LONDON -- 5G World -- Uncertainty hovers above 5G like a swarm of fleas over a mangy dog, and perhaps the greatest uncertainty of all is how much the whole thing will cost.
A few numbers have been bandied around. A Barclays report last year estimated that blanketing the US with a 5G network would cost about $300 billion. Deutsche Telekom AG (NYSE: DT) CEO Timotheus Höttges says doing the same across the whole of Europe will require investments of between €300 billion ($335 billion) and €500 billion ($558 billion). But Ovum Ltd. analyst Daryl Schoolar thinks T-Mobile US Inc. may be facing a bill of about $25 billion to construct a nationwide 5G network. And a 5G action plan from the European Commission reckons 5G upgrades across the region will cost about €57 billion ($64 billion), according to the Financial Times. (See DT Plots 5G Across Entire Footprint andT-Mobile 5G Plan Could Drive Capex to Record Highs.)
This diversity of opinions is partly down to spectrum. The Barclays estimate seems based on the notion of a nationwide 28GHz network, which would require lots of additional mobile sites and will clearly never happen. T-Mobile plans to use a range of frequency bands, including 600MHz airwaves that would let it cover wide areas with relatively little equipment (simply because signals travel much further in these ranges). (See Is T-Mobile's 5G Plan Just a Pipe Dream?)
But even with the certainty that Europe's operators will focus on using mid-band (3.5GHz) and low-band (700MHz) spectrum to support 5G, there is a surprising amount of cluelessness just two or three years in advance of 5G rollout. (See DT, EE, Orange, Vodafone Open Up on 5G.)
France's Orange (NYSE: FTE), for instance, expects capital expenditure to peak in 2018 or 2019 as it makes preparations for a 5G launch. Yet chief financial officer Ramon Fernandez admits that Orange stills lacks the "full picture" on 5G spending. (See Orange Sees 'Peak' Capex in 2018/19, Ups 2017 Guidance.)
Software could have a massive impact. If one sensibly assumes that deployments based on millimeter wave spectrum (very high frequencies) will be limited to hotspots like airports and shopping malls, then many operators may already have much of the physical infrastructure in place. The growing software "upgradeability" of today's network technologies could make the transition from 4G to 5G much more economical than previous hardware-based overhauls, agrees Hossein Moiin, the chief technology officer of Finnish equipment vendor Nokia Corp. (NYSE: NOK). (See 5G Spectrum to Cost Less Than 4G, Says Expert.)
Then again, the "active antennas" that come with 5G technology are likely to be very expensive, says Bengt Nordström, the CEO of the Northstream market research and consulting business. And investments in mobile edge computing to support lower latency services could also be substantial. (See Don't Count on 5G for a Capex Boost.)
Operators are left scratching their heads partly because -- despite all the excitement about service launches in 2019 or 2020 -- standards still haven't been finalized.
Forget about the roadmap of the 3GPP, the official standards body. Groups such as Facebook's Telecom Infra Project (TIP) and the xRAN initiative that involves Stanford University are now spearheading some of the most critical 5G-related work, as far as mobile operators seem to be concerned. A new TIP working group that features BT Group plc (NYSE: BT; London: BTA), Vodafone Group plc (NYSE: VOD) and Telecom Italia (TIM) , for instance, is trying to come up with a standard that could dramatically reduce the costs of virtualizing the radio access network. xRAN, whose members include AT&T Inc. (NYSE: T) and Deutsche Telekom, is eyeing a similar goal. Yet it remains unclear what impact these projects will have (if any) or how they will coordinate their activities with the 3GPP's. (See Facebook's TIP Seizes vRAN Initiative From 3GPP and DT: 5G Network Slicing Lacks Clear Definition.)
What makes this even more troubling is the prevailing doubt over the 5G business case. As this week's 5G World tradeshow in London has made abundantly clear, operators are not expecting to see much increase in mobile broadband revenues as a result of launching 5G.
"We still see growth [in traffic] without new revenues and would be foolish not to use [more efficient] technology," said Bernard Bureau, the chief technology officer of Canada's Telus Corp. (NYSE: TU; Toronto: T), in explaining the operator's main reason for investing in 5G. "In our country the competition is also really high and competitors are going to be launching 5G networks at the same time -- you cannot fight that even with the best LTE-A network."
Howard Watson, the CIO of BT, has an even more humdrum reason for investing, arguing in straightforward terms that 5G technology is largely about keeping up with traffic demands.
From an investment perspective this is about as exciting as being told that at least the rain won't get any worse. Berit Svendson, the CEO of Telenor Norway, seemed to recognize the dilemma when she said earlier this week that it would be hard to defend 5G investments without the possibility of "monetization."
The telco dream is that "massive IoT" (in plain English, connecting refrigerators and tennis rackets to the Internet) and "ultra-reliable low latency" (connecting self-driving vehicles and robot surgeons to the Internet) will generate new streams of revenue. But they probably won't be very lucrative unless 5G helps operators to become more than just connectivity providers. And there is no obvious reason why it should. (See 5G Guru Predicts Rollout Disparity.)
None of this means that investing in 5G is pointless. When the exhaust pipe falls off your car, you want the latest in exhaust-pipe design to replace it -- not something as old as the vehicle itself. But you also want to know just how much that exhaust pipe will cost.
— Iain Morris, , News Editor, Light Reading