The tricky—and really expensive—business of building 5G

If you build it, will they come?
If you build it, will they come?
Image: Reuters/Arnd Wiegmann
We may earn a commission from links on this page.

Simply put, 5G will make the ability to analyze colossal volumes of data cheaper and more accessible—at a scale of magnitude that turns what now sounds like a futuristic fantasy into a Main Street reality. It will also cost up to $1 trillion to build.

What will mobile wireless operators get in return for giving us this magical new reality of near-instant and ubiquitous information?

That, as it happens, is the biggest question looming over our 5G-ified future. Sure, the long-term chance to cash in is fairly clear. In addition to creating whole new markets for data services, 5G will also make it much cheaper for operators to provide those services.

But while 5G is still being figured out, operators could find themselves stuck in a kind of financial purgatory: having to invest a lot in building 5G without much in the way of return.

The 5G revolution

Based on the rollout so far, it would be easy to mistake 5G for yet another phase in the 30-year evolutionary march toward ever-cooler phones. Not so. Okay, yes—our phones will get way faster and more powerful. 5G will bring consumers other new things too—highly amazing virtual reality experiences for gamers-on-the-go, for example. And, slightly less dazzlingly, connections fast enough to supplant home broadband connections (and bring internet to places that fiber can’t reach).

Crucially, though, unlike mobile “generations” past, the real goldmine of 5G lies not in bringing the internet to people, but, rather to objects. The technology is specifically designed to work with devices beyond simply smartphones, says James Thompson, CTO of Qualcomm, one of the companies that pioneered the 5G standard—and “ultimately support new features we do not even envision today.” Not knowing what, exactly, those devices are, however, makes it hard for operators to plan their buildouts, let alone recoup their investments.

In some countries, notably China, governments are taking some of the strain off operators, subsidizing the enormous upfront costs. But in the US and most other nations, carriers face greater pressure to sign up customers to defray the cost of buildout. And with the majority of customers satisfied with their 4G connections, selling services to businesses—private corporate networks, for instance, or network security—will be key.

Operators have already begun exploiting 4G’s virtualized architecture to sell cloud-based private networks to enterprises, letting firms connect their offices and data centers over vast distances via their own private internet. 5G will enable carriers to build these proprietary networks vastly more cheaply and flexibly—but first they invest in the capabilities to provide those services.

Orange, for instance, is now France’s top supplier of cloud services and earns about 30% of its revenue selling its services directly to businesses, says Arnaud Vamparys, an executive at the Paris-based operator that provides mobile services in 27 countries worldwide. Even so, in the last year it has hired thousands of employees to staff its new cyberdefense division, he says, and made a slew of recent acquisitions to beef up cloud and network security offerings.

Virtualization is critical to cutting costs, particularly those that come from spending on hardware—and this new generation super-charges that digital shift. “5G changes the ability to virtualize our technology,” says Heidi Hemmer, VP of Technology at Verizon. “Putting mobile architecture in the cloud allows a seismic change in the services we provide.” The “pivotal point,” she notes, is the switch to a fully 5G virtualized core network (known in the industry as a “standalone” network).

But that’s not here yet—and probably won’t be a reality until at least 2025 (except in China, where one operator plans to launch the world’s first standalone 5G network next year).

Until then, operators will be doing the first phase of 5G: upgrading existing 4G core infrastructure with 5G technology so that it supports new 5G frequency. And that means before they can start to truly reap the benefits of 5G technology, operators will first still have to build a heck of a lot more hardware.

The vast increases in computing power and AI capabilities that underpin 5G architecture mean network hardware differs structurally from past generations. Instead of erecting hulking cell towers, operators will be installing many more small antennas to carry signals, including mounting these on “city furniture” (industry lingo for urban objects like utility poles, street lamps, manhole covers, bus stop vestibules). That will require spending billions on leasing the real estate for new cell sites (and, later on in the buildout process, mini-data centers).

And 5G network equipment itself costs a lot, in no small part because there are only three equipment providers: Nokia, Ericsson, and Huawei (though increasingly, Samsung is emerging as a competitor). The US and several other countries have banned Huawei equipment—upping the prices operators must pay even more. And since 5G will change the topography of network infrastructure, building it will also require spending on building new fiber-optic connections linking antenna stations to the core network.

The burdens of the past

What’s more, carriers can’t spend all of their infrastructure funds on 5G, since they have to keep building and maintaining networks used by the billions of customers for their existing 2G, 3G, and 4G networks. A McKinsey study estimated that operators in one unnamed European country would have to up network-related capital spending 60% from 2020 to 2025. Continuing to boost the capacity of 4G networks, in particular, is critical, since rising data use is already straining capacity. They’re also the biggest subscriber base.

“We have over 100 million subscribers on 4G and no intention of abandoning them. That’s still our bread and butter, and we’re still adding capacity and functionality to our 4G network,” says Hemmer, noting that with past mobile generations, it’s tended to take a few years before customers to start upgrading en masse. (For example, Hemmer says she only just convinced her parents to move off 3G. “They were very happy with their iPhone 4s,” she adds.)

