It is critical to the near term 5G business case that total cost of ownership is reduced in as many ways as possible. This is particularly true in markets where key expected drivers of new 5G revenues will not emerge at scale for some years, particularly the growth of enterprise services that will underpin operational transformation in many sectors – but not necessarily for some years to come.
Where additional revenues are not easy to find, but operators are still under competitive or regulatory pressure to build out 5G rapidly, clearly a new cost model is needed.
Over the next decade, the move to cloud-native networks, and the transformation of many telco processes, will help to deliver that new model, but in the near term, cloudification represents new cost and disruption for digitally advanced operators, and a future aspiration for the huge majority. By far the biggest contributor to reducing 5G TCO in the crucial next few years – before new revenues and cloud-based platforms emerge – will be increased sharing of network assets, from towers to data centers to active RAN.
Of course, sharing of passive infrastructure such as towers, and increasingly of fiber for x-haul and, in some markets, power systems, is already commonplace. But the commercial landscape is changing rapidly, especially in Europe. This driven by a combination of operators seeking to unlock the value of their site assets through carve-outs and IPOs, at a time when infrastructure stocks are highly rated; with the geographical and business expansion of several large pure-play towercos, such as Cellnex and American Tower.
These giants are driving scale into the market, which could reduce costs for operators as they rely more on third party infrastructure, but the telcos may also face a decline in supply chain competition almost as serious as they face among their RAN vendors.
Meanwhile, asset sharing must extend beyond sites and x-haul if it is to make 5G cost-viable for many operators. For operators that are starting to invest in the cloud that will support their virtualized networks in future, there is a dilemma over whether to build out that distributed cloud themselves (and monetize it where enterprise opportunities are strong), or rely on a shared or third party platform. Hyperscalers are the obvious partners for centralized clouds, and some operators are prepared to risk placing their network functions, not just their IT functions, in the public cloud, as Dish’s recent agreement with AWS highlighted.
But closer to the edge, where many operators will need to build to support 5G RAN, the partners may be different, and more specialized in building and running a cloud to support challenging telco functions such as real time 5G vRAN processes. Here, some towercos see a growing opportunity, though they will compete with hyperscalers and with enterprise edge providers.
Then there is the active network itself. Regulators are still wary of too much fiber access sharing, at least without strong open access rules, as Italy’s latest developments show. But they are warming to the idea of shared mobile networks, recognizing that for three or four MNOs each to build a separate 5G platform puts significant pressure on 5G economics and on access to infrastructure such as small cell sites, and this can delay or constrain the launch of 5G services.
Some governments aim to drive or manage the process themselves, though these attempts have rarely resulted in great success, as 4G era projects in South Africa, Russia, Kenya, Mexico and others have shown. Undeterred, however, there is a new plan for a shared 5G network in Russia, and Malaysia is also pursuing its plan for a single platform accessible to all the MNOs.
There is likely to be greater commercial success, and cost savings, from mutual RAN sharing agreements agreed by MNO partners, without intervention from government. RAN sharing will be fundamental to the 5G business case in countries where it is already established, from the UK to China, and to markets where it is becoming more attractive both to operators and regulators. The big question that follows from that is whether the towercos will expand their business into active networks?
Just as tower sharing deals between two or more telcos evolved, in some markets, into pure-play wholesale towercos, so the same could happen in the RAN, with active network sharing between two or more operators morphing into a model based on a neutral host wholesale RAN.
Cloudification should make active network sharing easier, and enable a single entity to manage RANs on behalf of several operators, from common infrastructure, with techniques such as slicing. That entity could be one of the operators, or a joint venture; a vendor such as Ericsson, extending its managed services business; a hyperscaler; or a towerco.
Neutral host RANs, as opposed to multi-operator arrangements between telcos, are likely to be seen first in small cells and enterprise or indoor networks, which are largely greenfield and where operators have often struggled to make the business case work. But it is not inconceivable that 5G macro RANs will, in future, be co-developed and shared, or even that they may be deployed and run by Cellnex or American Tower.