The study (which is linked in the BBC article if anyone wants to read it) says that there is a relationship between time and total energy expenditure. The shorter the time, the more energy can be used. At long durations (multiple weeks), it flattens out at about 2.5× BMR. Most of the data in the study comes from RAUSA runners, with estimates about other endurance events. The three weeks of grand tour are not sufficient to reach the flattening out stage.
I agree it would be nice to include estimates from Mark Beaumont’s or Amanda Coker’s records, but the trendline they have from their other events seems pretty robust, so I wouldn’t expect any surprises. I seem to remember looking at one of Amanda’s rides and being surprised by the low heart rate, which would be an indication of a fairly low energy expenditure.
The interesting bit is that the authors link this limit to the amount of food that can be turned into energy, long term. In other mammals, the limit is usually due to heat dissapation, but this didn’t seem to be the case with humans/the activities considered. The other interesting bit is that it is just slightly higher than the TTE during pregnancy, which might be another evolutionary driver for this limit.