The global arms race powered by Artificial Intelligence has already begun — and it’s accelerating
Artificial Intelligence has arrived on the battlefield, and the world will never be the same.
In September 2025, during a military parade in Beijing, President Xi Jinping welcomed two very special guests: Vladimir Putin of Russia and Kim Jong-un of North Korea. The three watched, side by side, a demonstration that captured the attention of the entire planet — Chinese autonomous drones flying in formation alongside fighter jets, with no human pilot at the controls. The message was clear, direct, and calculated to be seen by everyone, especially the rivals across the Pacific.
On the other side of the ocean, the Pentagon went to high alert. American defense and intelligence officials concluded that the United States program for unmanned combat drones was falling behind China’s. And Russia also appeared to be ahead in building facilities capable of producing advanced drones at scale. The response came quickly. Anduril, a California-based defense startup founded by Palmer Luckey, began manufacturing its own AI-powered autonomous drones at a factory on the outskirts of Columbus, Ohio — three months ahead of schedule — as part of America’s effort to close the gap with China.
But that’s just the tip of the iceberg. What’s happening goes far beyond a rivalry between two superpowers. We’re in the middle of a global arms race powered by Artificial Intelligence, with countries investing billions in systems designed to operate independently — identifying, tracking, and striking targets without any need for human intervention. And the questions about who controls these machines, and under what rules, are still far from being answered. 🤖⚔️
What are AI-powered autonomous weapons and why they change everything
When we talk about autonomous weapons, we’re not talking about science fiction or robots from the distant future. We’re talking about systems that already exist, have already been tested, and in some cases have already been used in real conflicts. An AI-powered autonomous drone can take off, navigate complex routes, identify a specific target based on programmed parameters, and execute an action — all without any human being needing to press a single button.
This capability, which seemed far off ten years ago, is now a reality that multiple countries are racing to master and scale. And the examples go beyond drones. The current competition spans unmanned fighter jets that coordinate attacks at speeds and altitudes few human pilots can match, as well as central AI-managed systems that analyze intelligence and recommend strike targets in a matter of seconds.
The fundamental difference between a conventional drone and an autonomous drone lies in decision-making. In traditional models, a human operator sits at a base, sometimes on the other side of the world, and controls the equipment in real time. In AI-powered systems, the drone itself processes environmental information, recognizes visual or behavioral patterns, evaluates context, and acts independently. This means it can operate in environments where communication with human operators is blocked, disrupted, or simply unfeasible — a massive tactical advantage in electronic warfare scenarios.
The technology, as American officials have pointed out, doesn’t need to pause, eat, drink, or sleep. It promises to transform warfare by making combat faster and more unpredictable than anything we’ve ever seen.
And here lies one of the most sensitive points of this entire discussion. When a machine makes the decision to attack, who is responsible for the outcome of that action? That question still has no clear answer in international law, military treaties, or humanitarian conventions. While governments and organizations try to debate regulatory frameworks, autonomous drone factories are operating at full speed, producing equipment that goes far beyond what any existing regulation can keep up with. 🚀
The arms race nobody can hit pause on
The scene in Beijing in September 2025 served as a symbolic trigger, but the arms race focused on AI was already well underway before that. In 2017, Putin himself declared that whoever led in Artificial Intelligence would become the ruler of the world. In 2024, Xi Jinping stated that technology would be the primary battlefield of geopolitical competition. And in January 2026, U.S. Secretary of Defense Pete Hegseth ordered all branches of the U.S. Armed Forces to adopt AI, saying it was necessary to accelerate like never before.
Receive the best innovation content in your email.
All the news, tips, trends, and resources you're looking for, delivered to your inbox.
By subscribing to the newsletter, you agree to receive communications from Método Viral. We are committed to always protecting and respecting your privacy.
The United States and China, the two largest military powers on the planet, are at the center of this contest. But the race has widened significantly. Russia and Ukraine, now in the fifth year of war, are seeking any technological edge possible. India, Israel, Iran, and other countries are investing heavily in military AI. France, Germany, the United Kingdom, and Poland are rearming amid doubts about the Trump administration’s commitment to NATO.
Every nation is trying to stockpile the most advanced technological arsenal possible, in case they need to face drones against drones and algorithms against algorithms in ways that human beings simply cannot keep up with.
Billions on the table
The numbers are staggering. The Pentagon requested more than $13 billion for autonomous systems in its most recent budget and has already spent billions more over the past decade, though the total is hard to track because AI funding is spread across numerous programs.
