However, data hikes will not just be driven by consumers. Firms also predict a rise in so-called machine-to-machine (M2M) communication, that will connect the mobile networks to an array of inanimate objects – from bins that will signal when they are full to electricity meters that will constantly call in to the utility company.
By the end of this year, Cisco predicts that the number of mobile-connected devices will exceed the number of people on earth, and by 2017 there will be more than 10 billion.
No wonder the chairman of the US Federal Communications Commission recently declared: "The clock is ticking on our mobile future.
The illusion is that the airwaves, like the atmosphere they pass through, are effectively limitless. We can’t see them, they can travel in any direction and link any two points – why should they be limited? Yet, in practice they are as hemmed in as a motorway through a city.
Radio spectrum is a limited resource, strictly farmed out by national and international regulation. At the moment it is all spoken for by the military, mariners, aviation, broadcasters and many more – all the way up to the very extreme of useful frequencies at 300 gigahertz.
No-one can get more bandwidth without someone else losing out. The 4G spectrum auction that recently began in the UK, for example, is the equivalent of adding a new six-lane motorway to the existing wireless infrastructure (itself already running at 10-lanes), built on virtual land vacated by old-fashion TV broadcasts.
It helps, but will only keep the expansion going for a certain time. Which is why mobile operators, and their rivals, are gearing up for major spectrum negotiations at the International Telecommunications Union in 2015. The so-called WRC-2015 conference aims to carve up the available spectrum amongst different competing uses. But an overriding priority is identifying and allocating additional frequencies to mobile services.
Already, the stakeholders are preparing their positions. Ericsson's Afif Osseiran, project coordinator for the European consortium Metis, says the ITU conference “will be a crucial moment for laying out the spectrum needs for the 2020s.”
But industry will not just rely on these delicate negotiations to secure its future. Much of the advance in the past 20 years has not been about how many of these wireless “lanes” we have, but how efficiently we use them.
Like a newly built motorway that’s used by just a few cars, the first generation of phones were incredibly wasteful of the spectrum they used. Capacity was wasted in the same way as the gaps between vehicles represented lost transport opportunities.
In going from 1G to 2G, there was a 1,000-fold increase in capacity, mostly not because of the new radio lanes added in, but because more traffic was squeezed onto those lanes.
And in going from 2G to 3G, capacity rose another factor of 1,000: digital techniques managed to squeeze out yet more of the empty space.
But with the latest generation of tricks being rolled out in 4G (actually described as 3G Long Term Evolution by developers), the industry is running out of ways to improve the efficiency further.
These limits that determine how much information can be transmitted were established in the 1940s by the American engineer Claude Shannon. Although his employers, the Bell Labs of AT&T telephone, were interested primarily about the limitations of telephone wires, Shannon's equations can be used equally for radio transmissions.
And mobile experts generally accept that the limits to data flow revealed by Shannon’s formulae are close to being reached.