My initial reaction was "this cannot work". So I looked at their website, which is mostly puffery and other flowery language. But to their credit, they've got two ~~studies, err papers, err preprints, uh~~ PDFs, one of which describes their validation of their product against wind tunnel results.

In brief, the theory of operation is that there's a force sensor at each part where the rider meets the bike: handlebars, saddle, and pedals. Because Newton's Third Law of Motion requires that aerodynamic forces on the rider must be fully transfered to the bike -- or else the rider is separating from the bike -- the forces on these sensors will total to the overall aerodynamic forces acting on the rider.

From a theoretical perspective, this is actually sound, and would detect aero forces from any direction, regardless of if it's caused by clothes (eg a hoodie flailing in the air) or a cross-wind. It does require an assumption that the rider not contact any other parts of the bike, which is reasonable for racing bikes.

But the practical issue is that while aero forces are totalized with this method, it provides zero insight into where the forces are being generated from. This makes it hard to determine what rider position will optimize airflow for a given condition. To make aero improvements like this becomes a game of guess-and-check. Whereas in a wind tunnel, identifying zones of turbulent air is fairly easy, using -- among other things -- smoke to see how the air travels around the rider. The magnitude of the turbulent regions can then be quantified individually, which helps paint a picture of where improvements can be made.

For that reason alone, this is not at all a "wind tunnel killer". It can certainly still find use, since it yields in-field measurements that can complement laboratory data. Though I'm skeptical about how a rider would even respond if given real-time info about their body's current aerodynamic drag. Should they start tacking side to side? Tuck further in?

More data can be useful, but one of the unfortunate trends from the Big Data explosion is the assumption that more data is always useful. If that were true, everyone would always be advised to undergo every preventative medical diagnostics annually, irrespective of risk. Whereas the current reality is that overdiagnosis is a real problem now precisely because some doctors and patients are caught in that false assumption.

My conclusion: technically feasible but seems gimmicky.

For 3k pounds.... id rather go to a wind tunnel

Ps you can test aerodynamics with golden cheetah's aerolab if you take accurate weather measurements. Its not going to be as accurate as a wind tunnel but it can help with positions or different kits.