Ready, set, go! The AI arms race is on its way.
Written by Marc Bouwman. Edited by Koen van der Pool
In the history of mankind, mastering transformative technologies has always been the key to power. Agriculture led to permanent settlements, immense population growth and the rise of civilizations. Gunpowder gave unprecedented military dominance to civilizations to rapidly conquer land overseas. The printing press made knowledge widely accessible and allowed the spreading of ideas, sparking cultural, religious and scientific revolutions. Whenever a new technology proliferated, it spread out exponentially, becoming more widely accessible and cheaper to the public and benefitting those who were early adopters the most.
Artificial intelligence is no exception. AI is already past the ‘hype’, proving to be just as – if not more - transformative and disruptive as its technological predecessors. (Suleyman and Bhaskar, 2023) AI isn’t just a digital technology like the cloud or the blockchain. It has the potential to be the first time in history that humans create something that is more intelligent than themselves. This is a revolutionary technology of which we cannot even comprehend the long-term outcomes and ripple effects. As of now, Al can already detect early-stage cancer (Pesheva, 2024), produce deepfakes to influence political campaigns and operate drones to hit targets at pinpoint precision (Kirichenko, 2024).
“It has the potential to be the first time in history that humans create something that is more intelligent than themselves.”
Therefore, nations understand that those who are leading the race to master AI, have the power to become and remain a global giant. The AI arms race is influencing geopolitical power structures, creating new leaders, alliances, and rivalries. This intensifying competition has resulted in a race in which a winner’s podium most probably will be shared between the United States, China and Europe. Billions of dollars, yens and euros are being poured into tech companies to foster innovation in AI. But what are the consequences of an arms race, and will it benefit the greater good?
What does history tell us?
Who wants to predict the future must study the history they say. To understand the emerging AI arms race, we look back at one of the most significant technological arms races: the nuclear arms race. This race started between the US and Germany during World War II. If you happen to have seen the movie “Oppenheimer”, the story will sound familiar. Nuclear fission was discovered in the 1930’s, after which theories followed on the possibility of an atomic bomb. A technological weapon so powerful, it would end wars instantly and make a nation so terrifyingly strong that no other nation would even dare to start a fight.
As there were fears during World War II of a German atomic bomb project, the US responded by putting hundreds of thousands of people to work on the ‘Manhattan Project’ costing billions. (Metcalfe, 2023) The first ever nuclear detonation took place in the desert of New Mexico on July 16, 1945. A true eureka moment for nuclear physicists. It then took the US military just 21 days to bomb the cities of Hiroshima and Nagasaki, killing almost 220.000 people. Japan surrendered, ending World War II.
Non-Proliferation Treaty (NPT) being signed in London in 1968.
Nations, terrified and intrigued, instantly understood the essence of nuclear weapons for maintaining geopolitical power and influence. Multiple countries started funding nuclear weapon programs, where the climax occurred between the US and the Soviet Union during the Cold War. Strategic alliances were formed between nations, like NATO and the Warsaw Pact.
It triggered regulations and treaties like the Non-Proliferation Treaty (NPT) with the objective to prevent the spread of nuclear weapons and promote the peaceful use of nuclear energy. (Freedman, 2025) Next to military proliferation, nuclear fission drove investments in science and technology, transforming economic structures, education and even cultural perspectives.
“Similarly, AI has the potential to be a foundational technology, like nuclear fission, that provides geopolitical power to those leading the race.”
As exemplified before, AI is already making its entrance in military capabilities, improving healthcare to prolong lives, and influencing political viewpoints. And we are just getting started.
AI arms race in full swing
Just two and half years ago, US-based startup OpenAI launched ChatGPT. It sparked global interest, setting the stage for competing startups, as well as regulatory bodies. The European Union introduced the AI Act a year later. This regulatory framework showed the union’s commitment to responsible use of the technology, ensuring it aligns with human rights. It received, and still does, receive criticism for hampering innovation on the continent (Davies, 2024). However, looking back at the treaties of the Cold War, I believe early-stage regulations are for the long-term good.
As a power move, after President Donald Trump’s inauguration, he announced Project Stargate, an initiative to invest 500 billion dollars in AI, aiming to maintain their technological supremacy. In comparison, the Manhattan Project, crucial for American victory, costed 2 billion at the time, which is equivalent nowadays to 27 billion dollars, a bit over 5% of Stargate. In response to Trump and the ‘anti-innovation’ criticism, Ursula von der Leyen pledged 200 billion euros to European AI innovation during the summit in Paris. (Von der Leyen, 2025)
In silence, China worked on their own projects. A startup named Deepseek launched their own model in January 2025, scoring similar to ChatGPT on several metric tests. The significant part of Deepseek is the fact that it was created with ‘only’ a few million, which is a fraction of the cost it took to create US models, something that American tech CEO’s like Musk and Zuckerberg previously said to be impossible. This sent a shockwave to US stock markets, plummeting the value of tech companies such as Nvidia and indicators like the S&P500. However, when looking at history, it’s not that big of a surprise. Remember: disruptive tech naturally spreads out exponentially, becoming cheaper and more widely accessible to the public. This is exactly what happened with the introduction of Deepseek.
