The entire notion of full modularity in the design of a vehicle is one of those things that is a great idea in principle but of limited practicability. The chief problem with modularity is that it requires common interfaces, and those interfaces can and will drive design decisions which add weight and complexity in order to provide the flexibility for different applications. As an example, there is in theory no reason every passenger car couldn’t use the same fuel tank, or at least a “modular” tank that mounts in the same way with different volumes based upon the engine fuel consumption and desired range. Practically speaking, that would drive many other aspects of the vehicle design, from suspension to chassis stiffness to the visual styling of the vehicle. So every car design essentially has its own uniquely conformal fuel tank.
The notion of being able to have swappable cells or banks within a battery pack is another thing that may seem desirable in principle but unworkable in practice. In addition to the issue of needing to match performance aptly described by scr4 (which is not some minor pedantic issue; it’s a fundamental necessity), just having to design a battery pack to be capable of being readily serviced in situ is not practical or even desirable. As I noted previously, in many electric cars the structure of the pack that houses the cells forms an inherent part of the vehicle structure because of how large it is, and therefore can’t have large access doors or other service ports. The enclosure also has to be able to resist intrusion by water and other liquids, penetration by road objects, and protect the cells in an impact in order to prevent gross leakage and shorting that could result in a fire (and the electrolyte and matrix materials in modern NiMH and LiPo/LiON batteries is both flammable and toxic), so you don’t want hatches or covers that can pop off or be left open during inexpert servicing. As the cells age, their performance degrades in predictable ways that are tracked by modern battery power management systems; suddenly swapping in new cells or banks is going to disrupt that and result in poor performance and reduced lifespan. And from a liability standpoint as a vehicle or battery manufacturer, I wouldn’t want to permit an owner or uncertified mechanic to swap OEM cells out with poor quality aftermarket or counterfeit cells that could pose a fire or explosion hazard. And even with OEM cells they need to be installed correctly and soldered or mechanically fixed in place to assure safe and reliable functionality.
You’ll note that there are many systems in a vehicle that are serviceable only at the major component level, e.g. you can buy a complete water pump but not individual components like valves or motors within the pump. This is because it is not only easier and cheaper to design a unitary system for initial installation or later servicing, but also because line replaceable single unit systems are inherently more reliable and less prone to being improperly installed or serviced. This would appear to be more wastage–you can end up throwing away a “perfectly good” pump just because of a small crack in a connector or a broken impeller–but the intent is to actually assure the same reliability in a unit replaced in the field as one installed during original assembly.
As a mechanical/aerospace engineer and occasional designer of things that go bang/wham/crunch/chop, I’ve work on a number of systems intended to be modular. Sometimes modularity makes sense, but often it is imposed as a requirement at the top level without real consideration for the significant design compromises to accommodate a theoretical need for modularity that provides little actual benefit or value in practice.
The discussion regarding the unsprung weight of hub motors has already been adequately addressed, so I’ll only make the point that while electric vehicles offer a number of efficiency benefits and potential reduction in service (using a minimal amount of fluids, pumps, valves, flexible belts, and other moving mechanical systems that tend to wear or fail prematurely if not adequately services) they have substantial deficits in many respects to internal combustion engine-powered vehicles, are not immune from environmental damage or degradation from poor service, and are still on the steep part of the technological maturity curve, as witnessed by many of the service issues seen with Tesla and others. That is not to say that it isn’t a welcome technology that is worthy of investment and development, and the essential fungibility in energy source by itself is a significant advantage (e.g. battery electric vehicles can be charged with power developed from coal, gas, sun, wind, nuclear fission, geothermal, genetically enhanced superhampsters) versus internal combustion engines which require a certain class of liquid or compressed gas hydrocarbon fuel, but they’re hardly a panacea to all of the service and safety issues faced with modern internal combustion engines which are a very mature and surprisingly reliable technology despite all of the moving parts and fluids.
Stranger