“[DeepSeek] is a wake-up call to the American AI industry… The last administration sat on their hands and allowed China to rapidly develop this AI program” ~ Karoline Leavitt, White House Press Secretary
On January 27th 2025, China’s DeepSeek shocked the world and caused a sudden plunge in the U.S. stock market. While DeepSeek’s impact has been subdued, its implications are still being debated. Louise Marie Hurel, a researcher at the Royal United Services Institute’s (RUSI) Cyber and Tech Programme, recently wrote a report that elaborates how DeepSeek has reshaped the global AI industry.
In the beginning, the author pointed out that DeepSeek’s R1 model highlights several major developments in the global AI race. First, a cost-effective AI model running with legacy chips. Second, China’s tech resilience in bypassing U.S. sanctions on advanced chips. Third, the rise of an open AI model that challenges the existing proprietary mode. Moreover, DeepSeek has reignited national security concerns with regard to America’s technological dependency on China.
How have Chinese companies adapted to the DeepSeek moment? Undoubtedly, DeepSeek has given China a renewed national pride and technical confidence in China’s superiority in the global AI battleground. The diffusion of DeepSeek’s model soon accelerated as numerous city governments and state-owned enterprises in energy and telecommunications declared incorporation of the model into their information systems. Other companies are facing government pressure to follow suit. Sitting at the epicenter of the AI shockwave, DeepSeek itself is working on the next R2 model, which is scheduled to be released later this year.
How have Chinese companies responded to the post-DeepSeek moment? DeepSeek has emboldened some Chinese AI companies to copy its success story and re-create their own “DeepSeek-like moment”. For example, in March 2025 Butterfly Effect, a local startup, released Manus AI, claimed it to be “the world’s first general AI agent”, only to receive mixed reviews from media and users. On the other hand, China’s big tech companies, such as Alibaba, Tencent, Huawei, and BYD, have sought to integrate DeepSeek’s open model into their existing products. This national sentiment has also emboldened Huawei to move forward on its Ascend chips and further improve the manufacturing processes.
How have American companies reacted to the DeepSeek moment? OpenAI first expressed apprehension that DeepSeek misused the company’s proprietary model for training, and thus was suspected of infringing on OpenAI’s intellectual property rights. Some American tech giants, like Meta, played down DeepSeek’s low-cost efficiency. Other tech giants, such as Microsoft and Amazon, chose to incorporate DeepSeek’s R1 model into their AI platforms and cloud services, allowing app developers to harness a library of AI models in designing and testing. At the same time, Apple seemed to benefit from the DeepSeek shock as the iPhone family could serve as a platform to distribute smaller AI models.
What are DeepSeek’s implications for the Indo-Pacific region? With ambiguity and pragmatism, many countries in the region have adopted a hedging strategy between the U.S. and China. Singapore and Malaysia have collaborated with OpenAI and Alibaba in developing local AI capacities, and worked with Nvidia to establish domestic AI infrastructure. Meanwhile, the Indo-Pacific companies can partner with both American and Chinese companies while use DeepSeek’s R1 model as a benchmark to conduct testing and optimization at smaller scales. Furthermore, companies in the region can leverage DeepSeek’s open weights model to trigger more innovation in other open models or AI applications, such as local large language models (LLMs) or Big Tech LLM architectures.
In conclusion, the author acknowledges the fact that China is so far the largest beneficiary of the DeepSeek moment. China has cleverly used DeepSeek to galvanize the entire local AI supply chain, accelerate diffusion of the R1 model, and incentivize local semiconductor companies to make advancement in technology. However, China’s hubris has caused undue obsession with fast delivery rather than quality, as well as inflexible responses to America’s export controls and import tariffs. On the American side, economic coercion has eroded Silicon Valley entrepreneurship and spooked venture capitalists. Moreover, relying on sanctions as the sole means to check China’s innovation progress may backfire, making China a more competitive rival.
Clearly, both American and Chinese AI companies are asked to make technological breakthroughs in a highly VUCA (volatile, uncertain, complex, and ambiguous) environment, which is rife with trade wars, protectionist policies, and shifting supply chains. As the author stresses, multinational tech companies need to balance opportunities against risks and engage with both America and China’s AI ecosystems, so as to promote homegrown AI innovation and play crucial roles in shaping AI’s next trajectory of development.
The article is based on the Royal United Services Institute (RUSI)’s “DeepSeek: What it Means for Chinese and US Companies’ Strategies in the AI Race”, published by Fletcher Security Review and authored by Louise Marie Hurel, a researcher at the RUSI Cyber and Tech Programme. Read the full report