Tech companies from the US and China have built a substantial lead in creating and commercialising large language models (LLMs), or GenAI. The US has had a pronounced head start in the GenAI race, building on decades of AI leadership, but the GenAI power map could change.

A new report, “How CEOs Can Navigate the New Geopolitics of GenAI,” from the BCG Henderson Institute and BCG’s Centre for Geopolitics, looks at the current state of play and the implications for both corporate leaders and policymakers.

Lang continued to say: “For corporate leaders who are integrating GenAI into their business operations, relying solely on GenAI supplied by companies in the US or China could pose serious risks due to the possibility of regulations, data requirements and availability of the models all being vulnerable to shifts in government policy.”

Two AI Superpowers…For Now

The two AI superpowers, the US and China, are currently the only players with robust access to, and control over, sizeable portions of the entire GenAI value chain. They produce the most intellectual property (IP), and they have the largest AI talent pools. They have some of the richest data ecosystems in the world and the most data centre infrastructure capacity, as well as access to capital.

Two Chinese companies, Alibaba and the GenAI startup 01.AI, contribute over one-quarter of the world’s top open-source LLMs. Established tech giants Baidu and Tencent have also released high-performing models, as have a new generation of GenAI startups, the so-called AI tigers. Top Chinese LLMs have substantially reduced the gap compared to state-of-the-art alternatives in the last year.

Who Could Emerge as a New GenAI Player?

This US–China story has fuelled a ‘two superpowers’” narrative, but momentum is building in other parts of the world. Other countries could catch up. The European Union can position itself as a significant GenAI ‘middle power’ by combining the strengths of its member states. The EU is already home to a nascent GenAI startup ecosystem and has the second-largest global AI specialist talent pool.

Saudi Arabia and the United Arab Emirates (UAE) can rely on substantial sovereign wealth funds and focused investments to rapidly grow their AI talent bench, expand data centre infrastructure, and develop high-performing LLMs.

Japan and South Korea can build on their established tech ecosystems, significant R&D investments, and expertise in hardware to position themselves as GenAI middle powers.

Canada and the UK are home to a sizable share of the world’s top AI scholars, have produced many notable AI models, and have published some of the most important AI research papers. Israel, despite its smaller population, has a high concentration of AI specialists and has developed competitive models. For these countries, opportunity lies in accelerating research breakthroughs.

A Call to Action for Leaders

If all options for GenAI originate in just two countries, it could lead to disruptions in GenAI’s availability due to geopolitical shocks. Corporate leaders will need to build the “geopolitical muscle” required to sense coming shifts and adapt their operating models.

AI sovereignty may soon become a critical source of national security, economic value, and soft power. Governments of countries or in regions with potential to become GenAI middle powers should consider what it would take to claim the space.

“This race is far from over,” said Nikolaus Lang. “Both business and political leaders will need to navigate a dynamic and high-stakes environment, shaped by regulation, policy and technology. Whether the middle powers—and which ones—will establish a claim to the competitive supply of GenAI at a global scale is one of the biggest unknowns. How this plays out has the potential to rewrite the rules of the game.”