AI on the Pitch: What Euro 2024 Really Showed Us

In our previous investigation, we traced how UEFA had been laying the groundwork for an AI-driven transformation of football. Euro 2024 was where those blueprints met reality: ten German cities, 24 nations, 51 matches, and more artificial intelligence running beneath the surface than any tournament in history. What follows is what we found.

There is a moment every football fan knows intimately. The ball nestles into the back of the net, the striker wheels away, arms wide, mouth open, and then everyone freezes. The assistant referee's flag is up. Or maybe the VAR light is on. Or maybe nothing happens at all, and 80,000 people are left suspended in collective uncertainty, waiting for a machine to tell them whether joy is permitted.

At UEFA Euro 2024 in Germany, that relationship between football and technology reached a new frontier. What unfolded across ten German cities last summer wasn't merely a tournament. It was, quietly, a live demonstration of what happens when artificial intelligence is woven into the fabric of professional sport at the highest level.

The Match Ball That Tracks Itself

Before any tactical analysis or referee intervention, the technology began with something deceptively simple: the match ball itself.

• The Adidas Fussballiebe, a $170 sphere made partly from recycled polyester, corn fibers, sugar cane, and wood pulp, contained an Ultra-Wideband sensor that tracked its position with centimeter-level accuracy up to 400 times per second. A separate motion sensor operating at 500 Hz transmitted precise movement data to video officials within seconds of any incident. Twelve to twenty-four network antennas positioned around the stadium triangulated the ball's exact three-dimensional location at all times.

The result was a ball that, in a very real sense, knew where it was. And crucially, it could tell everyone else within 20 milliseconds.

This was the first time connected ball technology had appeared at a European Championship. The implications were immediate. Every touch, every pass, every shot generated a data signature. Machine learning algorithms processed that raw information in near real time, calculating ball speed, pass distance, and identifying the precise moment of contact, the kind of granular information that once lived only in a coach's memory or a commentator's gut feeling.

How AI Ended the Offside Argument

Nothing in modern football generates more debate than the offside call. A shoulder, a kneecap, a fraction of an armpit: the margins have become so fine that even freeze-frame television replays leave fans arguing for days.

Semi-Automated Offside Technology, deployed across all Euro 2024 matches, attacked this problem methodically. Ten specialized cameras tracked 29 separate body points on every player simultaneously. Combined with the connected ball's ability to pinpoint the exact moment of contact, the system could determine a player's position at the precise instant the ball was played, not the moment the camera happened to capture a frame.

The difference is not trivial. Traditional VAR offside checks required officials to manually draw lines across grainy freeze-frames, a process that could take several minutes and still feel inconclusive. SAOT compressed that process dramatically, producing decisions that were both faster and geometrically precise. Play resumed sooner. Arguments, at least the ones about the data, became harder to sustain.

VAR in Germany: Precision Gained, Spontaneity Lost

The Video Assistant Referee system at Euro 2024 was supported by a Football Technologies Hub based at the international broadcast centre in Leipzig, with four dedicated Video Operations Rooms staffed by two assistant video referees and three video operators per match. Their mandate was narrow by design: goals and their build-up, penalty decisions, direct red cards, and cases of mistaken identity.

What made the system notable wasn't the technology alone but the structure around it. All significant VAR decisions were shown on stadium screens so fans inside the ground could understand what was being reviewed and why. It was a deliberate acknowledgment that transparency matters, that 40,000 people sitting in the cold deserve to know why the celebration they just had is being taken away from them.

Still, the criticism never fully disappeared. VAR's detractors argue, with some justification, that the technology has traded one kind of injustice for another. Refereeing errors, once forgotten by the next match, now linger in high-definition replays and algorithmic verdicts. The spontaneous eruption of a goal, that pure, unrepeatable surge of noise, gets clipped at the knees while officials stare at monitors in Leipzig.

The question the technology has not yet answered is not whether AI can make better decisions than a linesman. It clearly can. The question is: what is football actually for?

From the Boardroom to the Training Ground: How Clubs Are Using Data

Away from the pitch, AI had already been reshaping football for years before a ball was kicked in Munich. At Euro 2024, the squads assembled by competing nations were products of a scouting revolution that is still accelerating.

Brentford's recruitment operation uses computer vision to process match footage and surface players based on expected-goals metrics and shot-quality statistics, rather than reputation or word of mouth. It was this approach that identified Neal Maupay before most clubs knew his name. Brighton, operating under a similarly data-driven philosophy through their Starlizard platform, built a squad piece by piece, using thousands of data points per player to assess suitability, injury risk, and tactical fit simultaneously. Moisés Caicedo and Alexis Mac Allister, both sold for enormous fees, were products of that system.

At Liverpool, a collaboration with Google DeepMind produced TacticAI, a system that analyses thousands of historical corner kicks using geometric deep learning to generate high-probability set-piece sequences. In testing, coaches preferred the AI-generated setups to human-designed ones nine times out of ten. The tool is now used in real matches.

Sevilla FC, meanwhile, adopted an NLP-powered scouting assistant built on Meta's Llama model. The system digests vast scouting report databases and synthesises them using retrieval-augmented generation, compressing weeks of manual reading into seconds of intelligent summary. A scout flying to watch a left back in Bratislava now arrives already knowing what the data says about him.

Football's Unanswered Question

The honest answer is that nobody fully knows what comes next. Euro 2024 demonstrated that AI can make refereeing faster, scouting smarter, and training more precise. It has not yet been demonstrated that it makes football more beautiful, or more human, or more worth watching at 3 a.m. in a foreign city because something about it still feels alive.

The researchers who documented the tournament's technology put the essential tension plainly: who is a football game actually for? Investors and broadcasters have one answer. The man in the upper tier who queued for ninety minutes in the rain has another. The ball, tracking itself 400 times a second across a German summer, does not have an opinion on this. But someone will have to decide.

Images created by AI.