Updated: Jun 1, 2020
This time I was triggered by this article in Telecom TV – asking why infrastructure sharing is so difficult. I can confirm that it IS difficult – for a number of reasons, some rational, some more irrational.
I made my first attempt at infrastructure sharing already in the early 90s during the rollout of 2G networks. This was between the two network operators in Norway at the time, i.e. Telenor Mobil and NetCom (today Telia Norway). I was the CTO of NetCom – and my counterpart was my peer in Telenor Mobil. After some initial discussions it soon fell apart – so it never happened – and to date, no similar discussions have been held between the two local business units.
After my three years in NetCom, I moved back to Telenor Group, where I worked for more than two decades – and it is interesting to note that Telenor has made a number of network sharing agreements across the globe, of various types, but not in Norway. Having spent a number of years in the international industry as well, there are some operators that go all the way on infrastructure sharing, but many are still hesitating or strongly objecting.
So why is this so difficult? There are various reasons – and some are quite fundamental in terms of who you are. As a network operator, you get spectrum and a license to operate – and most operators have big network departments. This was the case for me and my peer in the early 90s – and we were in the early phases of network rollout. In other words, we were competing on network coverage (which was the main technical KPI at the time – capacity or bitrates were less relevant in 2G – unlike what we see around 5G currently). Irrespective of G, however, operators still compete on being first to market – at least when the G is new. This applies not only to operators but also to handset suppliers and governments – in a large “pissing contest” around being first. You may also backtrack to several of my previous articles relating to 5G, e.g. this one on “G”s, geopolitics and business – plus one or two others.
So, in the early days of a “G”, there is a first-to-market and coverage competition between operators. After some time, when all parties are in a more stable mode, then the rationality of sharing cost may compensate for the competitive factor. In fact, when nobody has a substantial lead in coverage, then infrastructure sharing might be open for discussion. After all, the ultimate coverage for all will be obtained faster and better. The problem is only that when this time has passed, then the scope for cost sharing might be less.
There are various types of infrastructure sharing, some requiring more cooperation than others – and in some cases, the only reasonable way to do it is to make a separate network and a Joint Venture, with equal control of the JV. More simple ones involve just sharing towers or masts, with both (or all) operators installing network equipment on them – and there are also options in between, e.g. sharing only the RAN – with each operator having its own core. An even simpler form of tower “sharing”, which has been used by most operators ever since 2G (also in Norway) is to lease site access from each other. At least you save time – and you might move Capex into Opex.
A more irrational aspect is also that, with infrastructure sharing, each operator has less control of the network rollout – which involves not only the network department but also the marketing department – as they normally want a say on what coverage to have. In sum, network sharing involves all departments in an operator, i.e. the CTO, the CMO, the CFO – and probably also the Chief Strategy Officer and maybe others. This cross-functional perspective makes it more complicated – and the end result might depend on who is the dominating person – and on how much the CEO or Board drives a decision.
The Telecom TV article also brings up the question of horizontalization and disaggregation of the value chain – as the technology itself is increasingly horizontalized – and since the siloed vertically integrated business model in the telecom market structure makes infrastructure sharing difficult. This comes back to my initial comment on who you are as a network operator. Do you operate a network? Do you sell communications (or other) services? Do you develop your own services in-house? … or through partners? Are you a retail player? … or a wholesale player? … and more.
My view is that every operator needs to be clear on this – and if you are a vertically integrated operator, then network sharing is probably more of a challenge – but it can be done! In fact, I was personally involved in an attempt to split Telenor Mobil in Norway into a network operator part and a service provider part in the late 90s – which eventually failed due to a strong wish to maintain a vertically integrated operator at the time. My view is also that, in particular for 5G, the whole ecosystem will be significantly changed (see also my earlier article on “who will eventually benefit from 5G”. The point raised on disaggregation is therefore a very pertinent one. I foresee that, within five years’ time, the industry structure around 5G will be very different from today.
So much for the operator perspective on this … If you are a regulator, you will most likely consider how to provide the best services for the general public, how much competition you want in the market, etc. Regulators want as much coverage, capacity and broadband as possible – and as fast as possible -– and they need a regulatory regime that maximizes investments to this effect. They also need to avoid collusion between players in the market. This is when it gets difficult. Operators need payback on their investments in 5G. This means they need to save cost as far as possible – so infrastructure sharing should be an obvious area to look into. Operators will also hesitate to build out non-profitable areas – and may in many cases want to see a private-public partnership, i.e. public money in some form, from the governments.
Some governments have suggested single national wholesale networks, which I would not recommend – simply because they would never be properly funded, as they most likely will have to be funded through national budgets – and there would be no competitive pressure to make the networks as good as the end customers want. A reference case for this could be the example of the failed TETRA networks in Europe for public services. Infrastructure sharing should be based on voluntary commercial agreements. There might be cases in countries where the regulator might need to ensure equal opportunities for all players though.
To sum up: Operators need to overcome their resistance to infrastructure sharing – and they need to consider it early. For 5G, it will most likely be a commercial necessity, especially with mmWave spectrum. Regulators, on their side, should encourage investment and be prepared to chip in on the financial side – if they want extensive 5G broadband in their countries. Regulators should not impose infrastructure sharing though!