Atlantic Council Strategic Insights Memo September 18, 2024 Assessing China’s AI development and forecasting its future tech priorities By Hanna Dohmen

 Atlantic Council 

Strategic Insights Memo

September 18, 2024

Assessing China’s AI development and forecasting its future tech priorities

By Hanna Dohmen

TO: Policymakers and technology policy strategists

FROM: Hanna Dohmen

DATE: September 18, 2024

SUBJECT: Assessing China’s current AI development and forecasting its future technology priorities

In July 2024, the Atlantic Council Global China Hub (AC GCH) and the Special Competitive Studies Project (SCSP) convened experts and policymakers in the second of a two-part private workshop series to gather insights into China’s technology priorities today and in the future. Participants discussed Beijing’s posture on artificial intelligence (AI) development and deployment today, including the hurdles China’s AI industry faces amid US-China technology competition, as well as Beijing’s policy priorities over the next decade. This memo summarizes insights gathered during the workshop.

Strategic context

In today’s strategic competition between the United States and China, both countries seek to bolster their nations’ innovation ecosystems and enhance their ability to develop and deploy breakthrough technologies. The United States is committed to maintaining US technological leadership in the long term, as Secretary of Commerce Gina Raimondo demonstrated at the Reagan National Defense Forum in December 2023, when she stated that “America leads the world in artificial intelligence. America leads the world in advanced semiconductor design, period . . . We’re a couple years ahead of China. No way are we going to let them catch up. We cannot let them catch up.”

China’s strategic focus has long been on “self-reliance and self-improvement (自立自强).” In fact, on June 24, 2024, Chinese President Xi Jinping delivered a speech at a major Chinese science and technology (S&T) conference in which he emphasized this longstanding ambition: “Since the 18th Party Congress [in 2012], the Party Central Committee has promoted the implementation of the innovation-driven development strategy in an in-depth way, proposed the strategic task of accelerating the construction of an innovation-oriented country (创新型国家), established the goal of building China into an S&T powerhouse by 2035, continuously deepened S&T structural reform (科技体制改革), fully stimulated the enthusiasm, initiative, and creativity of S&T personnel, and vigorously promoted the building of self-reliance (自立 自强) in S&T.”

To sustain its own growth in technology leadership, the United States has concentrated its efforts on computational power and AI. Thus far, a key pillar of the US strategy has been to slow China’s progress in developing advanced-node semiconductors, a critical input needed to power AI. US export controls on advanced compute, semiconductor manufacturing equipment, and supercomputing—as well as regulations that will prohibit and monitor US investments in Chinese AI, quantum computing, and semiconductor companies—are part of a broader strategy to maintain US leadership and slow China’s progress.

As each country advances its own agenda, the implications of this competition will continue to shape the future of technology development and geopolitics. Given the rapid advancements in AI, the current strategic environment is complex and fast evolving. As such, it is critical to not only assess the current state of technological competition, but to also look ahead at future technology priorities both in the United States and in China.

Benchmarking China’s AI progress and challenges

One of the key questions in this strategic competition is China’s position in AI development relative to that of the United States. However, workshop participants highlighted that this depends entirely on the lens through which one views competitiveness. Should competitiveness in AI be assessed by the size of models and processing speeds? Should it be about which ecosystem can leverage AI to deliver the most tangible economic benefits in terms of revenue growth and operational and efficiency improvements?

While participants agreed that these questions are critical when considering the long-term objectives of either country’s strategies, current assessments primarily focus on which models are the biggest and fastest. Here, views range from estimating that China’s AI model development is six to twenty-four months behind that of the United States. For example, in June 2024, Kai-Fu Lee, the chief executive officer of the Chinese AI startup 01.AI, claimed that the company is six to nine months behind US AI leaders but is catching up rapidly. Some experts, including Joe Tsai, Alibaba co-founder and chairman, suggest that China’s AI companies are “possibly two years behind” US AI companies.

