It is fashionable to talk about the race to 5G. This theme started last year, and has picked up this year with many opinion articles published on the topic. Perhaps the one that got most attention is the white paper by Deloitte: The chance to lead for a decade. The paper compared US networks and investments to that of other countries, with a focus on China. Of the several comparisons made, one in particular got me thinking: can we use the number of cell sites to qualify who’s winning or losing a race? Never mind whether or not the whole notion of a 5G race is relevant (I’ll save this for another post), but is the number of sites, which is a proxy to coverage and capacity indicators, a legitimate metric? In this, I don’t mean to pick on one element of the argument. This question is more fundamental: should coverage and capacity remain the sole indicators for wireless services?
China ahead on number of sites – and that’s that!
According to the Deloitte, China has 14.1 sites per 10,000 people vs. the US’ 4.7 (3x). Alternatively, China has 5.3 sites/square miles vs. the US’ 0.4 (13x). At face value, this is a big disparity, but one that’s well known. The larger of the US carriers like AT&T and Verizon have on the order of 60,000 cell sites, each. That pales in comparison with China Mobile’s 500,000+ LTE sites alone!
Need to consider spectrum
But that does not tell the entire story. There are many nuances that weaken such comparisons. For instance, spectrum allocations must be considered. Deloitte did not make any discussion of spectrum.
China has allocated a total of 635 MHz of spectrum for mobile services. Chinese operators use a mix of technologies: LTE, TD-SCDMA, CDMA, GSM and WCDMA technologies. LTE operates in the high frequency bands primarily 2.3 and 2.5 GHz where a total of 190 MHz is operational (30% of allocated spectrum).
Compare this to the US where 778 MHz of spectrum is allocated to mobile services (not including any CBRS spectrum). This is 23% more spectrum in the US than in China. Moreover, US networks are quickly migrating to LTE as operators turn off 2G GSM and CDMA and refarm 3G spectrum to LTE services. This provides an advantage in spectral efficiency. US LTE networks operate in frequency bands between 700 and 2100 MHz providing coverage advantage over networks in 2.3 and 2.5 GHz which have to rely on more sophisticated antenna systems to achieve the desired coverage (and hence are more expensive).
The population factor
Now consider that China has 4x the population of the US (1.38 billion vs. 325 million) which Deloitte factors into the argument. However, China’s urban population stands at 57% in comparison to 81% in the US. With such high rural population, China needs to build more towers to meet its coverage objectives. This is not the case in the US where only 19% of the population lives in rural areas. Since both countries have similar land mass (China is only 2.5% bigger than the US), China needs to build more towers. The limited availability of alternative infrastructure in China even increases the pressure to build cell towers.
These are additional factors to consider which add up to require China build more sites. But the point I’m try to make is that we cannot use number of sites, and investment into site construction as a sign on who’s winning or losing the ‘5G race’. I actually think we need to move beyond that to create new metrics.
Need new metrics!
The industry uses coverage and capacity to set comparative benchmarks. Operators use these metrics for marketing while regulators use them to gauge the state of the industry (e.g. broadband defined as > 25Mbps, minimum coverage requirements, etc.).
But increasingly coverage and capacity will miss the point on assessing the state of wireless. Mobile service reached saturation in both coverage and number of subscribers. Moreover, differentiation in 5G will be in the core network on which services will be built. Virtualization is a cornerstone technology that will differentiate service providers. How different service providers play that card will be one to watch.
In short, where it comes to 5G, we need to focus on different metrics. This could be problematic as coverage and capacity are easy to measure and understand. I have discussed this in a recent article with my colleague Peter Lyons of the World Economic Forum: 5G mobile is nearly here – but we should share networks to make it affordable.
I will expand more on this in future articles as the question implications on how operators provide wireless services.