China, according to researchers, is spending comparable amounts and has used financial incentives to push the private sector to develop military AI capabilities. Russia, for its part, has invested in drone and autonomy programs, using the war in Ukraine as a testing ground to refine these technologies on an actual battlefield.
Palmer Luckey, founder of Anduril, stated that Russia, China, and the United States are building AI arsenals as a deterrent, following the logic of mutually assured destruction — the same concept that defined the nuclear era. The comparison, by the way, has been a recurring one: the current buildup of autonomous weapons has been likened to the birth of the nuclear weapons era in the 1940s. 🔍
The Ukrainian laboratory and the lessons the world is learning
Russia’s invasion of Ukraine in 2022 turned theory into reality.
Outmatched in numbers, funding, and equipment, Ukraine held off Russia with an improvised arsenal of cheap technology. Racing drones used by hobbyists were adapted to strike Russian positions on the front lines, eventually becoming more lethal than traditional artillery — and in some cases gaining autonomous capabilities. Remotely controlled boats kept the Russian fleet in the Black Sea virtually paralyzed.
Russia adapted too. Its Lancet drone, initially piloted by humans, began incorporating autonomous targeting features. According to American officials, Russia is building Lancets capable of loitering in the sky and choosing targets autonomously.
As Michael Horowitz, a former Pentagon official involved in autonomous weapons development, put it, four years of brutality on Ukrainian battlefields served as a laboratory for the entire world.
Recently, Ukraine began sharing its massive volumes of combat data with Palantir and other companies so that AI systems can learn to fight wars more efficiently. In Europe, governments seeking to reduce their dependence on the American military apparatus absorbed the lessons. In February 2026, Germany, France, Italy, the United Kingdom, and Poland announced the joint development of an anti-drone air defense system.
The role of defense startups in this new era
One of the most surprising shifts this new era has brought is the growing role of private companies — especially startups — in developing military technology. For decades, the defense industrial complex was dominated by giants like Lockheed Martin, Raytheon, and Boeing, companies with multibillion-dollar contracts, slow processes, and enormous bureaucratic structures. Today, the pace of innovation that Artificial Intelligence demands doesn’t fit that model.
Anduril is the most emblematic example of this transformation. Founded in 2017 by Palmer Luckey — the same guy who created the Oculus VR — the company built a completely different proposition: instead of waiting for the government to define what it needed and then trying to meet that demand, Anduril develops its own technologies, tests them, validates them, and then offers finished products. This model flipped the traditional logic of the sector and has proven far more efficient in terms of speed.
And it’s not just in the United States. In China, Beijing pushed commercial tech companies into defense partnerships through a strategy called civil-military fusion. Private companies were integrated into military procurement processes, joint research, and other work with defense institutions. At the Zhuhai airshow in 2024, Norinco, one of the country’s leading defense manufacturers, unveiled multiple AI-capable systems — including an entire brigade of armored vehicles and drones controlled and operated by Artificial Intelligence. Another piece of equipment on display was a 16-ton jet-powered drone designed to function as a flying aircraft carrier capable of launching dozens of smaller drones in mid-flight. 💡
Project Maven and computerized warfare
The story of how the United States got here passes through a project called Maven. In 2017, Lieutenant General Jack Shanahan helped create this initiative within the Department of Defense to incorporate AI into military systems. The problem was clear: the American military was collecting so much data — drone footage, satellite imagery, intercepted signals — that nobody could process all of it in a useful timeframe.
The goal was to work with Silicon Valley to build software capable of rapidly processing those images for intelligence purposes. Google was initially chosen to help, but when employees at the company discovered the project, they protested internally, arguing that a company that once adopted the motto don’t be evil shouldn’t be helping identify targets for drone strikes. Google ultimately walked away from the project.
In 2019, Palantir, the data analytics company co-founded by tech investor Peter Thiel, took over Maven. Today, the system is powered by commercial AI — including a military version of Claude, the chatbot from Anthropic — and is capable of analyzing intelligence from multiple sources, generating priority-ranked target lists, and recommending weapons, essentially eliminating the gap between identifying a target and destroying it.
At a conference livestreamed by Palantir following the American and Israeli strikes on Iran in February 2026, a Pentagon official demonstrated how the system worked: a satellite feed showed a warehouse. With mouse clicks, an officer selected targets in real time. Within seconds, the AI software suggested a weapon, calculated fuel and ammunition requirements, estimated the cost, and generated an attack plan.
