THE ALGORITHM OF GLOBAL LEADERSHIP

In the context of the ongoing strategic competition between the US and China, the development of sophisticated artificial intelligence (AI) is becoming a key priority. AI has become an integral part of national security at all levels, from the economy to military capability. In the future, this link will only become stronger, to the point where they will be inseparable.

It is evident that the United States leads in this field, having developed advanced and large-scale models for processing big data, as well as innovative chip developments. However, this advantage is not without its limitations, and China is making significant inroads into the American market. In light of Donald Trump's visit to Beijing, there has been a resurgence of calls for Washington to put aside its differences and adopt a unified national strategy in the field of AI. Otherwise, the US will not be able to confidently surpass China in every way.

Beijing is unwilling to yield. Since 2017, it has had a comprehensive AI development plan, and it aims to become the world's dominant technological power by 2049. Significant resources are being allocated to these endeavours, including the development of AI technologies that are more efficient and cost-effective than their American counterparts, as well as the pursuit of a leading position in open-source AI, cloud infrastructure and global data ecosystems.

In the Communist Party's latest five-year plan, AI is identified as a key enabler for strengthening the country's army and economy. It is evident that there is a genuine battle underway between Washington and Beijing for control over global digital infrastructure, both in the present and with regard to the future.

Legislative uncertainty

In light of this, American experts have highlighted that the US approach to AI is fragmented and occasionally contradictory. Legislation in this sector is inconsistent across states, resulting in ambiguity for businesses and the emergence of a small number of dominant companies. It is estimated that in 2026 alone, state legislatures introduced more than 1,700 bills related to artificial intelligence. In light of these considerations, there is a growing impetus to devise a cohesive national action plan that would encompass measures to develop energy infrastructure, streamline the permitting process for data centres, enhance cybersecurity, boost local production, and facilitate workforce training.

For instance, several experts have expressed concerns that advanced AI models (such as OpenAI's ChatGPT and Google's Gemini) are overloaded with security requirements and checks, as concerns about deepfakes and worker displacement are constantly voiced. It is evident that these concerns are not unfounded. However, they carry a significant risk of slowing innovation, experimentation and research, cutting-edge solutions, open-source development and market entry for small companies. In contrast to the tech giants, these companies do not have extensive legal departments, and it is more cost-effective to invest in business development and technological innovation.

Consequently, the administrative barriers that hinder American companies' presence in emerging markets, along with the restrictions on their capacity to contribute to the host country's development, could potentially pave the way for China to dominate the digital infrastructure of numerous countries worldwide. It is inevitable that these countries will soon realise that US computing power in AI is not at all irreplaceable.

In this sense, China is successfully following its strategy. For instance, it employs open-source AI, whereas the US is moving towards closed, tightly controlled systems. Conversely, if intermediary countries begin to rely on Chinese cloud service providers, this could provide China with unprecedented access to global data for training AI. Should China gain a significant advantage in data volume, it is highly likely that the United States will lose its leading position in the field of AI. It is important to note, however, that US closed models (from OpenAI or Google) continue to dominate commercial corporate deployments.

There is also a growing view that China already leads in developing the most efficient open-weight artificial intelligence models, which raises concerns about privacy, censorship and America's ability to shape the future of global AI. Whereas 'closed' models are only available online via websites or application programming interfaces, 'open' models are released as a set of weights that users can download and run on their own computers.

Electric power as a crucial component

Another significant vulnerability in the American AI chain relates to the underlying infrastructure, particularly in the energy sector. The substantial energy demands of artificial intelligence highlight the need for enhanced energy efficiency and security measures. According to recent estimates, over the next decade, total US demand for electricity from data centres, industrial electrification and electric vehicles could increase overall consumption by more than 50% compared with today's level.

The operation of various AI-related and 'green' economy facilities, as well as industrial growth, requires a power system with sufficient capacity and redundancy to ensure fault tolerance. Electricity has never played such a pivotal role in shaping the future of the United States as it does today. Data centres currently account for 7% of US electricity consumption, and this share continues to grow. It is estimated that the US will need to increase its electric power capacity by at least 50 gigawatts over the next few years in order to support the development of AI. Other estimates suggest that by 2030, AI's electricity needs alone could exceed the net generating capacity of the US power grid.

At the same time, consumers are increasingly concerned about rising electricity prices and the ageing grid. A substantial proportion of US grid infrastructure is over 25 years old, and major modernisation work will take at least ten years. It is evident from the available data that a significant proportion of distribution transformers are approaching the end of their service life. This indicates that they are likely to fail in the near future. Extreme weather conditions can also contribute to this, which has recently become a frequent occurrence worldwide. It is evident that if this issue is not resolved in the US, American companies will begin to relocate their operations overseas, which will also place them in a vulnerable position.

It is worth noting that the issue could be resolved by implementing small modular nuclear reactors, which require a relatively small amount of land and less water. It is recommended that they be situated in close proximity to data centres and industrial facilities in order to ensure a reliable and consistent power supply. For instance, Virginia is home to 25% of American data centres (more than 100 of the 504 known worldwide) and 32% of the state's electricity is already generated by nuclear power.

In the American narrative, the competition with China in AI is already being presented as a struggle for democratic values. It is generally accepted that these values should determine the shape of future technologies. However, it is evident that the issue is quite distinct. The entity that establishes its AI systems as the global standard will gain significant influence and leverage over other parties, necessitating a universal adherence to its rules.

It is evident that AI could emerge as a significant theme in the 2028 US presidential race, driven by widespread public concerns regarding the technology. According to a recent survey, many Americans are concerned about the impact of artificial intelligence on the economy and the labour market. A majority of Americans (71%) believe the pace of AI development is too fast. A recent survey indicates that nearly two-thirds (64%) of Americans believe that the likelihood of AI bringing economic benefits to everyone is small or very small.

While younger Americans express less concern than their older counterparts about the pace of AI development and the distribution of economic benefits, they are more likely to worry that AI will replace jobs on which they and their families depend.

The outcome of the artificial intelligence race will therefore be pivotal in determining the technological standard, the architecture of global digital security, economic rules and geopolitical influence for decades to come. Ultimately, the ability to create a scalable, reliable and widely accessible AI ecosystem will be key to setting the rules of the game for the entire world order.

Conversely, the AI race may evolve into a multifaceted competition where the US and China excel in distinct areas of the technology stack. This could even contribute to global safety, including by further strengthening the interdependence between Beijing and Washington.