After tariff wars, Trump is opening a new front in US-China rivalry

The latest salvo in the US–China technology rivalry comes not in the form of tariffs but as a sharply worded memo on artificial intelligence. Just weeks before a planned summit between Donald Trump and Xi Jinping, the White House has accused China of orchestrating industrial-scale efforts to extract the intellectual property of American AI systems. The charge, laid out by Michael Kratsios, director of the White House Office of Science and Technology Policy, underlines how AI has become the central arena of strategic competition between the two powers. What was once a race for innovation is increasingly becoming a hot contest over control, security and global influence. The timing of the memo indicates that AI tensions could overshadow even high-level diplomacy.

Also Read: A quiet weapons crisis is building up within the US military

The US and China are moving toward a hardened technological rivalry with global consequences. The details behind that rivalry reveal a far more complex and contested landscape than headline claims suggest.


At the heart of the current escalation is the claim that Chinese entities are using “model distillation” to replicate advanced US systems. Technically, distillation involves training a smaller model on outputs generated by a more capable one. It is widely used within the industry to optimise performance and reduce costs. What is new is the US assertion that foreign actors are doing this at scale, targeting proprietary systems developed by firms like OpenAI and Anthropic.

If distillation is treated as legitimate learning, the global AI ecosystem remains interconnected. If it is classified as intellectual property theft, it becomes grounds for sanctions and export controls. The memo indicates that the US is preparing to treat model distillation as intellectual theft. It also suggests a coordinated policy response of intelligence-sharing with private firms, potential blacklisting of offending entities and tighter scrutiny of technology flows.

This is where hardware enters the picture. Advanced chips produced by Nvidia are essential for training frontier models. The latest memo casts doubt on whether these powerful chips will be allowed for sale in China. The Trump administration gave ⁠a green light to the sales in January but with some conditions. Commerce Secretary Howard Lutnick said no shipments had yet been made.

Also Read: AI at war: The new arms race

DeepSeek and the shock of cost disruption

The anxiety in the US, as revealed by the White House memo, is rooted partly in what happened last year with DeepSeek. Its large language model demonstrated performance comparable to leading US systems but at a fraction of the training cost. This was a big technical achievement. It also challenged the core assumption that scale of capital and compute would remain decisive barriers to entry.

DeepSeek’s approach relied on efficiency gains, architectural tweaks and, critics argue, extensive use of outputs from existing models. David Sacks, then serving as President Donald Trump’s AI and crypto adviser, said DeepSeek effectively distilled knowledge from American systems. While such claims are difficult to prove definitively, they have shaped policy thinking in the US.

In February, OpenAI, the developer of ChatGPT wrote to US lawmakers that China should not be allowed to advance “autocratic AI” by “appropriating and repackaging American innovation.” Anthropic, the maker of the Claude chatbot, also accused DeepSeek and two other China-based AI laboratories of engaging in campaigns to “illicitly extract Claude’s capabilities to improve their own models” using the distillation technique.

The deeper concern is structural. If cutting-edge capabilities can be replicated cheaply, then export controls on hardware alone may not be sufficient to maintain a technological edge. This is why the current debate extends beyond chips to include data access, model security and even API usage.

China is closing the AI gap with the US

The Stanford University Institute for Human-Centered Artificial Intelligence 2026 AI Index report provides a more nuanced picture than headlines may suggest. It confirms that the US still leads in frontier model performance, particularly in reasoning-intensive benchmarks and multimodal systems. However, the gap has narrowed significantly in areas like natural language understanding and code generation. The gap between the top AI bots in the US and China is shrinking in Arena scores, a metric that shows relative performances of large language models. In May 2023, the U.S.’s top model, OpenAI’s GPT-4, led with more than 1,300 Arena points compared with China’s fewer than 1,000. By March 2026, the gap shrank from more than 300 to just 39 points. Anthropic’s Claude Opus 4.6 was leading China’s Dola-Seed 2.0 by just 2.7% points.

