Isn't the design intent of the 5G methodology more about maintaing useful throughout to individual devices in the presence of an increasing number of other devices (and other interference), than it is about massive throughput to single devices, in preparation for a predicted explosion of wireless IoT variable things? If so, then this result isn't a problem, we've just been primed to expect more by headline grabbing numbers coming out of artificial benchmarking experiments, that were never expected to be reproduced in the messy real world.