DeepSeek’s Disruption: Geopolitics and the Battle for AI Supremacy

The rise of DeepSeek signals a profound shift in the global AI landscape, challenging the foundations of US technological dominance. Until now, Washington’s AI strategy hinged on controlling access to high-performance computing and advanced semiconductors, enforcing export controls to constrain China’s innovation. DeepSeek’s ability to develop a powerful AI model with significantly lower computational costs upends this paradigm. The game has changed, and a new phase in the race for AI supremacy has begun.

The US Bet on Export Controls

Washington’s AI containment strategy relied on restricting China’s access to advanced semiconductor technologies, assuming that US tech firms could outpace Chinese competitors while maintaining a technological edge. This approach carried an inherent risk, in that it forced China to innovate with what it had. What few anticipated was the speed at which this would happen.

Founded by Liang Wenfeng, a hedge fund investor and AI researcher with ties to Chinese state-backed initiatives, DeepSeek has emerged as a formidable competitor to Western firms. Its key advantage lies in minimising computational requirements, sidestepping many of the constraints imposed by US restrictions. Unlike traditional AI models that rely on brute-force computation, DeepSeek employs a ‘mixture of experts’ approach, activating only the necessary computing resources for each task.

But could DeepSeek’s breakthrough also be a product of intellectual property theft? Reports suggest it may have trained its models on OpenAI data using a technique know as distilling. Given China’s history of data acquisition practices, leveraging such methods would align with its strategic goals.

Challenging US AI Dominance

For decades, US AI leadership rested on research excellence, corporate investment, and semiconductor superiority. Companies like OpenAI, Google DeepMind, and Microsoft set the pace, while NVIDIA supplied the high-performance chips necessary for AI training. DeepSeek challenges this dominance in three fundamental ways.

Firstly, it decouples AI from the concept of chip supremacy. DeepSeek demonstrates that cutting-edge AI is no longer exclusively dependent on massive computational resources and advanced chips. Secondly, it redefines cost structures by offering AI at a fraction of Western subscription costs. DeepSeek could disrupt the US AI business model and shift market expectations toward low-cost alternatives. Finally, by innovating around US trade restrictions, it erodes US leverage, showing that trade policies alone will be insufficient to contain China’s AI advancements.

AI as a Geopolitical Tool

AI is an instrument of soft power, and DeepSeek strengthens Beijing’s position. If China can offer AI-driven solutions at lower costs than its Western counterparts, it will become the preferred partner for emerging economies. DeepSeek’s accessibility makes it attractive to countries in Africa, Latin America and Southeast Asia, where high AI costs have hindered adoption.

China’s integration of AI into global infrastructure projects, particularly through the Belt and Road Initiative, could further extend its influence. If DeepSeek becomes embedded in key sectors such as finance, logistics and governance, China will shape AI standards and protocols in ways that challenge the US-led digital order.

Economic and Security Implications

DeepSeek’s rise underscores the deepening intersection of AI, economic influence and cyber security. AI-driven automation is transforming financial markets, supply chains and trade networks. If Chinese AI models dominate these areas, Beijing could strengthen its economic intelligence capabilities, gaining leverage in trade negotiations, resource management and financial forecasting.

The growing reliance on AI in economic decision-making also introduces security vulnerabilities. AI-driven financial systems and supply chains could be manipulated, exposing businesses and governments to new cyber threats. Additionally, as cost-effective AI models gain traction, supply chains may shift away from US-designed semiconductor technology. This could weaken Washington’s strategic influence over AI hardware production, necessitating a reassessment of semiconductor policies and investment strategies.

Strategic Responses for the US and its Allies

To maintain AI leadership, the US and its allies must rethink their approach. A primary focus should be on prioritising algorithmic efficiency by investing in AI architectures that emphasise performance optimisation rather than sheer computational power. DeepSeek has demonstrated that AI advancement does not solely depend on hardware superiority, and US firms are already analysing its methodology to develop their own adaptations.

Additionally, the US must reassess its semiconductor policy. Export controls alone are proving insufficient to curb China’s progress. Expanding restrictions to include lower-end chipsets – previously overlooked in AI regulation – may become necessary. However, given China’s strategic focus on these components, enforcing such controls will be a complex challenge.

Beyond hardware, safeguarding economic infrastructure is crucial. AI-driven risk assessments much be integrated into financial and supply chain strategies to mitigate vulnerabilities that could be exploited by adversaries. Given AI’s growing role in economic intelligence, Washington and its allies must ensure that key industries remain resilient to AI-enabled manipulation.

Finally, the US and its allies must develop a robust AI diplomacy strategy to counter China’s global outreach. DeepSeek’s success exemplifies the role that AI can play as a geopolitical instrument, particularly in the Global South. To prevent Beijing from dominating AI infrastructure and influence, Washington must offer competitive AI partnerships that present viable alternatives to Chinese technology. Expanding AI-driven development initiatives in emerging economies will be key to ensuring a more balanced global AI ecosystem.

Jevons’ Paradox and the New AI Paradigm

The AI race is entering a new phase, where efficiency rivals raw computational power. Microsoft CEO Satya Nadella has pointed out that DeepSeek exemplifies Jevons’ Paradox – a principle from industrial economics stating that technological efficiency often increases overall resource consumption rather than reducing it. By making AI more computationally efficient, DeepSeek is accelerating AI adoption worldwide and amplifying China’s influence in the process.

Some have called this a ‘Sputnik moment’ for AI – a wake-up call for the US and its partners to accelerate their innovation cycle. While the US retains an edge in fundamental AI research and technology, China has shown that breakthroughs in efficiency and accessibility can be equally disruptive. The AI contest is no longer a one-sided race. The battle for AI primacy was already well underway, but China just rewrote the rules of engagement.

© Tobias Feakin, 2025, published by RUSI with permission of the author

The views expressed in this Commentary are the author’s, and do not represent those of RUSI or any other institution.

For terms of use, see Website Ts&Cs of Use.

Have an idea for a Commentary you’d like to write for us? Send a short pitch to commentaries@rusi.org and we’ll get back to you if it fits into our research interests. Full guidelines for contributors can be found here.