This is an excellent question but a massive one to answer. Probably the length of a scientific paper. It'd involve thermo, engineering economics, etc.
To keep it short, yes there is a return to scale, but it's a diminishing one. Gas turbines run at about 55% thermal efficiency. Large of small it matters little, that limit is set by the blade materials' melting temperature (which sets longevity).
Your car typically runs at about 30%, its efficiency partially offset by heat losses to the cylinder walls (ie the larger the better), but is mostly set by engineering decisions other than fuel efficiency (one big cylinder has less surface heat loss than four little guys but would be unbalanced)
But there are so many legacy design decisions in an ICE that no longer apply if we have hybrid drivetrains and ammonia/urea injection (to mitigate NOx from high compression).
The Prius challenged a lot (but not all) of these decisions and remains, in my opinion, the most revolutionary car of the past 50 years.
I really believe that an fossil fuel car engine can get an efficiency within the transmission losses of the best gas power plant. But even if ICE development were frozen, hybrids still make more sense, from a CO2 POV, than EVs