There is no doubt that talk about 5G is still mostly hype. That’s not to say the technology isn’t being developed or worked on, but in practice no one has yet figured out a way to make money by deploying it, which has got to be the key starting point of any rational assessment.
Faced with that fact you can either ignore the topic entirely until it becomes a lot more concrete, or you can assume that businesses, regulators and governments don't always make rational decisions and that therefore some things will happen that may have deleterious effects on the market even though actual instances of 5G deployment will be few and far between.
One such issue, and one that is largely ignored by both operators and policy makers, is the threat to the mobile competition model that 5G represents. It's something I've already talked about in public, but haven't written about much, and I'd like to expound on that here.
Successive generations of mobile technology have always delivered higher speeds and lower latency. But the trade off was that you needed the cellsites to be closer and closer to the end users to get those benefits. 4G saw an unprecedented densification of cellsites, and while a lot of this was accompanied by fiber backhaul to these same cellsites, the proportion of fibered cellsites is still low in many markets. While 5G also opens up some lower spectrum bands, the rest of this piece assumes that the significant increase in performance will be in the higher spectrum bands and therefore require further cell densification.
These small cells will need to be connected to fiber, not only because that's how end users will get the most out of them performance wise, but because backhaul will no longer be the only game in town : other approaches like fronhauling will allow significant savings in the intelligence required at the cell site at the cost of much more bandwidth required to carry the traffic back to an edge facility that will process it. In other words, efficient small cell 5G technology cannot be dissociated from the ability to connect said cells to fiber.
And therein lies the competitive threat: in most countries, that underlying fiber network either does not exist (in countries where even FTTC hasn't been widely deployed) or is owned by a single player (usually the incumbent) who happens to be both a fixed and a mobile operator.
This is how things are likely to play out :
- 5G will first be launched in dense urban areas as an upgrade to 4G macro-cells. All existing market players will be able to participate, but the actual impact on customer experience will be marginal. The marketing will promote the hell out of 5G however until it becomes a household name and people outside of 5G coverage areas will want it.
- The incumbent will gradually roll-out small cells in less dense areas (still urban, but not so dense). It will reuse existing fiber deployed for FTTC or (ideally) FTTH and therefore shave considerable costs on its deployment. It will market 5G in those areas.
- Competitors who don't own similar underlying fiber aggregation network will be hard pressed to keep up. They will either have to invest massively to build their own fiber, or will fall behind. The former will hurt their profitability, the latter is much more likely.
- Gradually the incumbent will gain market share to a point where possibly one of its competitors will no longer have a viable market position. Then the universal paradigm of "three mobile players can compete and thrive in any given market" will die.
- The incumbent will then stop deploying 5G in markets where no one is likely to compete, which will widen the existing digital divide.
In a nutshell, free market 5G, if left to develop with no regulatory intervention, will lead to diminished competition in the mobile market, the reinforcement of the market power of the incumbent, and a widened digital divide.
There are solutions to the problem, I think. At least I can envisage a few and we have one example of a policy maker taking a pro-active stance (although details are hard to come by). I will detail these in the second installment of this two part post in a few days.