I can't speak for Mercedes and Tesla, and these observations are according to feel and anecdotal, but I paid close attention: I had a few junker Volvo's, a VW Jetta, and a Honda Civic, that I used to strap as much lumber and construction materials as I could fit on top (with straps that were specified for many times the weight, although only routed and tightened appropriately) without rubbing the tires pulling out of the hardware store, often with as many sacks of concrete inside as it could take, distributed evenly. (Yes, the Volvo's won the competition, but that's not at all my point.) My point is that I noticed over time that certain particularly heavy loads that did not make the tires themselves pop or rub anything nonetheless did cause normal handling of the vehicle unladen change before and after the heavily laden trip to the point where I am convinced the car was damaged by the amount of load. However, across multiple vehicles and load amounts, I also noticed that this point of excess was directly related to how long it was loaded, and how hard it was handled along its courses, AND MOST importantly I THINK (not sure here) it also had a steep curve of damage that was SOMEWHERE BETWEEN the max loading specification of the vehicle on the placard and the point at which the tires started rubbing, that if I stayed under this steep part of the curve, then the damage did not accrue quickly. Once I learned this point, I did some suspension and axle repairs, and got immensely more cargo tonnage throughput with almost zero problems, except the eventual breakdown of the transmissions.
In my opinion and experience, overall, vehicles not designed to carry more weight than specified WILL SUFFER WITH EXCESS WEIGHT. The key question is WHAT ARE THE ENGINEERING SPECIFICATIONS for what its maximum would be given very slight enhancements, and WHAT ARE those precise enhancements? I think there are so MANY evenly balanced parts with respect to weight capacity that that question most likely should be answered by well integrated engineers specific to the task of answering that question. I'm not saying that it's an intractable question, only that it's quite a thing to presume any particular answer or set of answers without expensive knowledge. That's just my very limited opinion and experience.
Having said that, it would be fun to see the numbers output of what parts would need to be modified and what those particular modifications would do with respect to capacity, after such extensive engineering. OTOH, that's what heavy cargo vehicles are for. I actually had dreams of using Tesla parts for a truck until I realized trucks have so many efficiency improvements over cars for heavy cargo that the existing companies already outfitting trucks with EV would probably be better suited to that task. There went my dreams.
Speaking of which, heavy is not always better. Garbage trucks ruin residential streets 4x faster than they should be ruined. Instead, a heavy semi-truck-with-trailer that parked in one spot per zone on one major heavy weight road with a bunch of automated car-weight robot cars that individually went down small areas (say, one half block each trip per car) to collect a smaller net weight of cargo (refuse) on each smaller car, which then returned to the mother ship to reintegrate the results (refuse) into the larger weight vehicle, would reduce the miles*weight on residential streets, greatly reducing wear and tear, as well as noise. This is exactly one idea I had for Tesla-type cargo carriage once the autonomous driving methods worked, but the robotic arm to collect the "cans" would need to be a light-weight one unlike the very heavy ones on regular garbage trucks. Take off the human-occupant part of a Tesla, and the resulting weight limits of even a car would probably be exactly the right amount for this type of distributed collection system. Even with the increased cost of complexity and parts, the savings on roadways alone would probably be way more than enough to offset the capital for that cargo system. I haven't done the calculations recently so I don't know if it's actually true, but it seemed like a big savings when I researched the numbers a year ago on this topic.