Deepseek's launch in January 2025 shocked the tech sector.
Participating countries are increasingly aware of the importance of keeping AI intel within borders in order to hold an advantage. Since the Biden administration, visa denials for Chinese scientists and technologists have spiked. Recent news reports surfaced stating China is implementing travel bans on their AI experts traveling to the US. (Mendoza, 2025) Similarly, during the nuclear arms race, physicists Oppenheimer and Bohr, were restricted from speaking to former university colleagues like German physicist Heisenberg, who eventually became head of the Nazi nuclear weapon project. Later, during the Cold War, espionage concerns led to heavily restricted travel between Americans and Soviets. All with the sole purpose of keeping technological know-how internal.
Lessons for the future
Learning from past technological arms races and understanding the events in AI that are rapidly following each other, we can make careful predictions for the future of how geopolitics will be reformed. In my opinion, the following three are the most prominent.
Firstly, we will see the emergence of strategic alliances and rivalries. It’s safe to say that relationships between the US, China and Europe aren’t particularly amazing. As history has proven, the rivalry could escalate to a potential ‘Cold Tech War’ where certain tech companies are not welcome to operate in other regions. Similarly to post-World War II power dynamics, where alliances were shaped around military nuclear technology, future coalitions could be defined by AI capabilities and shared goals. Collaborative research and development could lead to pacts, as we have seen with NATO, for example.
Secondly, the AI arms race in military warfare will accelerate. Less humans in the field will be required, but a higher percentage of GDP will be needed to invest in technology. The Manhattan Project is a fine historic example where a mass weapon is far more destructive than a large army. We will probably see autonomous drones, cyber attacks and AI-driven intelligence allowing states and/or nations that invest in these to become more dominant. This could lead to conflicts emerging quicker and, similar to Japan’s surrender, ending sooner. Thereby, just like post-World War II, a few dominant military nations will set the rules on a global stage.
Thirdly, AI has the potential to become more intelligent than humans. This state is called artificial general intelligence (AGI), and it’s a goal most proprietary AI companies worldwide have. (Venkatesh, 2025) The peak of the Cold War led to a stalemate, in which both parties understood the power of their nuclear weapons would wipe the other out. This led to negotiations and treaties like the Non-Proliferation Treaty regarding the reduction of strategic arms. Now think about AI as future technology that’s smarter than us, therefore having the potential to do things we humans cannot comprehend. This powerful yet daunting stage could cause a similar stalemate, leading to an agreed reduction in AI advancements by the countries possessing this technology.
“Responsible innovation is key, where progress isn’t driven by competition but by the collective goal of benefiting humanity.”
In an ideal world, we should see a balanced approach towards AI development. An AI arms race should make place for cooperation, transparency and accountability. Responsible nnovation is key, where progress isn’t driven by competition but by the collective goal of benefiting humanity. Right now, we are facing an uncertain future in which AI can become a tool for both progress and/or destruction. As with nuclear fission, mankind used it for progressive breakthroughs benefiting economies, but more notably, destructive purposes devastating the lives of many. Yet again, it’s up to us to choose which.
Suleyman, M. and Bhaskar, M. (2023). The Coming Wave.
Pesheva, E. (2024). New AI tool can diagnose cancer, guide treatment, predict patient survival. HMS Communications.
Kirichenko, D. (2024). The Rush for AI-Enabled Drones on Ukrainian Battlefields. Lawfare.
Metcalfe, T. (2023). What Was the Manhattan Project? Scientific American.
Freedman, L. D. (2025) Treaty on the Non-Proliferation of Nuclear Weapons. Britannica.
Davies, P. (2024). EU AI Act reaction: Tech experts say the world's first AI law is 'historic' but 'bittersweet'. Euronews.
Von der Leyen, U. (2025). EU launches InvestAI initiative to mobilise 200 billion euros of investment in artificial intelligence. European Commission.
Mendoza, C. (2025). China Blocks Its AI Experts from U.S. Travel, Escalating Tech Cold War. Medium.
Venkatesh, C.R. (2025). Understanding Artificial General Intelligence (AGI): A Complete Overview. Dotcom Infoway.