One second. That is how long it takes to scan a player and generate the 3D model that will determine whether a goal stands at the 2026 FIFA World Cup. The speed is almost beside the point. What matters is what the scan produces: imagery precise enough to settle offside disputes that have fractured fan trust in football’s most contested technology for years.
The broader context, according to the announcement, is an organisation that has fundamentally changed how it runs its flagship event. Previous World Cups distributed operational weight across local organising committees. For 2026, FIFA is running operations directly, across three countries — Canada, Mexico, and the United States — with 104 matches, 48 teams, more than 180 broadcasters, and an expected global audience of 6 billion people. That is not an incremental increase in complexity. The jump from 32 teams and 64 matches in Qatar was a structural change, and the operational model had to change with it.
The AI tools unveiled at Lenovo Tech World in Hong Kong are best read as responses to that pressure, not additions to a system that was already working.
The tool every team gets, regardless of budget
Football AI Pro is a generative AI assistant that will be available to all 48 competing teams. It is built on FIFA‘s Football Language Model, trained on hundreds of millions of FIFA-owned data points, and delivers pre- and post-match analysis across text, video, graphs, and 3D visualisations. It supports multiple languages. It will not operate during live play.
The access argument behind it is direct. A wealthy footballing nation fields a dedicated analytics department. A team appearing at its first World Cup does not. Football AI Pro is designed to give every squad the same analytical starting point. Whether the output is equally useful to both — or whether the gap simply shifts from data access to data interpretation — is a question the tournament will answer.
Deploying that system consistently across 48 teams, three host countries, and a weeks-long match schedule is the kind of infrastructure challenge that sits at the centre of Lenovo‘s enterprise AI positioning, which helps explain why the partnership extends well beyond hardware.
Transparency as the real purpose of better footage
The updated Referee View uses AI stabilisation to smooth body-camera footage in real time, eliminating the motion blur that made the original version difficult to follow during fast passages of play. The first iteration was trialled at the FIFA Club World Cup last year. The 2026 version is a technical step beyond that.
It will produce better broadcast material. That is the framing in the announcement. But the more durable purpose is officiating transparency. VAR has generated persistent controversy not only because fans dispute decisions, but because the imagery used to communicate those decisions has often been unclear. Sharper, stabilised footage from the referee’s perspective changes the evidentiary record that audiences and broadcasters are working from.
If it shifts how fans receive officiating decisions, it functions as a governance tool. That is a different category from broadcast enhancement.
The 3D avatar system operates on the same logic. The existing semi-automated offside technology makes correct calls. The problem has always been communication — the lines are hard to read, the angles confuse rather than clarify, and fans dispute decisions the system got right. Precise 3D models, generated from one-second scans, produce imagery that is both more accurate and more legible. The system was tested at the FIFA Intercontinental Cup last year, with Flamengo and Pyramids FC players scanned ahead of their match.
Romy Gai, FIFA‘s chief business officer, described the operational challenge of a 48-team, multi-country tournament as one of complexity above all else. The technology announced this week is an answer to that complexity — but the World Cup in 2026 is where it either holds or it doesn’t.
This article is a curated summary based on third-party sources. Source: Read the original article
