“You know, I’m used to getting laughed at by people. I always laugh last,” says Richard Plugge, general manager of Team Visma-Lease a Bike.
“I’ve seen it many times that a few years later, or even a few months later, people are adapting the same things we do. I’ll give you one example: we don’t shake hands anymore in the Tour de France. We fist bump. When we started that, people were saying: ‘Haha, they’re crazy. But now everyone is doing it.
“Let’s see in two years who is laughing, and who is not.”
Partnerships are commonplace in cycling, a sport which relies on these deals to keep itself alive. But Visma’s latest agreement, with French company Mistral AI, has sparked a storm of interest, intrigue, and, in some places, derision.
According to the team’s press release, the partnership would see Mistral AI, named after the unpredictable winds off France’s Mediterranean coast, “applying its world-class AI expertise to enhance performance, optimize decision-making, and drive competitive advantage”.
But with Team Visma-Lease a Bike being one of cycling’s ‘super-teams’ who won two Grand Tours last year — and who boast star riders including Jonas Vingegaard, Wout Van Aert, and American climbers Matteo Jorgenson and Sepp Kuss — Plugge makes no apologies for attempting innovation.
Patrick Lefevere over AI-partner Visma | Lease a Bike: “Het bullshit-alarm gaat af”:https://t.co/WxWmoMjRGQ
— Raymond Kerckhoffs (@raykerckhoffs) February 28, 2026
Jokes flew on social media when Visma had a nightmare day at Omloop Het Nieuwsblad on February 28 — but were notably quiet when the team’s young British sprinter Matthew Brennan won Kuurne-Brussels-Kuurne one day later.
“Cycling teams are not hundred million pound businesses with huge teams of analysts,” explained Patrick Broe, Visma’s head of strategy, on his Lanterne Rouge podcast. “If you can use (AI) for a performance advantage, it’s something teams have to go for — (for) training, race strategy, pacing, nutrition.
“Every file from every rider has second-by-second data — temperature, altitude, power, heart rate, speed, ventilation — and so the amount of data points on each rider is giant. To leverage that properly, you need AI tools and machine learning.”
To this end, AI use is not totally new in cycling. Equipment manufacturers have used generative AI to develop concepts for a number of years, while several of Visma’s competitors have their own deals. Tadej Pogacar’s UAE Team Emirates, for example, receive data from Emirati firm G42, while Lidl-Trek work with American company ServiceNow.
Visma-Lease a Bike general manager Richard Plugge is confident he will have the last laugh (Kurt Desplenter / Belga Mag / Belga / AFP via Getty Images)
Plugge also raises the example of the “Control Room”, developed with title-sponsors Visma, a software firm, which centralised live race data by using AI software, as an example of its previous use within the team. The Control Room was eventually banned by Tour de France organisers ASO, citing technical regulations.
Visma’s new partnership with Mistral, however, appears to be the most extensive within the peloton. The team has developed a reputation for finding young talent — 20-year-old Brennan is the latest example — with Plugge desperate to defend that advantage.
“We are known as the Moneyball team,” says Plugge. “We started looking at scouting from this perspective around eight years ago — that’s how we found Jonas (Vingegaard) — but in the meantime, the progress and innovation in the world have grown so fast.
“So one of the first things we’re working on is updating our scouting. We think Mistral can play a big role for us in finding the successor of Wout or Jonas.”
How?
“You’re asking a non-technical guy,” he laughs. “Well, of course they need pure power, but there are also other aspects that make a really good rider. I want to know, before I hire a talent, what the probability is of them becoming that really good rider.
“The closer we get to a 100 per cent score, the better for me, because if there’s a misalignment there, we’ve paid a lot of money. Even one euro is a lot of money if it’s a failure.”
Can AI technology help Visma find the next Vingegaard or Van Aert? (Jasper Jacobs/Belga Mag/AFP via Getty Images)
But perhaps the most eyebrow-raising element of the partnership is its potential to influence race tactics. As well as the riders’ own numbers, Mistral’s technology is also able to analyse the pictures of race footage — helping to quantify the turbulent interior of the peloton in a way that challenges the human eye.
“So, we might look at a race like Omloop,” explains Plugge, which had taken place four days earlier. “Of course we’d always do an evaluation, but because there are so many data points, and also bias from me, or a sporting director, or a rider, it’s hard to tell what really happened.
“But from a data perspective, you can often have a better evaluation of a race. So you can make better choices if you can objectify those opinions that we might have.”
Examples include establishing whether to join a certain breakaway break, when to attempt to form echelons and split the peloton in crosswinds, or optimising the perfect timing for an attack.
“You could attack 80km before the finish line, be on the screen and ahead of the game for the next 10km, and even if you don’t win, everybody will say: ‘Oh, you did a great race.’ But was it really worth it?
“There’s a lot we cannot see because it’s happening in the peloton, but we have the data of all the riders, how much they’re pushing in certain sections. Maybe they’ve lost a lot of energy doing that, and although watts per kilo are of course important, how you use them is also important. Using the images and our power data, we can work out the efficiency of that attack, and answer whether they should have done something differently in the future or not.”
In other sports, AI has already played a role in off-pitch factors — but has generally stopped short of genuinely influencing live decision-making. That may change in cycling, where currently, teams have limits on how they are allowed to use the live data they generate, but which does not specifically preclude teams’ abilities to make in-the-moment strategic decisions.
Visma’s record of unearthing talent looks like it has continued with the rise of Matthew Brennan (David Pintens / Belga Mag / Belga / AFP via Getty Images)
At the moment, Visma use Mistral’s output as evidence to help the ultimate decision-maker, rather than the decision-maker in and of itself — but Plugge is open that this structure may change.
“There will be a tipping point moment, probably,” he says. “We are building this, and at some point, it may become a decision-making tool rather than an evaluation tool. It’s a method and a way of thinking — and it could be that there will be a moment we base our decisions on it.”
How have riders reacted to this? That they might question its output seems obvious. There has always been a sense that they are the only people who can truly know what happens in the bunch — but Mistral’s technology claims to challenge that assertion.
“They only get the result, let’s put it that way,” says Plugge. “But today they trust everything. Whatever (head of racing) Grischa (Niermann) says in the race, where he is better informed, and hopefully will have facts as well as opinions, will be from a point where he is better placed to give instructions. That’s better for riders.”
But it is undeniable that the AI boom comes with two competing emotions. Excitement is natural, yet for most, it sits alongside trepidation. Not for Plugge, who raises the film Moneyball, about the Oakland A’s’ embrace of analytics, as an example of how resistance will always be present before adaptation. He laughs at the suggestion that Broe is Visma’s version of Jonah Hill.
“We have to look towards the next step in innovation,” he says. “Patrick started our Moneyball system, but this is building on that, not just in scouting but performance and nutrition. People might see threats — that’s obvious in every innovation. But I’m looking at the positives, how we can use it better, how we can use it in the best possible way.”
