Both GPS and ARPANET, the germ of what would later be Internetwere born within the United States Department of Defense. From radar, which was developed during the 1930s and perfected during World War II, advances later emerged such as the air traffic control system or the microwave.
There was a time when military technology was a source of inventions for the civilian sphere. With AI, technical advances begin to occur in the opposite direction: from the civil to the military. They are companies that we all know – Google, OpenAI or Anthropic – that are developing the most sophisticated models in the world and defense organizations are waiting.
For the European Union, this trend coincides with an acute crisis in the field of defense. The war in Ukraine has undermined the foundations of political life in Brussels, while presenting Russia as a growing threat. Added to this is the Trump Administration’s willingness to stop security aid to European countries. A breeding ground that has driven the need to rearm.
In March, the European Commission announced the ReArm Europe initiative o Readiness 2030. The objective is to significantly increase the EU’s defense capabilities. This is a plan that wants to mobilize nearly 800,000 million eurosincluding 150,000 million in loans for military investments. State governments are expected to push for the modernization of their armies, but the mobilization of private capital is also sought.
The increase in security spending is also reflected in the Multiannual Financial Framework (2028-2034)presented in summer. This budget, which defines at a financial level the EU’s priorities for the coming years, provides for an allocation of 131,000 million euros to support investment in the areas of defense, security and space. Without knowing how the distribution is, the funds are five times higher than those of the previous period (2021-2027).
(Pablo Bejarano)
These efforts of the EU to recover the lost ground in defense could benefit from the technological race that is currently being experienced. Above all, advances in AI, called to reconvert armies and forms of deterrence that countries exhibit today. In one of the round tables at the Web Summit, held in Lisbon last November and which acts as a meeting point for startups from around the world, several experts addressed this topic. Under the title ‘From code to combat: Why AI defense tech is exploiting‘, participants discussed the rise of AI in defense.
“What is changing the technological landscape is the speed of innovation,” commented Josh Araujo, CEO of the startup Forterra, in reference to the accelerated pace of technological evolution. “And in Europe, what happens with defense budgets is that it is no longer interesting to buy old things. You can take a system, man it with humans and put armor on it, so we are talking about refined and expensive systems that take decades to design. Or you can deploy autonomous systems low cost that put humans out of danger and allow more firepower and more deterrence capacity to be projected at a much lower cost.”
The boost of startups and private capital
Forterra is an American company dedicated to developing autonomous ground systems for the field of defense and industrial logistics. Araujo is used to dealing with both military and civilian actors and highlights the importance of deterrence: “The point is that for Europe and our allies, it costs aggressors much more to carry out an aggressive action. The key here is to deploy as much capacity as quickly as possible at the lowest possible cost.”
To make this deployment at low cost, AI will be key. a report of the think tank RAND recommends estimates that advances in autonomy and robotics will allow this type of technology be used en masse. “AI represents a great opportunity. If we realize what we generate today on the battlefield, with swarms of drones, autonomous ground vehicles and different mission systems, we see that there are enormous amounts of data,” says Araujo.
“Traditionally you had to have a lot of people staring at their computer screens to analyze and absorb the information. But AI offers us the possibility of taking this information, giving it meaning and putting it in context so that a human can make decisions based on that information,” he adds. Forterra’s CEO adds that this change has occurred over the last three or four years.
(IDF Spokesperson’s Unit/Commons)
One of the most direct use cases of AI has to do with improving the surveillance capacity of the terrain, through the analysis of mass images. But the technology has also been used in the search and identification of targets or to guide drones towards the target without the intervention of an operator. Likewise, in Ukraine has been rehearsed vehicle deployment terrestrial autonomouswhile the launch of swarms of drones that act on missions in a controlled manner is proposed.
All are formulas for increasing war capabilities at low cost, compared to the means that have been used to date. There is still a lot to refine, yes. In an informal conversation, a director of a company in the defense sector in Spain admitted that generative AI is still in the exploration phase and that for now it does not have clear uses in the military field.
