Video consumption, connected homes, health, cities and cars, as well as augmented reality, automated factories and artificial intelligence all feed our insatiable appetite for data. While each of these developments deserve a separate discussion, cumulative global data traffic is expected to grow at a compound annual growth rate of 24% from 2016 to 2021, according to a Cisco estimate.1

This surge of data will need to be collected, connected, processed and analyzed. Which technology will be the winning enabler of this growth in data? In our view, the build-out of the ultra-fast 5G wireless technology will be a critical centerpiece. We believe that certain global telecom companies will be initially disadvantaged by the increasing 5G-related capex requirements, but the value chain of hardware, components, connectivity solutions, software and data centers will stand to benefit over the lifecycle of this transition. Will this surprise anybody? Indeed, we think it will, as very little if any 5G-related revenues are currently priced into the beneficiary companies’ earnings guidance or analyst projections. Below, we explore the technology behind 5G, which helps us understand which companies along the value chain should be the early beneficiaries.

Global data traffic forecast

Source: Cisco VNI, as of June 2017.

What is so special about 5G? It delivers peak speeds 100 times faster than 4G. And latency—the time it takes for data to be stored or retrieved—is expected to be 50 times lower than 4G.2 This technology is key to supporting such applications as autonomous driving, remote surgery, or even augmented reality headsets for firefighters entering a burning building. All of these situations require expediency and precision of data transfer, though not in equal measure. For instance, monitoring 100 security cameras at once requires high data capacity, but not necessarily super low latency. On the other hand, remotely flying a drone depends on a low-latency connection. Under the current 4G standard, telecom companies provide largely the same experience to its users, effectively limiting the capabilities of both the security cameras and drone mentioned above that have needs different from the average consumer. 5G helps solve this problem by introducing the concept of network slicing, an offering made possible by the opening of new spectrum bands for mobile uses. With the new spectrum and other enabling technologies, telecom providers will have the ability to optimize and sell various slices of the network with different characteristics to various users based on the parameters required. This concept will likely become the new business model for telecom providers to, over time, monetize their upfront network investments that are needed today.

Source: Verizon, company presentation, as of November 2017.

The evolution to the ultra-fast 5G is made possible by improvements in spectrum availability and efficiency. This process has been long underway, with the 3.6GHz band the first to be officially reserved for 5G uses at the Mobile World Congress (MWC) in 2015 and more wave frequencies already being considered to be reserved at the MWC in 2019.3 This additional availability of spectrum is important because these new, wider bands of spectrum being considered create more bandwidth, a critical component to achieving faster data speeds. For 4G, most of the currently licensed spectrum around the world is in the sub-3GHz range, and the best an operator could hope for would be to band together 100MHz of otherwise limited spectrum through carrier aggregation. However, the capacity created by the opening of new bands to mobile uses means that operators may be able to use blocks as large as 1,000MHz.To that end, country-specific spectrum auctions are moving forward to license some of the wave frequencies being considered. In the United States, for instance, the Federal Communications Commission (FCC) is expected to begin the auction process for the 28GHz band of spectrum in November 2018, with the auction of the 24GHz band planned to follow directly after.4 These auctions are in turn enabling commercial trials to speed ahead globally, with 111 trials already conducted primarily using the 1-6GHz and 24-29.5GHz bands.5

Spectrum relevant for 5G wireless access

Source: Global Mobile Suppliers Association, as of January 2018.

The one drawback of these higher spectrum frequencies compared to the lower bands used for 4G is the poor quality of propagation, meaning the signal quality erodes more easily with distances and physical obstacles. So while large cell towers do the heavy lifting of current wireless networks, 5G network architecture will increasingly rely on a dense network of small cells, with some as miniaturized as a backpack. They will appear on streetlights, utility poles and other existing infrastructure, adding capacity to dense urban areas, residential neighborhoods, stadiums, universities and other public places. These small cells will utilize two emerging technologies to help enable throughput improvements for 5G—multiple input multiple output (MIMO) and autonomous beam forming. Massive MIMO means that instead of one antenna facilitating a channel to transmit and receive data for a household or device, clusters of antennas will allow up to 100 channels to send bits of data simultaneously, thereby delivering a higher overall network capacity. Meanwhile, autonomous beam forming is the attempt of the network base station, through a broad network of small cells, to more accurately track your phone in order to precisely transmit a signal with the best quality possible. These new technologies and network density will be key to delivering a high-quality 5G offering in a future state, where there will be three times more devices than people by 2021, according to Cisco.6

A closer look at the future wireless world…

Source: Crown Castle International, as of April 2018.

