I actually have the Mörbylånga table at home and I find it very sturdy. One thing which the pictures doesn't show is that there are two supporting beams under the table, which provides the necessary strength to not collapse on first touch. Obviously, I have not done the actual test, but I will try to remember and report back if the table ever breaks.
I attribute it to the speed test I server which is picked up - it is not the ISP provided one, so I assume there will be some variation of the measured bandwidth. Manual tests against the ISP provided instance is much closer to 10gbit, but I haven't bothered to automate it yet to graph it.
On other hand, my understanding is that fiber is not immune from congestion problems - the only difference is at what point it happens.