Tatiana Dias is a fellow at Tech Policy Press.
Miami, June 24, 2026: WORLD CUP, BRAZIL vs. SCOTLAND. General view of the VAR screen at the Miami stadium for the match between Brazil and Scotland at the 2026 World Cup. Photo: Rodolfo Buhrer/AGIF
In this World Cup, a spectacle of sensors, cameras and artificial intelligence has made technology part of the sport’s appeal. The AI-powered World Cup has been celebrated for bridging the gap between smaller and larger teams, and for making the tournament “smarter,” “more inclusive and more accessible,” as Lenovo CEO Yuanqing Yang announced at CES 2026.
Lenovo is FIFA’s technology partner and provider of the technology that is increasingly part of gaming. Players’ bodies have been rendered as 3D avatars. We can see them on the screen, when the VAR system is analyzing the position of a player. Every team has an AI assistant at its disposal: based on thousands of data points collected during a match, the Football AI Pro platform, launched by FIFA with Lenovo, produces tactical reports and strategic recommendations in real time.
The Adidas match ball has motion sensors that record everything and transmit it in real time to the VAR system. Sensor data is combined with player positions and video to review incidents on the pitch. The promise is more precise and faster refereeing.
The effects can be seen on the field and in the results. Croatia were eliminated by Portugal after a goal was disallowed because the ball detected a touch from a player in an offside position through VAR review. Iran fell in the group stage when a goal was disallowed due to a millimeter offside also identified by the VAR. Egypt was subsequently eliminated by Argentina after the VAR detected a foul committed by an Egyptian player before his shot on goal.
So far, there have been more than 100 VAR interventions in the World Cup, an increase that FIFA anticipated. It was a decision by Pierluigi Collina, president of the FIFA Referees Commission, to reduce the time lost during matches. Meanwhile, criticism has also grown that soccer, a sport defined by creativity and improvisation, is becoming too much of a game of precision.
The World Cup most dazzled by technology is also the most lucrative in history. And the two phenomena go hand in hand in the datafication of football.
However, the techno-solutionist spectacle of this World Cup is part of something bigger. Beyond the well-known complaints that technology is killing subjectivity and improvisation in football, can anyone imagine Maradona’s Hand of God surviving a VAR review? — this process is building a new and profitable industry that is transforming modern football, powered by sensors, artificial intelligence, betting and, crucially, invisible labor.
The political economy of football data
For Rafael Grohmann, a professor at the University of Toronto, the conversion of football into data is “a striking example of how financialization and datafication travel together.” Grohmann recently set out to investigate the world of tech workers in football.
In his 2022 book Expected goalsjournalist Rory Smith describes one of the starting points of the datafication of football: the founding of the data analysis company StatDNA in 2009. Its founder, Jaeson Rosenfeld, also co-founded Digital Divide Data, a company that outsources data and technology work in Laos and Cambodia with the aim of “helping the world’s poorest to benefit from information technology by creating sustainable social enterprises” and provided labor to StatDNA.
According to a 2014 article in The Guardian, Digital Divide Data workers looked at match footage and coded it into specific data, such as the position of goalkeepers and the striker’s favorite foot. In 2012, StatDNA was bought by Arsenal, the English club that pioneered the adoption of data analytics.
Jaeson Rosenfeld now serves as an advisor to FIFA on advanced analytics.
According to Grohmann, over the last decade, as transfer fees have skyrocketed and England’s Premier League has become the richest league in the world, the process has consolidated, accompanied by an influx of American money into Europe’s top leagues. Today, more than half of Premier League clubs are majority owned by American individuals or companies.
The datafication of football is creating a new industry, one modeled after the global AI industry: lucrative, concentrated and unequal. Most clubs lack the ability to collect and process data like Arsenal did, making data providers the profitable end of this industry.
Companies such as Hudl, SkillCorner and Sportradar promise to develop “winning strategies”, “empower smarter decisions” and “increase fan engagement” through AI-powered football, and count national teams as well as clubs, federations, media companies and betting and prediction markets among their clients.
Physical data generation is dominated by Sony, which owns a number of sensor, wearable and computer vision companies, including Hawk-Eye, the operator of the technologies deployed in this World Cup, including VAR.
As with that pioneering company, Grohmann’s early research shows that football datafication follows the same logic of inequality as the global AI supply chain: rich countries capture the money, while Eastern Europe, Africa, South Asia and Southeast Asia are left with the low-paid work of data annotation.
The stories that data doesn’t tell
On her first day of work as a soccer data recorder, Ashley Flores of the Philippines was faced with one task: watching the 2014 World Cup semi-final, the one in which Germany crushed Brazil, playing at home, 7-1. Flores wanted to be a soccer player, but the chances of achieving it in the Philippines are slim. So he became a “tagger”, a data labeler feeding the new and profitable data-enabled football industry.
Flores’ story appears in the introduction to Expected goalsBook by Rory Smith from 2022. Behind what some see as a technological revolution lies an economic and labor chain that reveals a lot about how a subjective human process like football is converted into quantifiable data.
At the company where Flores worked (Packing Sports, the Manila subsidiary of German firm Impect), data taggers spent their days watching games and recording every move. They trained with the game 7-1, counting how many rivals were bypassed, either through a dribble or a quick pass.
This is because Brazil’s historic defeat was not a typical match. Considered apart from the score, the numbers suggest two evenly matched teams, and even a slight advantage for Brazil, which had 52% possession and more corner kicks and shots than Germany.
“It was the first match that really highlighted the difference between our data and what you normally see,” explains Lukas Keppler, CEO of Impect, whose clients include Bayern Munich and Paris Saint-Germain, in Expected goals.
So instead of traditional match statistics, Impect offers “Packing”, a metric “that measures the value of actions and events in football games.” With the promise of better data, scorers are trained to count opponents between the ball and goal, judge the distance between players and their markers, and estimate the degree of pressure on the player in possession.
Thanks to this human work of collecting match data, Impact offers more than 1,200 ways to evaluate a player. The information is organized into detailed analyzes that can be used for tactical decisions, player signings, team selection, investment strategy or betting, according to their website.
According to Rafael Grohmann, in the United Kingdom, the increase in online betting was a phenomenon that drove the transformation of football. “It was betting in England that drove further investment in football data science,” explains Grohmann. Today, the betting market and prediction platforms like Kalshi and Polymarket are part of this industry.
“It’s pretty clear that the data I record is for betting,” a data recorder in Rio de Janeiro told Rest of World. Speaking anonymously, he said he is paid around €60 per match to record the action, money he considers reasonable. “They need real-time data to adjust their probabilities throughout the game,” the worker said. “So I can’t send the data late or I risk losing part of the payment.”
The technological spectacle presented at this World Cup promises to eliminate human error from the game. But it does not eliminate the human; it relocates and hides them, while transforming the game into monetized data.
This is not the peculiar fate of football. It’s the general trajectory of many industries: turn behavior into data, turn that data into a priced asset, and send the bill for the conversion to the Global South. Football simply had the misfortune, or luck, of being the place where the operation became a spectacle, broadcast to billions, with commentary and replays.