Cameron Stanley, the Department of Defense’s chief digital and artificial intelligence officer, called what Maven was doing revolutionary. Human involvement, he said, boiled down to left click, right click, left click.
Emelia Probasco, a senior researcher at Georgetown University’s Center for Security and Emerging Technology, cautioned that claims about Maven may be overstated and that much of the American advantage came from the scale of available data and the skill of the people operating the system. She added that China likely already has something similar — and, in fact, in her analysis of thousands of procurement documents from the People’s Liberation Army, she found Chinese systems that mirrored the American ones.
The questions that still don’t have answers
With all of this evolution happening in parallel, across multiple countries at the same time, questions arise that go far beyond military strategy. One of the most urgent is human control over lethal decisions. According to the original New York Times report, both China and Russia are experimenting with letting AI make battlefield decisions on its own. China is developing systems for dozens of autonomous drones to coordinate attacks without human input. Russia is building Lancet drones that can loiter in the sky and choose targets autonomously.
The only relevant AI arms agreement between China and the United States was signed in 2024 — and it’s merely a non-binding commitment to maintain human control over the decision to use nuclear weapons. Other countries, like Russia, have made no commitment at all. Liu Pengyu, spokesperson for the Chinese Embassy in Washington, stated that China has proposed international frameworks for governing military AI and called for a prudent and responsible approach to its development. The Pentagon and Russia’s Ministry of Defense did not respond to requests for comment.
Another point worth paying attention to is the risk of accidental escalation. When two countries have autonomous systems operating in the same space, any unexpected interaction can be interpreted as an act of aggression — and the speed at which this equipment operates is far greater than the human ability to intervene and de-escalate the situation. In exercises conducted since 2020, researchers at the RAND Corporation explored how autonomous systems could accelerate escalation and weaken human control — with alarming results. In one scenario, a system operated by the United States and Japan responded to a North Korean missile launch by autonomously firing an unexpected counterattack.
General Shanahan himself, who retired in 2020 and is now a researcher at the Center for a New American Security, said the race he helped start keeps him up at night. There is, in his view, the risk of an escalation spiral where each side fields untested, insecure, and unproven systems because it feels the other side is hiding something.
Proliferation as an additional threat
And then there’s the question of proliferation. Unlike nuclear weapons, which require enormous infrastructure and hard-to-obtain radioactive material, the components to build autonomous drones are widely available, cheap, and increasingly accessible. AI technology is becoming widespread, opening the door for countries like Turkey and Pakistan to develop new capabilities. What is currently a dispute among the world’s largest powers could, in just a few years, transform into a much more decentralized and harder-to-control reality.
Some experts argue that framing this as an arms race may be inaccurate. Michael Horowitz noted that AI is a general-purpose technology like electricity — and nobody talks about an electricity arms race. In his view, AI is transforming the military the same way electricity, computers, or the airplane did in their respective eras.
Within the Trump administration, however, the push for AI weapons has taken on an almost evangelical tone. Last month, the Pentagon classified Anthropic as a security risk, in part because the company wanted to limit the use of its technology for automated weapons. Jacob Helberg, Undersecretary of State for economic affairs, declared at a conference in Washington that the United States would win the AI race. 🌐
Deterrence or imminent danger
Palmer Luckey of Anduril argued that the buildup of AI weapons could, paradoxically, prevent wars between major powers. The logic mirrors that of the Cold War: if both sides know what the machines are capable of, neither would risk finding out in practice.
But deterrence assumes rationality — and autonomous weapons are designed to move faster than human reason can keep up with. That is the central tension defining this historical moment. The speed of autonomous systems has led to inadvertent escalation in simulated exercises, and there’s no guarantee that similar situations won’t happen in the real world.
The dynamics may echo the Cold War, but experts warn that the AI era is different in crucial ways. Startups and investors now play a role just as important as universities and governments. The technology is widely available. And what’s emerging is a continuous innovation race with no obvious finish line.
The arms race powered by Artificial Intelligence is not a future threat. It’s already happening — in factories in Ohio, in hangars in Beijing, on battlefields in Eastern Europe, in research labs scattered around the world, and at tech conferences in Washington. What remains to be seen is how humanity will decide — or fail to decide — to set limits on this technology before those limits lose all meaning.
