NATO builds an AI 'Kill Web' to stop Russian attackers before they penetrate its territory

The article, written by Luca-Marie Hoffmann for BILD and published by Business Insider as part of the Axel Springer Global Reporters Network, describes a new NATO defense strategy on its eastern flank called the Eastern Flank Deterrence Initiative (EFDI). According to documents obtained by BILD, the initiative explicitly identifies Russia as the potential adversary against which this system is designed.

The core idea is that the Alliance no longer relies solely on tanks, fighter jets and troops as a first response to aggression, but is deploying a vast digital network of thousands of sensors, drones, satellites and artificial intelligence along its entire eastern border, which now stretches from Finland, through Estonia, Latvia, Lithuania and Poland, to Romania on the Black Sea, also including the border with Belarus. The goal is to detect any attack as early as possible and to stop the aggressor before it penetrates allied territory.

This approach marks a shift from NATO's traditional doctrine, based for decades on 'deterrence by punishment' (repelling and then recovering lost territory with conventional forces). The new layer relies on the concept of 'deterrence by denial': preventing the attack from succeeding in the first place, not just punishing it afterward. According to Major Matt Blubaugh, spokesman for the US Army in Europe and Africa, the EFDI 'does not replace tanks, artillery, fighter jets or soldiers,' but rather seeks to 'preserve their combat power and give commanders more time and decision advantage.' Assets such as the Leopard 2 and M1 Abrams tanks, HIMARS rocket launchers, artillery and F-35 fighters will remain the backbone of the defense.

The technological core of this network, called the 'Kill Web' in the documents, is a system of interconnected nodes —satellites, reconnaissance drones, radars, ground sensors, cameras and electronic surveillance assets— designed so that if one node fails, others immediately take over its function. All that information gathered by NATO members would flow into a shared digital network, analyzed in real time by artificial intelligence, greatly streamlining a process that previously required a drone to report to a headquarters, the headquarters to analyze the data, and a firing order to then be transmitted through the chain of command to a military unit.

At the center of this open-system architecture would be Palantir's Maven Smart System (MSS), described as the artificial intelligence 'brain' of the EFDI, tasked with processing sensor data from all domains to speed up decision-making. The system would also integrate technologies from other defense contractors, such as Perennial Autonomy's AI drone interceptor Merops, along with capabilities from companies like RTX, Rheinmetall, Saab, Lockheed Martin and Boeing, all connected through the so-called 'EFDI Data Backbone' into a single sensor-to-shooter network. NATO sums up the principle in three steps: 'See first. Decide first. Shoot first.'

Another significant change is that the first line of contact would no longer be made up of soldiers. The plan envisions a forward zone in which drones, ground robots, sensors and other unmanned systems —such as the Hunter Wolf unmanned ground vehicle mentioned in the article— would confront the attacker first, absorbing the initial impact to buy time and preserve the combat capability of NATO's conventional formations. Even so, Blubaugh insists that 'at the end of the day, you still need soldiers, tanks and planes to secure and hold ground.'

The article explicitly notes that this concept is inspired by the lessons Ukraine has learned on the battlefield against Russia, particularly the massive use of low-cost drones, robotic systems and sensors to offset Russia's advantage in troop mass and speed of advance. The implicit comparison is that NATO would be replicating, in its own way, the powerful detection and strike network Russia has developed in its war against Ukraine.

It should be noted that the original Business Insider text is marked as subscriber-exclusive content ('This story is available exclusively to Business Insider subscribers'), although the full body of the article, with all its sections, was in fact available in the download used for this summary. Nevertheless, given the restricted-access notice on the page, readers should be warned that the article may be subject to a paywall if they try to read it directly on the site.