Bundesliga places AI at centre of its media strategy

SVG

article here

Introducing the Bundesliga’s latest showcase of technology innovation Deutsche Fußball Liga (DFL) CEO  Steffen Merkel put the emphasis firmly on AI.

“AI is not just another tech trend for us,” Merkel said. “It is becoming a fundamental capability that runs through the entire organisation, just as it does in many other industries—including sports and football.”

The event, held in Dortmund over the weekend ahead of the Franz Beckenbauer Supercup, featured presentations, a tour of Boroussia Dortmund’s Signal Iduna Park and VIP boxes stuffed full of new toys.

“I’m a firm believer in trying things early and learning from every experiment,” said Merkel, who oversees the league’s media and digital strategy. “Sometimes a prototype becomes a product; sometimes it doesn’t. But there is always added value in the process.”

More than 43 million people in Germany—almost half the population—follow the Bundesliga, spanning all demographics, each consuming content differently. Internationally, the diversity is even greater.

“This makes it clear why technology is essential to scale, localise, and personalise content for fans everywhere,” he said.

With around 400 employees across the organisation and roughly 250 dedicated to content production, the Bundesliga faces a scale challenge familiar to many rights holders. “When it comes to our international fan base and localisation, it’s impossible to do that manually. So AI is a huge opportunity for us to close the gap.”

The raw material feeding AI are 200 million data points collected for each match. Event data, positional tracking and 3D tracking feed innovations such as Bundesliga Match Facts powered by AWS, semi‑automated offside detection and automated event recognition. “Match data has become ever more intertwined with the media product,” he added, describing a future where real‑time insights shape fan engagement, media production, and increasingly the sport on the pitch (through AI assisted coaching and scouting for example).

From a study into AI commissioned by the league two commercial trends stood out. “The ones with the highest probability are personalised content and content scaling,” explained Professor Sascha Schmidt, Chair and Director of the Center for Sport and Management (CSM) at WHU – Otto Beisheim School of Management. Experts expect “a seamless integration of individually curated content in real time” to be standard by 2030, alongside the ability to “double the amount of Bundesliga content produced with the same resources.”

AI officiating support generated the strongest divide. “There was no clear consensus,” Schmidt said, with desirability scores sharply split.

The study also found that clubs recognise AI’s potential but feel “only moderately prepared” to capture its full value due to limited resources, talent and know‑how.

“Our task as a league is to unlock these potentials and help clubs do the same,” Merkel added. “Some clubs are already well equipped; others—especially in Bundesliga 2—benefit from us acting as an early‑stage accelerator.”

Bundesliga tests localised world feeds

The Bundesliga has taken a significant step toward reshaping how global football audiences experience live matches, unveiling early results from a proof‑of‑concept (POC) exploring localised world feeds. The initiative, led by Tabea Wagner, Senior Manager Live Product at Bundesliga Media, aims to deliver tailored match broadcasts for international markets without compromising the integrity of the league’s central production.

“As a rights holder, it’s our responsibility to deliver a top‑notch media product—especially for fans who aren’t privileged to sit in the stadium,” Wagner said. “We want to be the most fan-centric league – but if a Bundesliga match is not shown because there is no local commentary we cannot create any engagment.“

Today, the Bundesliga distributes one English world feed to dozens of markets with different languages, cultures and viewing habits. While broadcasters often add local commentary, resource constraints mean localisation is inconsistent.

“Our concept is straightforward,” Wagner explained. “We want to create many experiences based on the same feed. To do that, we add new layers—graphics localisation, commentary localisation, and eventually content localisation.”

Four factors convinced the league that the timing was right: rising fan demand for personalised experiences; rapid advances in AI translation and cloud workflows; an industry shift from satellite to streaming distribution; and growing momentum among other rights holders pursuing similar strategies. The Bundesliga’s partnership with AWS provides the technical backbone.

The POC focused on graphics localisation and audio localisation. Bundesliga Media worked with workflow‑orchestration startup Logic, running on AWS infrastructure, to coordinate the process.

Graphics localisation was tested using two methods and four service partners

·        Template‑driven HTML5 graphics, translated in advance and triggered by metadata managed by Vizrt.

·        Computer‑vision models that detect, translate and replicate graphics in parallel streams managed by mobii.

Audio localisation also followed two paths:

·        AI translation of the existing English commentary into multiple languages provided by Camb.AI

·        AI‑generated commentary based on match data, including 3D tracking, from Valka.AI.

Providers were asked to localise graphics into Arabic and German, translate commentary into Spanish and German, and generate new English commentary.

What worked—and what didn’t

“From a technical perspective, it is working,” Wagner said. But challenges remain.

Latency is a major obstacle for live environments. Graphics localisation proved more complex than expected—not in quality, but in how it fits into existing infrastructure. And language localisation is “not yet market‑ready.”

“Latency must go down,” Wagner said. “Commentary needs more emotionality and authenticity. And grammar errors must be eliminated. If the commentary isn’t as good as a human, we won’t put it into production.”

What’s next?

The league will continue monitoring service‑provider development and gather feedback from media partners to build a clear roadmap. Wagner emphasised that the future may not require more content—just more relevant experiences based on the same content.

“Localisation is a possible path,” she said. “It’s too early for conclusions, but it’s worth exploring—and that’s exactly what we’re doing.”

At the Super Cup, another layer of sign‑language localisation was trialled in partnership with Synapse.

The more difficult question is whether personalisation should extend to the live match feed itself. Merkel acknowledges the tension.

“Given all the content and visualisation possibilities, the question becomes: do you want every fan to have their own individual experience, or do you still want a shared, joint experience? That’s the trade‑off. On one side, you have deep personalisation; on the other, the common experience that defines football culture.”

He describes this as one of the central debates in football’s AI future. “This is part of the journey toward AI. Finding the right balance between individualisation and togetherness.”