Asking AI to predict the 2026 World Cup winner is pretty much like walking into your local sports bar to solicit opinions from human soccer fans. AI models are not uniform in their predictions except in one regard: Kylian Mbappé of France will win the Golden Boot as the tournament’s top scorer.
There’s big money involved in predicting the FIFA World Cup winner. MarketWatch expects a record $60 billion to be bet on the games, a reflection perhaps of the competition’s expansion. The number of teams involved in this year’s World Cup expanded from 32 to 48 nations. There will be a total of 104 matches played across venues in Canada, Mexico and the United States. The initial group stage divides the 48 teams into 12 groups of four with the top two teams in each group advancing. The Round of 32 sees those top two teams plus the eight best third-place teams pitted against each other. The knockout stage is the win-or-go-home round that will yield a champion after 39 days of play.
Like humans, AI models appear to be influenced by past performances. During its 96-year history, only eight nations have won the World Cup title. They are: Argentina, Brazil, England, France, Germany, Italy, Spain and Uruguay. An old World Cup truism is that while individual matches are unpredictable, the champion usually isn’t.
That said, AI models are not aligned when it comes to picking a World Cup winner. Like human prognosticators, AI models generally favor France, Spain, and Argentina to be there at the end, with France given the edge by some AI models for the final win. But there is plenty of disagreement. One AI model expects Brazil to be in the final while another says the team ultimately will be a big disappointment. A number of AI models give the Netherlands, Portugal and Morocco a chance. AI analysis may also only be as good as the quality of the prompt.
One thing that is clear is how heavily AI models are being interrogated. For example, an AI analysis by the World Sport Network predicted the USA atop Group D in 209 of 210 simulations and being able to progress beyond the group stage. The USA made it into the semifinals in 23 projections but only two saw the team in the final and none predicted the USA as the winner. The AI analysis puts the USA in the category of dark horse contender.
One caveat for bettors may be AI’s ability to predict scores. Claude, for example, had a six-game winning streak in the group stages but was woeful when it came to predicting the score, getting just a single 1-1 draw correct.
There’s no doubt World Cup fans are using AI for guidance in their World Cup deliberations. But at the end of the day, maybe only one thing is certain: anything can happen. As of June 19, Argentina legend Lionel Messi has scored three goals to Mbappé’s two.
