[aerohead]

Example:
2012 Tesla Model S
* original Cd 0.26
* frontal area= 25.8333-sq-ft ( 2.4 meters -squared )
* 4,197- pounds, EPA test weight
-------------------------------------------------------------------------------------
* @ 155-mph, aerodynamic load is 170.7366- hp ( top speed )
* @ 155- mph, rolling-resistance is estimated 26.8693- hp
* @ 115- mph, total road-load horsepower is est. 197.6059-hp
--------------------------------------------------------------------------------------
* @ 85-mph, aero load is 28.1571-hp ( fastest legal speed in USA )
* @ 85-mph, R-R is estimated 14.7348-hp
* @ 85-mph, total road-load horsepower is est. 42.8919- hp.
--------------------------------------------------------------------------------------
* @ top speed, the 'engine' power requirement is 460% higher than at maximum legal speed.
--------------------------------------------------------------------------------------
* @ top speed, 'engine' heat flux is 460% higher than at maximum legal speed.
--------------------------------------------------------------------------------------
* @ top speed, 'engine' and battery cooling load is 460% higher than at maximum legal speed.
--------------------------------------------------------------------------------------
Had the Tesla Model S been specified for a lower top speed, it follows that, if it requires an 8%-drag cooling system for 155-mph, the Model S could easily have an adequate cooling system, with a fraction of 8% drag.
Ditto for Porsche's Taycan Turbo, and Turbo S, of 161-mph top speed.
------------------------------------------------------------------------------------
And since modern cars have ' multi-stage active cooling management of cooling...' , one would want to assign a 'spectrum' of cooling system drag, rather than a fixed value, like 8%.
--------------------------------------------------------------------------------------
Additionally, as already mentioned by members, back-to-back comparisons between ICE and BEV variants of the same vehicle, offer the greatest insight into the difference in drag, including cooling drag, between any models of interest.
--------------------------------------------------------------------------------------
Finally, I'll leave 33.333% of the argument unmentioned for the time being. Just to see if any of the 'experts' pick up on it ( be careful who you trust ).