5G: The Heat is Off
SEE LAST PAGE OF THIS REPORT Paul Sagawa / Tejas Raut Dessai
FOR IMPORTANT DISCLOSURES 917.747.1923 /919.360.6278
psagawa@ / email@example.com
May 5, 2020
5G: The Heat is Off
COVID-19 has eliminated network congestion, the salient motivation for carriers to deploy 5G. While subs shelter at home, using WiFi to link to residential broadband, traffic in network hotspots has plummeted. In response, carriers have scaled back their 5G deployment ambitions, circumstances that will ripple through the wireless ecosystem, hitting demand for network equipment, components, smartphones, and wireless services. (N.B. the US market, where the newly combined TMUS/S is pressuring rivals to spend, and China, where the government is driving deployment are exceptions) We believe consumer mobility will remain depressed, even after official measures to close economies are lifted, with many more months before traffic patterns return to their previous levels. Moreover, delays in the introduction of new flagship smartphones and the likelihood of a lingering global recession suggests that consumer demand for expensive device upgrades may be quite muted. In this context, carriers may cut marketing and capex but will likely see higher churn, fewer subs, declining ARPU, and sharply lower device sales. Disappointing revenues mean weaker margins and cash flows in a high fixed cost business. Network equipment OEMs and component suppliers will disappoint. The smartphone ecosystem will likely see reduced volumes, delayed new product intros, and a mix shift to lower price points lasting well into 2021.
- Initially, 5G is about network congestion. Carriers pitch 5G’s lurid maximum download speeds, but the truth is that ordinary 4G LTE is already as fast as most residential broadband connections – overkill for any mobile app. The real problem is congestion – in busy cells and datacenters – caused by relentless growth in wireless data traffic. 5G, able to pack more data into the same spectrum and to mix in a much broader range of frequency bands, can eliminate network bottlenecks with plenty of room for future capacity needs at a much lower cost than 4G. The true benefits for users, beyond reduced congestion, will come with time as apps start to exploit the reduced latency, edge processing, and low power potential of the standard.
- Social distancing removes congestion. Comcast reported that LTE data usage was down 19% vs. the previous month, while wireless traffic handled by WiFi was up 49%. Network speed test leader Ookla reports LTE download speeds up all over the country, indicative of reduced network congestion. More subscribers at home means fewer visiting the public areas that had previously been the busiest cells in the network. We do not expect mobile traffic patterns to return to their previous plane anytime soon, with social distancing (e.g. work-from-home, reduced travel, reticence to join crowds, etc.) remaining a factor well beyond official policy. In this context, the pressure to invest to clear bottlenecks is gone.
- Carriers pushing out 5G investment. While QCOM expresses confidence in 5G network deployment, upstream suppliers are usually the last to hear bad news – we remember the 2001 crash. ERIC offered more cautious guidance, citing delays in the EU, while poor carrier results portend further delays. Five EU countries have postponed their 5G spectrum auctions and almost everywhere, the process for siting and building new cells has lengthened. The US market is still strong, as the TMUS-S merger holds VZ and T’s feet to the fire, as is China, but even here investment may be slower than it might have been otherwise. We believe global 5G capex is unlikely to reaccelerate and carriers will not aggressively market the service until consumers are once again willing to travel and enjoy crowded venues.
- Carriers may not be a safe haven. Historically, telecom has been a favored defensive sector during recessions, but this time, circumstances warrant considerable caution. A considerably less mobile population hits data overage charges, roaming fees, and other highly profitable billings. Homebound subscribers may be more prone to cancel services and shift to lower priced plans. For high fixed cost network businesses, lost subs, lower ARPUs, and sharply decreased new phone sales are real problems that will threaten cash flows, and in worst case scenarios, dividend payments.
- The wireless equipment value chain will be weak. Telecom equipment suppliers (e.g. ERIC and NOK) will see orders slip to future periods. Upstream component suppliers (e.g. MRVL, XLNX, QCOM, QRVO, etc.) will see their orders fall. Test equipment sales (e.g. KEYS, NATI, etc.) will hit a lull. US tower companies will be spared the most of it, as the newly combined TMUS/S is likely to pressure its rivals to maintain spending at a better clip than international peers, but there is more downside risk than upside potential.
- The smartphone value chain will also be weak. QCOM reports that 1Q20 smartphone volumes were 20% lower than it had expected and projects a 30% deficit for 2Q. AAPL has reputedly cut 2H20 manufacturing orders by 20%, and while it is still expected to announce its 5G iPhone models at the end of September, availability will be limited for weeks. Extremely thin 5G coverage, voluntary isolation (staying and working at home), and general economic hardship will limit the consumer appeal of expensive new devices and cash strapped carriers are unlikely to step up to expensive marketing campaigns to push user upgrades. (N.B., we caution against using China, where the state can shape consumer behavior, as the model for recovery elsewhere) All of these things will weigh on consumer buying decisions – the smartphone market may not return to 2019 sales levels until 2022. This means considerable pain for smartphone makers, their manufacturing partners, and their component suppliers.
