How the United States is winning the AI race against China

In the competition for dominance in artificial intelligence (AI), the United States is ahead of China, according to new analysis from the Lowy Institute, with superior US computing power and frontier models proving the difference.

The Data Snapshot, authored by Jack Sato and Hunter Marston and entitled China, the United States, and the race for AI dominance, draws on the most comprehensive available data on AI model adoption. Combined with the findings of the forthcoming 2026 Asia Power Index, the Snapshot reveals that the United States is well ahead in AI capabilities, driven by its dominance in processing power and AI adoption.

The United States’ AI data centres have nearly 40 times the combined computing power of China’s, and its AI chips 31 times. It also has a commanding lead in website technology and apps.

“Most people who use AI rely on US models, giving the United States unmatched reach and greater influence over the technology’s global direction,” write Sato and Marston.

Meanwhile, China holds a clear advantage over the United States in both production and known reserves of critical minerals.

China leads in production of 30 of the 60 minerals identified by the United States as critical to its economic and national security, while the US leads in just eight.

“China’s ability to impose targeted controls on critical minerals poses a major risk for the United States,” write Sato and Marston.

“To counter this vulnerability, the United States has sought to cooperate with other countries to secure critical minerals. It has developed the Pax Silica, a 24-member international framework focused on AI-related supply chain resilience.”

Despite their relative competitive advantages, neither the United States nor China has been able to emerge as a clear leader on global AI governance.

The authors find that neither Washington nor Beijing is widely trusted to regulate AI, with attitudes across Asia varying.

“With the region split, Australia, as a Pax Silica member, is well placed to act as a bridge between the US-led bloc and countries seeking to balance AI safety with investment in their own capabilities,” write Sato and Marston.

KEY FINDINGS

1. The United States is winning the race in artificial intelligence (AI), driven by its vast lead in computing power and strength from its frontier models. US data centres boast nearly 40 times the processing power (or “compute”) of China’s, and AI chips 31 times. Top US models outperform China’s best, consistent with claims of a six-to-nine-month lead, though the margin is narrowing.
   
2. China dominates critical minerals and open AI models, but its overall AI reach is overstated. It leads production of 30 of 60 listed critical minerals, against the United States’ eight, giving it leverage over the supply chains on which AI depends. Its open models are also popular, accounting for 78 per cent of all US and Chinese model downloads. But most people use AI through websites and apps, where US models dominate, drawing more than 90 per cent of web traffic and more than 80 per cent of app use.
   
3. Neither superpower is widely trusted to regulate AI, opening space for middle powers such as Australia. Asian countries are split between the United States’ Pax Silica and China’s World Artificial Intelligence Cooperation Organization (WAICO), with none a member of both. Given its economic integration and institutional ties across the region, Australia is well placed to bridge this divide and help its neighbours lay shared foundations for AI governance.

MEDIA CONTACT
Andrew Griffits
Head of Media and Communications
+61 2 8238 9035
media@lowyinstitute.org (Opens in new window)