AI Nationalism & Geopolitics: The Chip Cold War

---

The US-China Rivalry

The central geopolitical dynamic of the AI era is the rivalry between the United States and China. The two countries are the world’s leading AI powers, and they are engaged in a competition that touches every aspect of AI — from the chips that power the models, to the data that trains them, to the talent that builds them, to the regulations that govern them.

The US approach has been characterised by “small yard, high fence” — a phrase popularised by Biden National Security Advisor Jake Sullivan (with Commerce Secretary Gina Raimondo as a prominent secondary exponent) to describe the strategy of maintaining the US lead in the most critical technologies (the “small yard”) while imposing strict restrictions (the “high fence”) on the transfer of those technologies to China. The export controls on AI chips, described in B08 — The AI Chip Wars, are the most visible manifestation of this approach. The CHIPS Act and the Stargate Project are the domestic investment side.

The Chinese approach has been characterised by self-reliance. China has invested heavily in its domestic chip industry (SMIC, Huawei), its domestic AI models (Baidu, Alibaba, Tencent, ByteDance, DeepSeek, and a newer generation including Zhipu AI, Moonshot/Kimi, MiniMax, and Baichuan), and its domestic AI infrastructure. The Chinese government’s “New Generation Artificial Intelligence Development Plan,” published by the State Council on July 20, 2017, set a three-step target (2020 / 2025 / 2030) with the 2030 goal of making China the world AI leader. The DeepSeek-R1 release in January 2025 was the most dramatic demonstration that China was closing the gap.

The rivalry is not just about technology — it is about values, governance, and the future of the international order. The US argues that AI should be developed in open, democratic societies, with transparency and accountability. China argues that AI should be developed in a way that serves the state, with social stability and national security as the primary concerns. The EU, with its AI Act, represents a third approach — one that prioritises fundamental rights and regulatory oversight over both innovation and state control.

---

The EU’s Regulatory Power

The European Union has positioned itself as the regulatory superpower of the AI era. The EU AI Act — adopted by the European Parliament on March 13, 2024, published in the EU Official Journal on July 12, 2024, and entered into force on August 1, 2024 — is the world’s first binding horizontal AI regulation, and it has the potential to set global standards via the Brussels effect (the phenomenon, coined by Columbia Law School professor Anu Bradford in 2012, whereby EU regulations effectively become global standards by virtue of the EU’s market size) — in the same way that the GDPR set global standards for data protection.

The EU’s approach is not primarily about building AI — Europe does not have a frontier AI lab comparable to OpenAI, Google, or DeepMind. It is about governing AI — ensuring that AI systems are safe, transparent, and respectful of fundamental rights. The EU’s regulatory power gives it influence that exceeds its industrial capacity, and it has made the EU the third pole in the AI geopolitical landscape.

The EU’s approach has been criticised by both sides. The US technology industry argues that the EU’s regulations are too restrictive and will stifle innovation, particularly in Europe. The Chinese government argues that the EU’s regulations are a form of protectionism — a way to favour European companies over foreign competitors. And some European commentators worry that the EU is regulating AI without building AI, which could leave Europe dependent on foreign technology.

---

India, Japan, and the UK

Several other countries are pursuing their own AI strategies.

India has launched the India AI Mission — a ₹10,372 crore (~US$1.25 billion) investment programme approved by the Union Cabinet, spanning five years, covering AI infrastructure, research, and skills. India’s strategy emphasises the development of AI for public goods — healthcare, education, agriculture — and the use of AI to accelerate economic development. India is also positioning itself as a “Global South” AI leader, advocating for inclusive AI governance and hosting the India AI Impact Summit at Bharat Mandapam, New Delhi, February 16–20, 2026.

Japan has been a significant AI player since the Fifth Generation Computer Systems project — a 10-year MITI initiative launched in 1982 (the main project phase ran 1982–1992; the coordinating body ICOT was formally dissolved in 1995). Today, Japan’s AI strategy emphasises the use of AI to address its demographic challenges — particularly its ageing population and shrinking workforce — and the development of AI in robotics and manufacturing. Japan established its own AI Safety Institute (J-AISI) on February 14, 2024, hosted at the Information-technology Promotion Agency (IPA) under METI (Ministry of Economy, Trade and Industry) — the third national AISI after the UK and US.

The UK has positioned itself as a leader in AI safety, hosting the AI Safety Summit at Bletchley Park on November 1–2, 2023, and establishing the UK AI Safety Institute shortly after. On February 14, 2025, the UK AISI was officially renamed the AI Security Institute, with the name change emphasising national-security risks. The UK’s strategy emphasises the safety and governance aspects of AI, leveraging its strong academic research base and its diplomatic influence. The UK has also been active in international AI governance, co-hosting the AI Seoul Summit (virtual leaders’ session May 21, 2024; in-person ministerial May 22) with South Korea, which produced the “Frontier AI Safety Commitments” signed by 16 AI companies.

---

Sovereign AI Infrastructure

One of the most significant developments in AI nationalism is the push for sovereign AI infrastructure — the idea that countries should build their own AI computing infrastructure, rather than relying on cloud services provided by foreign companies.

The push for sovereign AI is driven by several concerns. First, there is the security concern — if a country’s AI systems run on foreign cloud infrastructure, the foreign government could potentially access or disrupt those systems. Second, there is the economic concern — if a country relies on foreign cloud providers, the economic value of AI (the compute revenue, the data processing, the model training) flows to the foreign provider. Third, there is the sovereignty concern — if a country’s AI capabilities depend on foreign infrastructure, the country is not truly sovereign in the AI era.

Several countries are investing in sovereign AI infrastructure. France announced €109 billion (~$112.5 billion) in private-sector AI investment at the Paris AI Action Summit on February 10–11, 2025, with an emphasis on domestic infrastructure. India is building AI computing infrastructure through the India AI Mission. The UAE, Saudi Arabia, and other Gulf states are investing heavily in AI infrastructure, positioning themselves as AI hubs for the Middle East and beyond.

---

What It Means

AI nationalism is, in some respects, a natural response to the strategic importance of AI. AI is becoming the most important technology in the world, and countries that do not develop their own AI capabilities will be dependent on those that do. This dependence is not just economic — it is also political, because the countries that control AI will have significant influence over the countries that do not.

But AI nationalism also carries risks. The fragmentation of the AI industry into national blocs — with separate chips, separate models, separate standards, and separate regulations — could slow the pace of innovation, increase costs, and reduce the global benefits of AI. It could also intensify geopolitical tensions, as countries compete for control of critical resources and technologies.

The challenge for the international community is to find a balance between national interests and global cooperation. Some level of competition is healthy — it drives innovation and prevents monopoly. But too much competition — particularly competition that fragments the technology, restricts the flow of knowledge, and prioritises national advantage over global benefit — could undermine the very progress that AI promises.

The AI era will be shaped by the tension between nationalism and cooperation. The countries that find the right balance — that compete where competition is productive and cooperate where cooperation is necessary — will be the ones that benefit most from AI. The countries that get the balance wrong — that compete where they should cooperate, or that cooperate where they should compete — will be left behind.

---

---

---

Was ai nationalism inevitable — the product of forces too large to redirect — or was it a series of choices, each of which could have gone differently? The answer matters, because it determines whether the future is something that happens to us or something we make.