Jensen Huang reckons China could come out on top, not through flashier models but through the simple brute force of having more electricity to run them. The claim sounds bold until you stare at the figures and feel your eyebrows lifting.

China added 429 GW of new power capacity to its grid last year, while the US added only 51. The country now generates more than double the electricity and is filing 70 per cent of the world’s AI patents. Washington keeps hurling billions at a tiny clique of outfits such as OpenAI, Anthropic and xAI, while Beijing builds power plants, data centres and efficiency gains that let it scale like no one else.

Jitterbit chief executive and former Interpol, Bush and Blair government adviser Bill Conner says the stakes are about far more than bragging rights.

He said: “AI isn’t just a modern-day space race, it’s an AI trust race. Without robust infrastructure and a clear strategy to attract talent, regulations risk slowing innovation rather than guiding it while it accelerates elsewhere.”

Conner said: “The path forward is about balance, building accountable AI systems that are powerful and reliable, while scaling infrastructure and nurturing talent. Countries that master both rapid execution plus principled oversight will define the next generation of AI leadership.”

He warned that Western firms rushing to outdo each other often ignore the broader geopolitical mess brewing in the background. Evidence keeps mounting that DeepSeek’s infrastructure fits neatly into a wider Chinese defence alignment.

The risks of skewed incentives and sneaky data usage no longer live in theory, and the chase for quick wins could leave governments and companies with long-term exposure through data leaks and reputational burn.