Robert Kühnen
· 02.08.2026
The second stage covers 148 kilometres from Aigle to Geneva and has slightly more elevation gain than the first, but the hills are concentrated in the first third of the route and the final section is flat. This means that the sprinters’ teams have a good chance of controlling the race. A classic sprint finish is therefore the most likely scenario.
When it comes to cycling kit, ambitious female riders will be pushing their aero bikes to the start line and, of course, kitting themselves out accordingly – from aero socks to an aero helmet.
Even breakaway riders are best served by this type of bike. Riders who turn up at the start on a mountain bike are signalling that they couldn’t care less about this stage.
The run-up to the finish line is straight. 310 metres before the finish, the road curves slightly to the right. The home straight is seven metres wide.
Which team has the best bike on the start line today? We’re simulating a 250-metre sprint. Result: Team Visma is in an excellent position. The fast and lightweight Cervélo S5 has the lead in the sprint. The slower Cervélo R5 crosses the finish line 2/10 of a second behind – that’s a gap of almost four metres!
So details are important. That makes it all the more exciting to see what the riders actually ride.
An overview of the (almost) full line-up:
The table shows that aero bikes reign supreme. The fastest bikes in the sprint are aero bikes. No wonder, as the predicted top speed is 63 km/h. The second sorting criterion is weight. The Ridley Noah Fast, which is 900 g heavier, finishes 15 thousandths of a second behind the Cervélo S5.
Based on our own wind tunnel tests, we carry out simulation calculations for the Tour de France tech briefing. How TOUR tests: Aero road bike test in the wind tunnel.
We are investigating which wheels can offer a technical advantage in which situations. The variables we can control in the simulation include wheel weight, rider weight, the inertia of the wheels, the drag coefficient, the rolling resistance coefficient and the efficiency of the drivetrain.
To model ride times, we use realistic power outputs and rider weights, combine these with our wind tunnel data, and have the riders race virtually along selected sections of the route, which we extract from the official route data; the derived elevation profiles are key to this. The modelling also includes bends, which we can brake for realistically, and adjustable power profiles for different types of riders. This allows us to distinguish between hill climbs and proper final sprints. Taken together, this makes the simulation very realistic. What we cannot replicate are dynamic handling effects such as the individual behaviour of the wheels on different surfaces.
The journey times calculated for the sections of the route that are decisive for the race highlight the influence of the wheels – provided that the riders always behave in the same way in a given scenario.

Editor