- Trading does not reflect these risks. Many stocks with significant 5G exposure have outperformed the S&P500 YTD (-12.5%), including VZ (-8.4%), ERIC (-5.8%) NOK (-4.3%), and AAPL (-0.1%). Chipmakers have been hit a bit harder, but QCOM (-14.2%), SWKS (-17.6%), AVGO (-17.0%) and QRVO (-19.1%) are within shouting distance of market performance despite their outsized negative exposure to the pandemic. XLNX (-15%) has been the hardest hit, despite a partial offset from its strong datacenter sales. TMUS is an anomaly, up 13.4% YTD, but this is entirely due to the approval and closing of its S acquisition, which was unexpected by many investors and which offers substantial upside synergy.
Tomorrow Has Been Delayed Indefinitely
2020 was supposed to be the year of 5G. The leading-edge markets – China and South Korea – were already well along in network deployment, with the US, Western Europe, Japan, and other geographies beginning to roll out. Global telecom capex estimates were revised upward. The first 5G phones – from Samsung, OnePlus, LG, Huawei, and others – were hitting the market. Mighty Apple was known to be on track to deliver its first 5G iPhones for its normal annual introduction at the end of September. Analysts wrote of a phone replacement “super-cycle” that would claw back the declining global smartphone unit volumes and spur new services and peripherals, all driven by the excitement of 5G. Then the virus came.
The early signs suggest a significant slowdown in 5G network deployment, particularly in Europe. This makes sense, as the real impetus for carrier investment in the new standard is relief from congestion in network hotspots. Mobile data traffic had nearly doubled every year over the past several, demanding nearly constant investment to clear bottlenecks in the areas where people gathered in dense crowds – urban centers, shopping malls, airports, stadiums, etc. Because 5G can make use of a much wider range of frequency bands, because it can carry much more data per radio wave form, and because its antennae can focus signals much more precisely, it is much more effective and much less costly than 4G LTE in resolving congestion. Now, with subscribers sheltering at home, connecting via WiFi, and NOT gathering in dense crowds, congestion is no longer a pressing issue. Meanwhile, the process of deployment (e.g. site selection, permitting, labor, etc.) is far more burdensome during lockdown. Thus, the drop off in capex.
These circumstances have ramifications. For carriers, a static population takes the pressure off of capex but hits revenues – subscribers on Wi-Fi at home do not have data overage or roaming fees, do not go to carrier operated stores to buy new phones, and they may be more likely to churn or drop to a cheaper plan. The impact on equipment vendors is obvious – delayed deployment means sales are deferred. In a recession, consumers will be less likely to upgrade their phones anyway and spotty 5G coverage and delays for new flagship devices are more reasons to hold on another year. Delayed network deployments and new device introductions also hits demand for test equipment, back haul, tower space, and other products and services tied to network deployment. Finally, the demand weakness will hit upstream as well. Expect the market for components to be badly hit – smartphones parts were 25% of all semiconductor shipments by value in 2019.
Meanwhile, stocks in the group have not suffered during the crisis. Most wireless carriers, equipment makers, smartphone brands, and component suppliers have outperformed vs. the S&P500 (-12.5% YTD). Expectations are relatively rosy – VZ revenues are projected to decline just 3.9% in 2020, AAPL’s FY20 (ending in September) sales consensus is flat, and Qualcomm (also FY ending in September) is estimated to grow nearly 10% (in part due to parts destined for the iPhone 12, now likely to sell in minimum volumes in 4Q). We think these numbers have to come down significantly and would not want to own the stocks until they do.
5G is All About the Carriers
Last year, B.C (Before Coronavirus), we wrote about the rise of 5G. (The 3 Phases of 5G: Coverage, Density, and Applications) In that piece, we detailed our belief that the pressure to expand network capacity to serve surging mobile data demand in competitive markets would force carriers to step up CAPEX well ahead of the then stated plans (Exhibit 1). Essentially, while deploying 5G involves significant upfront capital spending, incremental capacity expansion would be dramatically cheaper with the new standard than with 4G LTE. Thus, the carriers that bit the bullet first would have significant performance and marginal cost advantages until their rivals followed suit. This would be a classic game theory “prisoner’s dilemma” that would assure that most of the world’s carriers stepped up their 5G investment whether they initially wanted to or not.
The pandemic changes the calculus for several important reasons. First, perceived network performance is a function of the capacity of the busiest cells in the network to handle traffic demands at the busiest times of the day (Exhibit 2, 3). As all of the users in a cell share the aggregated capacity of that cell, crowding equals congestion equals poor performance. Thus, gathering areas – busy urban centers, parks, airports, stadia, etc. – are the gating factor and the big reason that carriers wanted to offload traffic to 5G (Exhibit 4, 5). Obviously, gathering areas are not so crowded these days and may not be for quite a long time. Investing to resolve congestion can be put off for another day.
