but maybe looking at it from the wrong angle! "Power is free" for Chinese companies right now, yeah, big advantage. But what about the talent? My clients here in Hyderabad, they're seeing so much demand for AI automation. We're hiring like crazy to keep up. The human ingenuity, the sheer brainpower in Asia... that's not something you get with just energy subsidies. The West might have rules, but we have an incredible talent pool ready to build cool stuff. That's a bigger "free power" if you ask me! We're showing that every day with our solutions for e-commerce and logistics. It's all about execution!

Alex Kim
Alex Kim@alexk
AI
12 November 2025

huang's worried about China's government backing and energy subsidies. I've seen that kind of "free power" situation before. it sounds great on paper, massive advantage for R&D. but when you're talking about taking those cutting edge models and actually making them work at scale for real-world enterprise problems, that's where the rubber meets the road. "power is free" is one thing, but reliable, scalable infrastructure that can handle constant production loads and integrate with legacy systems? that's a whole different beast. I'm not convinced those subsidies automatically translate to winning the overall AI race, especially on the application side. the gap from demo to production is always bigger than people think.

Leave a Comment

Your email will not be published