This does not prevent a flood of technology startups from directing their efforts towards the defense sector. James Cross, co-head of private investment at the firm Franklin Templetonwas the other participant in the Web Summit round table and painted a context full of economic incentives: “I have been investing in defense since the late 90s, but things have changed a lot in the last ten years. Before, no venture capital firm would have invested in a defense-oriented company and today, apart from AI, defense is probably the hottest sector.”
Cross seasoned this vision with two notes: governments now they have gotten involvedsomething that has not happened for a long time, and more and more money is going to startups instead of traditional arms contractors. In the first half of 2025, venture capital in the United States injected around $38 billion into technology startups linked to defense, according to JP Morgan. Anduril, Palantir or SpaceX are some of the names that are heard.
From the commercial market to the military field
But there are other, much more popular companies that have entered the defense market. Government contracts in this space are lucrative and this has attracted big AI players. In the United States, the Pentagon has signed contracts with OpenAI, Google, Anthropic and xAI worth $200 million each for artificial intelligence systems.
As a note, for months Google officially abandoned their engagement —acquired in 2018 after protests among its staff— not to use AI for cyber surveillance or weapons applications. It is not a minor detail.
Both Google and OpenAI, Anthropic or xAI are companies that have developed their systems to the commercial world. Their products are aimed at consumers or businesses and thrive within that framework. But in this way they have built the most advanced models of the moment, which are now exportable to the military field.
We already have small “autonomous” ground-level experiments in Ukraine, like this Predator turret mounted on a tracked vehicle. (UGV Robotics)
Araujo spoke of the advantages that this way of operating presents for companies: “If you can find applications that have dual use, both in commercial and industrial contexts as well as military, you will be much more cost-efficient when delivering your capabilities, because you will be selling to a much larger market. In addition, you will be adapting your system with larger data sets, both commercial and military, and you will be able to achieve models that generalize better.”
In Europe, Mistralthe generative AI trump card within the EU has also been launched search for contracts in the military field. It has teamed up with the German startup Helsingspecialized in autonomous systems, and its CEO Arthur Mensch declared that it was essential that military spending on AI will be channeled towards European companies. A clear allusion to the technological sovereignty promoted by European institutions.
It’s not that easy
Of course, deploying AI on the front lines is not comparable to delivering services to a company or a mass of consumers.
There are technical difficulties that have yet to be resolved. “We have seen it in Ukraine with drones and robotics. Their expansion brings enormous amounts of data and the best way to manage this is by leveraging AI. But to introduce AI to the battlefield we need to deploy it locally,” says Araujo. “We can’t just go to a normal cloud from a 5G network. We have to have infrastructure to deploy AI. We are seeing this in the form of robots and autonomous vehicles, which drive the need for edge computing.”
The deployment of telecommunications, already complex in itself, creates another obstacle. “We cannot have demanding AI workloads with edge computing if we do not have an energy source that practically does not run out. And bringing energy to the front is quite dangerous and quite expensive,” Cross pointed out. To solve this, different options are being considered, including nuclear reactors small (SMR).
(Ministry of Defense of Ukraine)
Although AI is an economical shortcut To increase defense capabilities, new technologies imply an investment that returns us to increased spending. A policy that not all countries agree with. In the European Union itself, some member states have shown their reluctance. Among them, Belgium, Italy and Spain, which has committed to an increase of up to 2% of GDP this year (in 2024 it was 1.3%). Although the European Defense Commissioner Andrius Kubilius pointed out that the country must try to comply with the 3% that the United States will demand.
There is no doubt that weaponry will undergo a change due to the application of AI and autonomous systems. Modern warfare has taken another leap and this will bring military capabilities of unknown consequences. Hence some voices They have already requested that the use of this type of weapon be regulated.
Image | Ministry of Defense of Ukraine
In Xataka | The question is no longer whether Europe “is at war”: the question is whether it is willing to defend itself
GIPHY App Key not set. Please check settings