More striking is China’s dominance in scale metrics. It produces a larger share of global AI publications and patents, indicating both breadth and depth of research activity. In industrial deployment, China leads in robotics integration, particularly in manufacturing and logistics. The report also highlights a surge in domestic AI startups, supported by state funding and regional innovation clusters.

Yet the data also reveals constraints. Chinese models tend to lag slightly in reliability and alignment, partly due to stricter content controls that limit training diversity. Access to cutting-edge chips remains uneven, forcing firms to innovate around hardware bottlenecks. This is where techniques like distillation and optimisation become critical for the Chinese.

The urge to control: Manus and Anthropic

The case of Manus illustrates the tensions within China’s own AI ecosystem. Positioned as a next-generation AI agent capable of performing complex tasks such as property transactions, game development and financial analysis, Manus attracted significant attention. Its subsequent acquisition by Meta triggered immediate scrutiny from Chinese regulators. The response was swift and revealing. The commerce ministry launched an investigation into the deal’s compliance with national regulations. Soon after, authorities reportedly prevented Manus’ founder from leaving the country. This dispute revealed a broader policy stance. For China, advanced AI capabilities are too strategically important to be transferred to foreign entities. The episode also reveals a paradox. China encourages rapid innovation and global competitiveness, yet it imposes strict controls when that innovation risks slipping beyond state oversight. This tension could shape the future trajectory of its AI sector, balancing dynamism against political constraints.

The US approach is also not without internal contradictions. Tensions between the government and Anthropic highlight the challenges of aligning private innovation with national strategy. Disputes have emerged over issues such as military applications of AI and the extent to which companies should cooperate with defence initiatives. Anthropic, which positions itself as a safety-focused developer, has reportedly resisted certain forms of government engagement, particularly those connected to weaponisation. This reflects a broader debate within the US tech ecosystem. While policymakers emphasise national security, companies remain wary of reputational risks and ethical concerns.

Such friction complicates the US position. Unlike China, where the state can more directly steer corporate behaviour, the US must negotiate with powerful private actors whose priorities do not always align with government objectives.

The hidden battlegrounds

Beyond algorithms and chips, two less visible factors are shaping the AI race -- talent and energy. Talent remains an important variable. The US continues to attract top researchers from around the world, benefiting from its universities and innovation ecosystem. However, China is rapidly expanding its domestic talent pool, producing large numbers of AI engineers and offering incentives for overseas experts to return. The competition is no longer just about attracting talent, but about retaining and mobilising it at scale.

Energy is emerging as an equally critical constraint. Training and operating advanced AI models require enormous computational power, which translates into massive electricity consumption. The US faces challenges related to grid capacity and environmental regulations, particularly as data centres proliferate. China, with its ability to rapidly build infrastructure, has an advantage in scaling energy supply. These factors rarely feature in public debate, yet they could influence long-term outcomes. A country that can harness both talent and energy at scale will have a structural edge in sustaining AI growth.

Despite their rivalry, the US and China are converging in one important respect.Both are moving toward greater state involvement in AI. The US is expanding export controls, considering sanctions and increasing public investment. China is tightening regulatory oversight while continuing to direct resources toward strategic sectors.

At the same time, their approaches diverge in execution. The US relies on a decentralised innovation ecosystem, which fosters creativity but can lead to fragmentation. China’s centralised model enables coordination but risks stifling independent experimentation.

This combination of convergence and divergence makes the competition more unpredictable. Each system has strengths that could offset its weaknesses, and vulnerabilities that the other side may exploit.

What the future might hold

The immediate trajectory points toward deeper decoupling between the US and China. Separate technology stacks, restricted flows of data and talent and competing regulatory frameworks could define the next phase of the AI race. The Kratsios memo is a step in that direction, indicating a willingness to escalate beyond rhetoric.

However, complete separation is unlikely. The global nature of AI research and the interdependence of supply chains create strong incentives for continued, if limited, interaction. The challenge will be managing that interaction without triggering escalation. Ultimately, the US-China AI rivalry is entering a more mature and contested phase. It is no longer just about who builds the best models. It is about who controls the systems, resources and rules that will shape the future of artificial intelligence.