Technical performance and metrics of Chinese models are an important measure of progress, but workshop participants emphasized how a more holistic view of the AI stack and broader innovation ecosystem is necessary to contextualize technological advancement. Three prevailing dynamics that will set the tone for future AI development in China emerged from this discussion: AI ecosystems, compute infrastructure, and regulatory landscapes.

More players, less capital

Workshop participants noted how China’s AI model development ecosystem differs significantly in scale and structure from that of the United States. In the United States, a small number of big players—such as OpenAI, Meta, Google, and Anthropic—dominate the field. These companies leverage their partnerships with hyperscalers and have access to the necessary compute needed to power their AI development and deployment. In contrast, China has a much larger number of AI companies developing models, which participants said is leading to a dilution of investment and compute resources. For example, as of August 2024, the Cyberspace Administration of China has approved a list of more than 180 large language models (LLMs) for general use, illustrating the broad swath of Chinese tech companies fighting for domestic market share.

Not only are these companies competing for a slice of the market, but they are also competing for funding amid an economic slowdown and a downturn in China’s VC industry. Participants stressed that while many Chinese startups have attracted investments from big tech companies, such as Alibaba and Tencent, many investors remain skeptical about AI startups’ abilities to generate revenue in the short term. In search of economically productive investments, many Chinese venture-capital firms are looking to diversify their risk by pooling resources, suggesting a more dispersed funding environment. Given both funding and hardware constraints on Chinese AI developers, participants suggested that China might succeed in advancing a few companies or AI labs by pooling resources, but these efforts will need to be selective and targeted, reducing the likelihood of substantial returns. Ultimately, participants suggested that this environment in China’s AI market is likely to lead to increased industry consolidation.

US export controls loom large

US export controls, and those of allied countries, are affecting China’s access to advanced computing resources, imposing significant constraints on both AI training and inference (i.e., AI model development and deployment). As China is unable to legally acquire leading-edge AI chips such as NVIDIA’s A100, it increasingly needs to rely on its own domestically designed and manufactured alternatives. Huawei’s Ascend 910B is China’s closest competitor, though reports suggest it lags in performance for training LLMs. While Chinese chip designers have made notable progress, China’s production of these chips is significantly constrained. These resource challenges make both AI training and inference for Chinese companies more expensive and less efficient. Participants suggested that these challenges could particularly hinder the deployment of AI models at scale in China.

Asserting control versus fostering innovation

Participants also discussed the impact of AI regulations, particularly China’s censorship standards, on AI development. Participants highlighted that there are two possibilities for how China’s strict regulations can impact AI innovation. On one hand, these regulations could hinder China’s ability to develop competitive AI models by imposing strict controls on the outputs of models. Conversely, it is also possible that if China navigates these challenges effectively, Chinese AI developers might gain valuable insights into how to make AI models safer. Participants believe the former is likely to be true, but this remains an open question.

China’s forward-looking technology priorities

In addition to understanding what China’s current AI development looks like, it is also important to consider the country’s strategic priorities for future technologies. The discussion highlighted that, looking ahead, AI will be one of the key elements to developing and advancing future technologies.

Future manufacturing: Looking toward the future, participants believe that China’s motivations for advancing AI predominantly center around enhancing industrial efficiency, particularly improvements in manufacturing and automation. In early 2024, seven Chinese ministries and government bodies, including the Chinese Ministry of Industry and Information Technology and the Ministry of Science and Technology, released a guidance document that identifies six “future industries” as priorities for China’s industrial policy. This document emphasizes that China should “seize the opportunities of a new round of S&T revolution and industrial transformation, focus on the main battlefield—namely the manufacturing industry—to accelerate the development of future industries, and support the advancement of new-style industrialization (新型工业化).”

In light of its current economic slowdown and demographic challenges, China emphasizes converting new technologies, including AI, into economically productive applications. Participants highlighted that there is less focus on LLMs for chatbots in China; instead, there is more focus on the industrial applications that LLMs can help advance and streamline. For example, the same guidance document suggests “utilizing artificial intelligence (AI), advanced computing, and other technologies to precisely identify and cultivate high-potential future industries.”

