It sounds not just infeasible, but comically infeasible. For the sake of argument, let's assume that it's easy to improve ICEs in general by 5%. You still have to apply that upgrade to all the existing engine designs, ideally as drop-in replacements, and then swap all those engines in the wild.
I was rather thinking of buying a tree seed that magically absolved you of a few tons of CO2, never mind that most don't grow to maturity, that it takes a century, and that after that century it falls over and releases that CO2 again. With that kind of math, you can make jet flights fully offset for something like 2% added ticket cost.
Just look at the greenhouse gas compensation airliners offer for an example of what Shell could reasonably have been up to.
I may also have had a classmate that worked at Shell and put on a really good show of how they're not all evil but also spending <insert objectively huge number> on good things. The last strand of that is severed now at least...
I assume the biggest polluters are not cars but big industrial operations or shipping. That seems like a smaller perhaps even realistic number of sites to target for some amount of efficiency gains. Especially if margins are too low in these industries for them to invest in their own efficiency.
