The Transformation of Defense Economies in the Age of Artificial Intelligence

I was talking to a distinguished computer engineer friend, who studied and worked in universities and companies that are the dreams of many young people anywhere in the world. He has now decided to study business administration (MBA), even though he holds a master's degree in politics and technology, and has shifted his specialization to artificial intelligence, not in programming, but in the management and economics of artificial intelligence (yes, there is a specialization with this name now).

What’s important is that he told me about an American company founded by "Palmer Luckey" after he sold his previous company Oculus VR to "Meta" for 2 billion dollars in 2014, and then established a new company called Anduril Industries. Believe me, this company has not just changed the shape of the defense industries in the world... it has actually changed it.

I will talk about it later, but I would like to discuss today a topic related to the future of defense economics and artificial intelligence.

Let’s think together about the next phase of the evolution of defense economics. Many of us look at artificial intelligence as a new tool to increase the efficiency of weapons, but the truth is that its importance lies in the fact that it changes the essence and shape of defense economics fundamentally.

To understand this transformation, we need to answer an important question: what does it mean to have an "algorithm-based defense economy" instead of an "economy based on factories of iron and nails"?

In traditional defense industries, value was created inside the factory; the factory that makes the airplane or tank or missile. But today, value is created in code, data, and algorithms that make the plane or drone think in a smarter way, or perform a task completely different from others even if they are of the same model. This is a radical economic transformation, because you are no longer selling a piece of iron, but selling a cognitive system.

If we go back just ten years, the term "operational autonomy" meant a decision-supporting system. Today, the world has moved towards systems called Human-Out-of-the-Loop systems, meaning that the decisions to kill or engage are made autonomously without direct human intervention.

This development is practically creating a new economic layer within the industry that we can call the defense artificial intelligence layer.

The most important aspect is that the companies controlling defense artificial intelligence are not the same ones producing airplanes or radars, but rather independent new companies entering the equation. Therefore, we can consider it a new layer in the industry, such as: Palantir, Shield AI, Helsing in Germany, or Anduril that we started talking about. We are talking about software and cybersecurity companies that have started to compete with traditional arms factories themselves.

You might now wonder: why aren’t the big old companies doing this work? A valid and difficult question at the same time, and departments like technology management and innovation in many universities have been trying for years to answer it.

Returning to the essence of our topic: this transformation creates what economists like me (damn my humility) call the unbundling of the defense value chain, which means separating the material industrial system from the cognitive digital system, and that is a field of research and development in itself.

The state that realizes this meaning early (hear me, state!) can create for itself technological and economic advantages at the same time.

To understand the importance of this change, let us look at some numbers.
In the United States, more than 70% of defense artificial intelligence contracts have gone to companies that have never been part of the traditional defense bureaucracy, but to startups. Goodbye to the giants of the old industry and the antiquated companies in the market, as we are now talking about companies that are only five or ten years old. This is indeed dangerous.

In Europe, the European Defense Agency has launched the AI4DEF program to reduce dependence on American technologies in smart applications, which is a massive program focusing on sensor fusion algorithms and autonomous control — also through startups.

As for China, it is doing the exact opposite; it integrates the civil and military sectors into one framework known as civil-military integration, and focuses on linking artificial intelligence with missile and drone systems through national supply chains. This means that China systematically arms any civil artificial intelligence technology.

From an economic perspective, this means that the defense sector has entered a new phase that can be called "turning knowledge into a commodity". This phase may be familiar in other sectors, but it is the first time in history that defense enters this phase. We’ve been accustomed to seeing defense as a "heavy industry" requiring giant companies and huge investments, but now we are talking about a new reality: every line of code and every machine learning model has become of real monetary value in the defense market.

This will change the power dynamics in the defense industry and will also change the nature of defense funding itself, because you no longer need huge industrial capital to build a tank or an ammunition factory, but need human and cognitive capital.

Bringing all these data together, it can be said that the defense industry is gradually transforming from a metal economy to a knowledge economy.

Here we arrive at a more dangerous question — a question that could be the subject of master's and doctoral theses:
How do countries that are still building their defense industrial bases deal with this transformation?
Do they start with iron and electronics? Or do they directly enter into software, smart control, and embedded systems?

My perspective is that the shortest way for any country outside the club of the greats is to enter through the door of defense artificial intelligence directly.
Not because smart weapons are easier — but because they are much more complex — but because the capital barrier is much lower. You don’t need billions for factories, but laboratories and scientists, data and flexibility in thinking, which are the resources that enable a new generation of countries to create a digital relative advantage instead of the traditional industrial advantage.

Countries that have not yet understood the meaning of smart defense economics will look at the world in just five years and discover that their pursuit of iron was a wrong and costly decision, and that those who preceded them were the ones who began building algorithms and code.

We are facing a truly astonishing scene: artificial intelligence originally started as civil, then was adopted by defense industries, and is now developing it to later transfer it back to civil sectors to benefit from it.
And those who understand this logic are building today defense artificial intelligence systems not only for war but because these systems will extend later to all civil industries.

I hope that "very poor" countries, whose "people are also very poor", realize this change and make a real transformation, to bring back their children scattered around the world, so they build an industry with their minds, not with giant factories and huge investments. The opportunity is really in front of you, and even very poor people can exploit it.