Robotics: China’s grim demographic outlook has also driven the country’s deployments of robots. The country’s working-age population is rapidly shrinking and its birth rates remain concerningly low. Some estimates suggest that China’s working-age population could decrease by an annual average of 0.83 percent between 2022 and 2035. In part to address the economic concerns prompted by these demographic shifts, China is focusing on increasing productivity through industrial robots. Chinese firms deployed nearly three hundred thousand industrial robots, while Japan and the United States deployed approximately fifty thousand and forty thousand robots, respectively. Indeed, China’s installation of industrial robots has increased by around 13 percent since 2017. The United States’ robot growth rate, however, pales in comparison at just 4 percent over the same period. Participants suggested that this will continue to be a significant priority for China in the coming years. Moreover, humanoid robots—machines with physical features and behaviors that resemble those of humans—are a key area of robotics that participants expect China to prioritize.

Biotechnology: Participants also highlighted China’s focus on biotechnology. Specifically, China is stressing the need for innovations in cell and gene technology, synthetic biology, and bioengineered breeding, as well as medical services empowered by technologies such as AI. This once again emphasizes China’s ambitions to utilize AI to advance other critical and emerging technologies. As a result, participants argued, the United States must put greater emphasis on its own biotechnology advancements. Biotechnology presents complex risks due to its diffuse applications and potential health benefits, making it a crucial strategic area for both the United States and China.

Fundamental research: Beijing is also redoubling its investments in fundamental research, recognizing systemic weaknesses in developing scientific and technological breakthroughs. In a June speech, Xi said, “although China’s scientific and technological undertakings have made significant progress, its original innovation capabilities are still relatively weak, with some key core technologies being controlled by others, and there being a shortage of top scientific talent. There is an urgent need to further enhance the sense of urgency, intensify efforts in scientific and technological innovation, and seize the strategic heights of technological competition and future development.” In March, Beijing announced that it was raising national research and development spending by about 10 percent, signaling how fundamental research will be a rising priority amid geopolitical tensions.

Participants argued that while scientific collaboration with China holds significant potential benefits, it is essential to navigate it carefully to avoid contributing to military applications. The challenge lies in balancing collaboration with security concerns, particularly in areas prone to dual-use technology risks. China boasts strong scientific capabilities and some of the world’s leading scientists, which underscores the importance of engaging in strategic research partnerships while safeguarding against potential military exploitation.

Conclusion

AI serves as the central thread linking China’s strategic focus across various emerging technologies, including advanced manufacturing, robotics, biotechnology, and many more. China’s broader future technology priorities reflect a comprehensive approach to leveraging AI in diverse fields, from smart manufacturing and quantum computing to biotechnology and space exploration. The country’s heavy investment in these areas demonstrates its commitment to achieving technological sovereignty and economic resilience. For the United States, this highlights the need for continued investment in AI and biotechnology, as well as careful management of international research collaborations to protect national security interests and maintain US technological leadership. Ultimately, the evolving technological landscape underscores the importance of AI as a key driver of technological progress and competition on the global stage.

About the author

Fellow

Hanna Dohmen

Nonresident Fellow

Global China Hub

China

Chinese

Acknowledgements

This strategic insights memo was written and prepared with the support of the Atlantic Council’s Global China Hub and the Special Competitive Studies Project.

The Special Competitive Studies Project (SCSP) is a nonpartisan, nonprofit initiative with a clear mission: to make recommendations to strengthen America’s long-term competitiveness as artificial intelligence (AI) and other emerging technologies are reshaping our national security, economy, and society.

Global China Hub

The Global China Hub researches and devises allied solutions to the global challenges posed by China’s rise, leveraging and amplifying the Atlantic Council’s work on China across its fifteen other programs and centers.