Imho you're missing the point, or at least, you've stated a valid conclusion but the point lies below.
Sure, by tightly integrating you are getting material benefits (size, consumption, optimization, etc). But by doing this, you go further down into high-tech, which means complexity and energy intensive build/repair/dispose [1]. At this point i'm reformulating your tradeoff, in which you choose the material benefits.
Now, one part of the right-to-repair movement comes from "low-tech" underpinning and objects the material benefits. Indeed, the point of a smartphone is not to crunch some machine instructions, it's to communicate with friends/family/world, interact with network services etc. Crunching code is just a byproduct of reaching this goal. It's even the same with supercomputers which one could think as performance-oriented: their goal is not to do the peta-flops, it's to solve some simulations at a given precision level. Nothing is built just for fun (or should): if you build a tool you better have a problem. And if you forget the goal, your tool will most likely be garbage. This is the objection of the r2r/lowtech movement: it's an assertion that a lot of high-tech has lost the sight of it's actual use and hence is hindering progress more than anything else.
This is a direct opposition with the pov which is widespread in modern western tech thinking that "if you build the tool they will use it". Obviously this pov has some depth to it, it's not completely nonsense (in the sense that creative and explorative processes are useful, i know i'm myself doing research in foundations of math). But this pov is also very extreme, and is in fact a corner-stone of the offer-driven production that is hegemonic today (with market economies). This same offer-driven production is also posing many social problems because of the need for aggressive marketing, problems with long-term stable availability of basic goods, with the focus on big/fast/shiny. We've gone too far into that direction and that a part of the reason for the pollution and resource exhaustion crisis (being themselves big parts of the climate and biosphere crisis going on rn).
The position of r2r/lowtech is obviously not to go to the other extreme which would be completely planned production based on somehow knowing what needs to be done (which we already see cannot really work and is very conservative by nature). But there has to be a balance. My position: the bulk of the energy-intensive and stable mass-manufacturing of basic known-to-be-useful-and-desirable goods should be done by planning, with open standards, user-groups, cheap repair, and some fraction of the resources should be reserved to the free-for-all market-driven exploratory research. Obviously this poses a couple problem with the fat cats running mass-manufacturing and other commodity services.
Hope this was a balanced explanation of the position.
[1] The matter will be ordered on a smaller scale so: (1) to build it you need to put a lot of energy into putting everything in order (entropy!); (2) to repair you both have to diagnose a complex thing and then act on it precisely; (3) to dispose of properly you need to put a lot of energy into separating everything again (entropy again!). Here i say "energy" in a large sense, this energy might be embodied as direct energy put into the process but also as human time, dependency on high-tech tooling (hence cost), etc. Another concept linked to this is "capital intensity".