But all those spiffy new 5G antennas can’t send out signals without waves frequencies to travel over. That brings us to another huge upfront cost to building 5G networks: spectrum. In a single auction for US spectrum in June, AT&T and T-Mobile spent just shy of $1.8 billion, combined. To understand why unoccupied frequencies are so valuable—and how 5G changes that calculation—it helps to understand a bit about wave physics.

Wave chasers

Lower frequencies, which is roughly similar to those used for 4G networks today, tend to travel far and cover a much broader geographic range. But their capacity is limited. Midband falls somewhere in between those tradeoffs—it’s often called the “Goldilocks” range of the spectrum—which makes it good for suburban areas.

Finally, there’s the feat of physics known as millimeter wave (mmWave). These ultra-high frequencies support a far greater volume of traffic at fast speeds, making it ideal for urban hubs like train stations or airports, or dense clusters of connections that you might find at wireless factories. The big challenge—and the reason this spectrum hasn’t been used in the past—is that these waves can’t travel more than a couple hundred feet and are easily blocked by building walls and other objects. Sending signals worked only if the antennas kept very still (which, obviously, isn’t feasible with mobile technology).

Thanks to advances in antenna technology and signal processing led by Qualcomm and other companies, the 5G standard opens up that spectrum to mobile networks. But the real-world application of this new technology has yet to be totally mastered.

Some countries, notably China, are avoiding that problem by focusing on the midband for their 5G networks, eschewing mmWave altogether.

The big upstart in the US market—or, as it calls itself, the “un-carrier”—is hedging its approach, exploiting another tech advance in the 5G standard: allowing operators to combine the capacity of a range of spectrum. T-Mobile boasts what it calls a spectrum “layer cake”—providing 5G coverage via a combination of low-, mid-, and high-band chunks of spectrum alongside each other. “This is a game changer because when you look at what’s unique [among US carriers], it’s the different frequency bands,” says Karri Kuoppamaki of T-Mobile USA. “It’s important to understand why it matters: frequency has a direct relationship with coverage.”

This is distinctive because viable lowband spectrum is already mostly used up by 4G networks, while a good deal of the midband is already claimed by satellite communications operators and the military. T-Mobile’s coveted slice of midband real estate comes by way of Sprint (pending a decision on a lawsuit concerning the proposed merger of T-Mobile and Sprint). It also has an unused chunk of lowband spectrum that it acquired a while back.

The twin titans of American mobile networks don’t have that luxury, which in part explains why their 5G launches have used mmWave. Relying on mmWave could prove risky: though successful rollout will provide customers the ultrafast speeds and connectivity promised in 5G marketing, the limited range makes it tough to market to consumers. And the technology itself is still a bit wonky (which has prompted some analysts to wonder if that’s why US operators haven’t bought even more mmWave spectrum).

Indeed, the tricky nature of mmWave is why AT&T chose to focus its 5G launch on business clients. It now offers 5G to enterprise customers in 20 cities and recently unveiled a wireless semiconductor plant in Austin, TX, in partnership with Samsung, via highband spectrum. The company is waiting to expand coverage nationwide when it deploys midband spectrum in the middle of next year.

Still, the end of all this profitless spending is in sight. 5G will also mean software does a lot more of the heavy lifting, creating opportunities for operators to find a much more diverse base of network architecture suppliers, and make them less dependent on their expensive relationships with the vendor troika. Things will get even cheaper when operators achieve that “pivotal point” noted by Verizon’s Hemmer: when 5G networks lift off the current 4G hardware-dominated architecture to head for the cloud.

The scramble for market share

With past generations, “you had to build more of the same hardware if you needed more capacity,” says T-Mobile’s Kuoppamaki. “But down the road you’re stuck with that hardware.” 5G’s software-based network architecture lifts this yoke, allowing carriers to tailor their services to a huge variety of data needs, making it possible to wring the most out of their network capacity. And, unlike existing networks, having to waste way less investment maintaining old infrastructure. “In a virtualized world,” says Kuoppamaki, “you can repurpose that [capacity] for other functions as needed, which helps with scaling and providing efficiencies.”

And, therefore, with earning a profit. Or that’s the hope, anyway. The transformational nature of 5G means that’s still very much up in the air.

Just as it creates new opportunities for operators to launch new business lines, it also creates new opportunities for competition.

For example, the scalability and mobility of 5G networks could make it possible—and profitable—for foreign operators to compete with each other in providing private network services on each others’ home turf.

An even bigger competitive threat will come from companies like Google, Facebook, Amazon, and Alibaba. The Big Tech titans are cash-rich, boast phalanxes of computer whizzes, and already have decade-long head starts in selling data analytics and cloud computing services to businesses.

Of course, as with everything 5G-related, there’s a lot that needs to be figured out before we get a sense how any of that will shake out. Unfortunately for operators, the resulting trials and errors alike will be done on their dime. But the longer that process takes, the more time they give their tech competitors to stake claims on that 5G goldmine.