Telecom companies in the United States are hoping that 5G will help them fight back against an onslaught of competition from cable companies, but will it be enough? Cable speeds are rising rapidly as cable companies upgrade old copper cables with high-speed fiber cables.7 With consumers demanding faster data, cable companies have been pushing fiber-based cables deeper into their networks, with the most ambitious laying plans to “future proof” their networks by connecting fiber cables all the way to individual homes, or at least neighborhood nodes. J.P. Morgan expects these speedy fiber cables to expand to reach 33% more U.S. homes over the next three years, bringing the total number of households passed by fiber cables to 40 million. This will likely outpace early versions of 5G technology, which J.P. Morgan anticipates will only cover 20 million homes in the densest parts of the top 25 U.S. markets over the same three-year period.8 With speeds of up to 1 gigabyte per second, fiber cable is a serious rival to 5G for anything that doesn’t need to move. But many things, such as cars and drones, do need to move. That’s why we expect wired and wireless networks to coexist and cater to different end users and end uses. The deployment of satellites will continue to raise speeds in areas not reached by cables or covered by a dense cell network. Cisco forecasts that by 2021 wired devices will account for 37% of global traffic, while Wi-Fi and mobile devices will grow to account for 63% of traffic, up from 49% in 2016.9 Fiber will play a key role in both, as the move to 5G infrastructure should also lead to greater use of fiber to connect the networks of small cells, baseband and radios, and cloud-based data centers.

Average global network speed forecasts

Source: Cisco VNI, as of June, 2017. Mobile speeds exclude 5G.

Upcoming catalysts in the space

Source: FCC, GSMA Intelligence, CAICT, Verizon, as of March 2018.

The first 5G standard was released in December 2017 and 5G is finally starting to make its presence known in 2018, as companies in Asia and the United States roll out early commercial trials of pre-standard 5G technology.10 China Mobile announced in March that it will trial 5G in Guangzhou, Hangzhou, Shanghai, Suzhou and Wuhan this year.11 South Korea will auction additional 5G spectrum in June 2018 and plans a commercial launch in March 2019.12 In the United States, AT&T plans to offer mobile 5G in a dozen cities in 2018,13 and Verizon will offer customer trials of 5G technology in five U.S. cities in Q2 2018.14 While most commercial launches of 5G are likely to come in 2020 and beyond, China has an advantage. Chinese operators have built out 4G sites at an awe-inspiring pace and have some of the lowest numbers of mobile subscribers per 4G site in the world. This gives China an advantage in moving 5G from concept to commercial reality, as China has already invested in a dense national 4G network. Most initial 5G deployments in China will be “non-standalone” (NR) in that they will use 5G spectrum but connect to existing 4G networks. While not a full offering with all the benefits of network slicing, it significantly eases the process of rolling out 5G. The United States will likely follow China’s lead, but additional network densification, and therefore investment, will be needed. The European operators are expected to wait for the standalone 5G specifications to be released in June 2018 and country-specific spectrum assignments to be made before rolling out the technology, though small-scale commercial trials are likely.15

Approval of the first 5G standard was a major step in beginning the build-out of the 5G infrastructure. Global 5G capex is bound to ramp up, albeit slowly initially and at varying speeds and timeframes across regions. We look for an inflection point and a recovery in global telecom capex starting in 2018. A rise in 5G network coverage from 2% in 2019 to 11% in 2020 translates to approximately $160 billion in expected mobile operator capex in 2020, while a larger footprint could require additional incremental spend.16 Telecom providers’ 5G business model is still somewhat uncertain, and the near-term spending expectations can weigh on the telecom stocks, as they did in the case of Asia telecoms already. We do, however, see value in select Asia telecoms, as capex-related underperformance is likely overdone. We are cautious on U.S. and European telecoms that are earlier in the 5G lifecycle. Instead, we prefer to position for the upcoming spending on 4G to 5G upgrades in companies that are the early beneficiaries of telecom capex. These include companies that provide telecommunication technologies, fiber optics, semiconductors, radio frequency filters and network function virtualization software. We also see longer-term value in servers. As rates move higher, we would consider adding tower companies that specialize in small cells, as well as data center REITs. Together, this value chain will have an important role to play in moving to the future state of ubiquitous high-speed connectivity.

Global data traffic forecasts by region

Source: Cisco VNI, June 2017.
1 Cisco, The Zettabyte Era: Trends and Analysis, as of June, 2017.
2 Huawei, 5G Vision, as of April 2018.
3 Global Mobile Suppliers Association, as of November 27, 2015.
4 U.S. Federal Communications Commission, as of March 27, 2018.
5 Global Mobile Suppliers Association, as of January, 2018.
6 Cisco, The Zettabyte Era: Trends and Analysis, as of June, 2017.
7 Cisco VNI, as of June 2017. Forward projections for mobile data speeds are based on third-party forecasts for the relative proportions of 2G, 3G, 3.5G and 4G among mobile connections through 2021.
8 J.P. Morgan, The State of U.S. Broadband, as of January 22, 2018.
9 Cisco VNI, as of June 2017.
10 3GPP, as of December 23, 2017.
11 China Daily, as of February 28, 2018.
12 GSMA Intelligence, CAICT, as of 2017.
13 AT&T, as of February 20, 2018.
14 Verizon, as of November 29, 2017.
15 GSMA Intelligence, CAICT, as of 2017.
16 GSMA The Mobile Economy 2018, as of April 2018.