Growing bandwidth demand is a genuine "problem" for access providers in one clear sense. If all access providers can sell is best effort, unlimited access, growing access is a problem, as costs grow, but revenues do not. Streaming video is the big driver of bandwidth consumption for most users, which is why everybody is hearing so much talk about the end of "unlimited access."
What also is obvious is that the problem actually is an opportunity for access providers, so long as access providers are able to charge for usage in some logical way. Absolute "metering," as with electricity or water, never has been popular with end users, and arguably depresses application usage.
Such notions are not reassuring for some other participants in the ecosystem, of course. Users will not want to pay more. Application providers understandably are worried about whether they will wind up paying access providers in some way, either for quality of service mechanisms or some other form of access tolls. Perhaps the bigger issue is potential abuse of market power, rather than pricing that is linked in some logical way to consumption.
But usage traditionally has been an important input for communications service pricing and packaging. Some might argue that the simplest, most logical way for access providers to participate in application system revenue growth is to simply tie retail access pricing in some way to expected growth of bandwidth usage.