In a working paper released by the Harvard Business School AI Institute in May 2026, researchers Hao Chen and Meg Rithmire unveil a transformative perspective on China’s approach to artificial intelligence. Titled “China’s Diffusion-Forward AI Strategy: Chatbots, Robots, and Political Economic Possibilities,” this research challenges conventional narratives that often frame the AI competition as a race for artificial general intelligence (AGI) leadership, primarily driven by Silicon Valley’s chatbot-centric innovation.
Instead, Chen and Rithmire argue that China is pursuing a fundamentally different vision, one that prioritizes embedding AI deeply into the physical economy. This “diffusion-forward” strategy could give China a strategic advantage in reshaping the future of manufacturing, robotics, and logistics, positioning it to potentially win the global AI race.
The “Diffusion-Forward” Strategy: AI as a Production Input
At the heart of the research is a key insight into the divergent AI philosophies between the US and China. The United States, shaped by a service economy and private capital markets, focuses on AI as a cognitive amplification tool, aiming for breakthroughs in AGI and software-driven productivity. By contrast, China treats AI as an input into production processes, deeply integrating it into factories, supply chains, and physical automation.
Chen and Rithmire write, “The priority has been to deploy AI in a diverse set of sectors and production stages in the physical economy.” Rather than viewing AI as an end in itself, China’s approach is utilitarian and practical: “AI was not imagined as an end in itself.” This focus on diffusion means the technology is being embedded broadly across sectors, from robotics manufacturing hubs in Shenzhen to logistics networks across the country.
The researchers encapsulate this contrast with a striking observation: “While Silicon Valley builds chatbots, Shenzhen builds robots.” This succinctly illustrates the differing priorities, chatbot software dominates American AI innovation, while China spearheads the physical automation revolution.
Institutional Strength and Investor-State Model: China’s Unique Advantage
China’s AI success is not just technological but also institutional. The government plays an active role as an investor, regulator, and strategic coordinator. By early 2026, the Cyberspace Administration of China (CAC) had registered over 1,177 generative AI services, with roughly 25% linked to state-affiliated entities. This signals a tightly integrated ecosystem where public and private sectors collaborate closely.
More critically, China employs what the authors call an “investor-state” model, where the state operates as an equity investor through over 2,000 government guidance funds channeling trillions of RMB into strategic sectors like semiconductors and robotics. This contrasts with the more hands-off regulatory stance typical in the US.
A compelling case study highlighted in the paper is UBTECH Robotics. The Liuzhou municipal government invested 200 million RMB in the company, which has accumulated 4,091 patent filings and spends an average of 62.4% of its revenue on R&D. UBTECH’s Liuzhou plant rapidly scaled production, going from manufacturing its first Walker S2 robot in June 2025 to producing its 1,000th unit by December 2025. This rapid industrial scaling exemplifies how China’s diffusion-forward strategy combines state capital with industrial capacity to accelerate AI-driven robotics. EastFrontier has previously reported on UBTECH Robotics’ large state grid deal and the company’s efforts to attract top talent.
Policy Initiatives Supporting AI Integration in the Real Economy
China’s “AI+ Initiative,” launched in 2024, is a policy blueprint calling for the deep integration of AI with the real economy. This initiative aligns closely with the diffusion-forward strategy by encouraging the deployment of AI across manufacturing, logistics, and other physical sectors. The HBS paper notes that this initiative is not just rhetoric; it translates into substantial funding and coordinated industrial policy.
The diffusion-forward approach’s emphasis on physical infrastructure integration suggests that China is building a foundation for sustained AI-driven productivity rather than chasing speculative breakthroughs in AGI. This pragmatic focus may yield long-term competitive advantages, particularly as global supply chains and manufacturing undergo digital transformation.
Challenges and Constraints: Concentrated Deployment and Economic Risks
Despite its strengths, China’s AI strategy faces significant challenges. The diffusion-forward model remains geographically concentrated, with AI deployment heavily focused in Beijing and Shanghai. This regional imbalance could limit the diffusion of AI benefits across the broader economy.
Additionally, the paper acknowledges economic risks. China faces fiscal strains from past industrial overinvestment, which could constrain future state-led funding initiatives. Furthermore, the acceleration of automation may exacerbate youth unemployment, a critical social issue, as machines displace human workers, especially in manufacturing sectors.
Chen and Rithmire caution that “The success of either country’s strategy will also depend on the accompanying economic and political strengths and fragilities and how well the institutional structures adapt.” This underscores that technological capability alone is not decisive; political economy and institutional resilience are equally vital.
Why China’s AI Strategy Could Win the Global Race
The diffusion-forward AI strategy marks a distinct strategic path for China that aligns AI development with the country’s industrial and economic strengths. By embedding AI into the physical economy, factories, robotics, and logistics, China is not merely innovating technology but reshaping the foundational structure of production and economic growth.
This approach could yield multiple advantages. First, it leverages China’s vast manufacturing base and government capacity to deploy capital efficiently. Second, it creates an AI ecosystem that is less vulnerable to the cyclical uncertainties of software markets. Third, by focusing on tangible productivity gains in physical sectors, China may build a more sustainable competitive edge than a narrow focus on AGI development.
The HBS research signals that the AI race is not a zero-sum battle over abstract intelligence but a multifaceted competition involving economic models, institutional arrangements, and sectoral priorities. The diffusion-forward strategy could prove decisive in the long run, especially if China manages to mitigate its economic risks and broaden AI deployment beyond coastal hubs.
Reframing the Global AI Competition
The Harvard Business School AI Institute’s research by Hao Chen and Meg Rithmire offers a fresh lens to interpret China’s AI ambitions. Rather than chasing the same AGI-driven path as the US, China’s diffusion-forward strategy integrates AI deeply into the physical economy, supported by robust state investment and industrial policy. This could enable China to transform its manufacturing and logistics sectors, positioning it as a formidable leader in AI-driven economic modernization.
As the global AI competition evolves, understanding these divergent strategies is crucial for policymakers, investors, and industry leaders. The future of AI may depend as much on how countries deploy and institutionalize AI technology as on breakthroughs in algorithms.
(Related: China’s Venture Capital Machine Breaks Records in Q1 2026, Fuelled by State-Backed AI Investment)