Everything nowdays includes microcontrollers or microprocessors, and often even in-silico (i.e. as hardware not software) implementations of things like decoders.
However it’s a huge range of those things and the vast majority doesn’t have the processing power and/or hardware petipherals to support “Smart TV” functionality.
For example that upscalling functionality can just be implemented in-silico in a separate chip or as part of the die of a SoC microcontroller for pennies, whilst the actual programmable part doesn’t have anywhere the power or memory needed for the code implementing a fancy UI (such as that for a VOD provider such as Netflix) because that would cost tens or hundreds of dollars more (just go check the price of a standalone TV box).
The economics of the thing nowadays do make it worth it for a TV manufacturer to add the extra hardware needed to make the thing a Smart TV (the kinda crap one only costs maybe $10 - $20 or so more) especially if they can use it to shove adverts in front of people to recoup it or sell “Smart TV” as a premium functionality, but that’s not at all the same as the HW for stuff like hardcoded algorithms such as upscalling being capable of running the software implementing Smart TV functionality.
Your “argument” is built on top of a serious misunderstanding of modern digital systems.