The US continues to lead in developing AI, but risks falling behind in deploying it at scale. China, once seen primarily as a follower, is positioning itself as a leader in implementation. If this persists, the consequences will be seen in the competitiveness of national economies.
When it comes to artificial intelligence, the United States still dominates the headlines – and, by most conventional measures, the technology itself. American institutions continue to produce a large share of high-impact AI research, and private investment reached over US$109 billion in 2024, nearly 12 times China’s total, according to the Stanford Institute for Human-Centred AI.
At the same time, the economics of AI are rapidly improving. Training and deployment costs have fallen dramatically in recent years, making large-scale adoption increasingly viable across industries. By these metrics, the US appears to be winning the AI race. But there is growing evidence this may not be the race that matters most. Because, while the US excels at building AI, China is moving more decisively to use it.
Across industries – from logistics to healthcare – China is not simply adopting AI tools. It is reorganising systems around them. By 2024, China had more than 600 million registered generative AI users and hundreds of models deployed across real-world environments, from hospitals to logistics systems. Adoption is not limited to experimentation; it is embedded in operations. This difference is not primarily about technological capability. It is about implementation.
In the US, many organisations are attempting to integrate AI into systems designed decades ago. Nowhere is this more visible than in logistics. The US trucking sector – responsible for moving roughly 70 per cent of the nation’s goods – has access to advanced AI tools but often struggles to translate that access into meaningful transformation. Instead of redesigning workflows, companies frequently layer AI onto legacy infrastructure built in an era of fax machines and dial-up internet, producing incremental gains rather than systemic change.
Economists have seen this pattern before. During the electrification of manufacturing, factories that simply replaced steam engines with electric motors saw little productivity improvement. Real gains only came when companies reorganised entire production systems around electricity – a dynamic often associated with economist Robert Solow and later expanded by Erik Brynjolfsson in the context of digital transformation.
AI may be following a similar trajectory. China appears more willing to make the leap into AI than the US. Companies such as Cainiao, the logistics arm of Alibaba Group Holding, have been built around real-time data integration, automation and optimisation rather than retrofitting older processes.
At the national level, large-scale investment reinforces AI as infrastructure rather than a discretionary tool. McKinsey & Co estimates that AI could generate more than US$380 billion a year in economic value in China’s transport sector alone, contributing roughly US$600 billion annually across the economy.
The result is a compounding advantage. Systems designed around AI generate more data, more data improves performance and improved performance drives further adoption. Over time, this creates a feedback loop difficult to replicate in more fragmented markets.
Part of the difference may also lie in deeper institutional structures. The US is, in many ways, a society shaped by lawyers. Its systems emphasise regulation, compliance and risk management – strengths that support accountability and trust but can slow large-scale transformation. Legal complexity, liability concerns and decentralised decision-making can make it difficult to redesign entire industries around emerging technologies.
China’s model operates differently. Its economy is more centrally coordinated and, in many sectors, more engineering-driven. Many, if not most, of the top political leaders in China trained as engineers. Decision-making often prioritises system performance, integration and long-term infrastructure development. When combined with state-backed investment and aligned policy, this structure makes it easier to redesign processes around new technologies rather than adapting them incrementally.
Viewed through a lens of a lawyerly society like the US, AI is seen primarily as risk and liability. When viewed through the lens of an engineering society like China, AI is seen primarily as possibility and improvement. This difference in perspective matters and helps explain the widening gap in AI adoption. According to Stanford’s AI Index Report research, 83 per cent of Chinese people have positive attitudes towards AI, seeing it as more beneficial than harmful. In the US, only 39 per cent of people have a positive attitude towards AI.
At the core of this divide is a question of trust. In the US, hesitation around AI is shaped by concerns over job displacement, increased workplace demands and data privacy. These concerns are not unfounded. But they do influence how quickly organisations are willing to commit to transformation. In China, AI is more often framed as a practical tool embedded in everyday systems that improve life and work, encouraging faster integration.
This creates a paradox. The US continues to lead in developing AI, but risks falling behind in deploying it at scale. China, once seen primarily as a follower, is positioning itself as a leader in implementation. If that gap persists, the consequences will not be abstract. They will appear in supply chains, production costs and the relative competitiveness of national economies.
The future of AI will not be decided solely in research labs or venture capital markets. It will be determined in warehouses, hospitals and transport networks – where technology either reshapes systems or quietly fails to deliver. In that sense, the most important question is no longer who builds the most advanced AI. It is who is willing – and able – to change enough to use it.
Republished from South China Morning Post, 5 May 2026
Matt Terrell
Matt Terrell is an assistant professor in the School of Communication and Media at Kennesaw State University in Atlanta.