Lidl-Trek and Decathlon CMA CGM adopt AI coaching platform across WorldTour squads

Lidl-Trek and Decathlon CMA CGM adopt AI coaching platform across WorldTour squads

Multiple WorldTour teams have committed to Vekta as core performance infrastructure for 2026, embedding automated data analysis and talent identification across men's, women's and development programmes.

2 min read

Lidl-Trek, Decathlon CMA CGM Team and Team TotalEnergies have adopted Vekta, the AI-powered training and coaching platform, across their WorldTour, ProTour and development programmes for the 2026 season.

The partnerships, announced on February 23, move beyond experimental trials. All three teams have appointed Vekta as a long-term component of their performance workflows, using it for training planning, race preparation and performance review.

They join existing partners Team Jayco AlUla, FDJ–SUEZ and Team AMANI on the platform, which launched in April 2025.

Vekta automates data synthesis, interval detection and session classification, reducing the manual processing load on coaching staff. Its talent identification tools use AI to surface performance markers in junior and development riders.

"The capacity to collect data has increased exponentially in recent years, far beyond what we can realistically analyse," Josu Larrazabal, head of performance at Lidl-Trek, said. "Working with Vekta allows us to spend less time processing data and more time on the human connection with our riders, which is what really makes the difference in the long term."

Paul Barratt, head of innovation at Decathlon CMA CGM Team, said the platform would support data-informed decisions across both WorldTour and development riders.

How the platform works

Vekta uses a Critical Power (CP) framework rather than the Functional Threshold Power (FTP) model common in legacy tools such as TrainingPeaks. CP defines the boundary between sustainable and unsustainable effort, while a companion metric, W-Prime, quantifies the finite capacity a rider has for work above that threshold. In racing terms, that helps predict how many high-intensity efforts a rider can sustain before fatigue sets in.

The platform's AI layer automatically classifies sessions by training stimulus, generates interval summaries and allows coaches to build structured workouts from text prompts, the firm said.

"These aren't trials or experiments, they're long-term commitments to modern, data-driven coaching and athlete development," Paul-Antoine Girard, Vekta's co-founder and CEO, said.

Cover image credit: Lidl-Trek/hardycc

Peter

Peter is the editor of Velora and oversees Velora’s editorial strategy and content standards, bringing nearly 20 years of cycling journalism to the site. He was editor of Cyclingnews from 2022, introducing its digital membership strategy and expanding its content pillars. Before that he was digital editor at Cyclist and then Rouleur having joined Cyclist in 2012 after freelance work for titles including The Times and The Telegraph. He has reported from Grand Tours and WorldTour races, and previously represented Great Britain as a rower.

Never miss a story

Get the latest cycling news, tech reviews, and race analysis delivered to your inbox twice a week.

Continue Reading