Second, the process of deploying a new network becomes more difficult and more expensive under lockdown conditions. Finding sites for new base stations, particularly where existing towers are inadequate,
Ex 1: Snapshot of Options and Outcomes of Wireless Prisoners’ Dilemma for 5G
Exh 2: Pre-Pandemic Mobile Traffic Density Forecasts in Major Global Urban Centers Predicted Rapid Growths in Traffic
Exh 3: Timelines of when operators would begin to run out of capacity in at least 50% of sites – before the pandemic
Exh 4: US Internet Performance During COVID-19: Avg. Mean Download Speeds
Exh 5: Growth Rates in US Internet Traffic Volumes During COVID-19 Pandemic
is difficult. Gaining necessary approvals and permits is difficult. Engaging workers to implement the site is difficult. Testing the installation is difficult. Significant delays are hardly surprising (Exhibit 6).
Third, carriers may need to conserve cash. Social isolation means much less travelling, and much less traveling means far less roaming traffic. For many carriers, particularly in Europe, roaming is one of the
Exh 6: Summary of Post-Pandemic Factors that alter 5G Deployment Pace
Exh 7: Cash on Hand and ST Liquidity Metrics for US Carriers, April 2020
most lucrative services they provide. Staying at home also syphons off total traffic as users connect at home via Wi-Fi rather than the carrier network, meaning much less revenue from overage fees. Subscribers staying well below their usage caps may also decide to drop down to a lower priced service tier, or even drop service entirely. Moreover, customers staying at home do not visit carrier shops and don’t buy new devices or accessories, while the carriers still pay rent and employee salaries (Exhibit 7).
Yes, but 5G is Faster than 4G…
The marketing pitch for 5G is all about speed. In practice, the average US download speed for 4G LTE is a bit over 35Mbps, short of a theoretical maximum that may be 10 times that speed (Exhibit 8). In contrast, 5G supports downloads at faster than 1Gbps in many cases. Who wouldn’t want that?
The real question may be WHY would anyone want that? First, pandemic conditions are seeing 4G speeds rising on relatively uncongested networks. Second, poor download performance is rarely caused by the capability of the network. Server congestion – too many users accessing the same application from the same host – is generally the culprit and would not be helped a lick by a faster interface. If not server congestion, then network congestion, which has improved dramatically during the crisis.
Third, there are no major applications that require speeds in excess of 100Mbps and most are fine with less than 10Mbps. For example, Netflix recommends 25Mbps for 4K video streaming and 5Mbps for HD. On a mobile device, it is difficult to notice any difference between these two standards, much less anything at higher resolutions. 100Mbps buys you 8K video streaming and is easily supported on relatively uncongested 4G LTE networks. While carriers have tried to sell the sizzle of these lightning fast speeds, they had been expecting to eat the steak themselves in the form of much higher aggregate capacity with which to serve that surging overall data demand. Except it is no longer surging.
Exh 8: Avg. US Download Speeds Ranked by Carriers Across the US for 4G LTE
Exh 9: SSR Criteria for 5G improvements in critical functions over 4G
Of course, 5G has other benefits for subscribers – greatly reduced latency, better support for edge servers to improve performance, flexible bandwidth provisioning to support low speed, low power internet of things (IoT) connections, new antenna designs that can help fill in poor signal shadows, etc. (Exhibit 9) Unfortunately, none of these things will be supported in the initial network offerings.
South Korea and China Went First
South Korea, the home of Samsung, and China, the home of an autocratic government hellbent on propelling its tech industry to 5G leadership, have largely built out 5G networks. In Korea, 10% of subscribers now use 5G capable devices and the reception has been mixed at best (Exhibit 10). To expedite deployment, Korean carriers built 5G atop their 4G infrastructure, using a technology called Dynamic Spectrum Sharing (DSS) to parse traffic between the two standards. This has resulted in significant latency problems with 5G connections often delivering far WORSE lag than 4G connections. Furthermore, 5G connections have proven far more battery hungry than 4G LTE, leading to further subscriber
Exh 10: 5G Subscriber Base in South Korea – growth is flatlining shy of 5M subs
Exh 11: Early 5G Deployments in South Korea Yielded Tepid Improvements
dissatisfaction. While connections speeds are clearly higher, many Koreans have chosen to disable their 5G capability, hardly a ringing endorsement of the standard (Exhibit 11).
Details in China are much less reliable. Carriers began offering 5G service last November, just ahead of the crisis, and the government offered projections that 110M Chinese (about 7% of wireless subscribers) would be on the standard by the end of 2020. This looks ambitious, given the impact of the of the pandemic on the economy and the possibility of secondary waves of infection, but the state’s ability to manipulate news flow and direct consumer behavior should not be underestimated. Nonetheless, even if the Chinese forecast proves accurate it seems unlikely that consumer experiences in that country are likely to have any bearing on behavior anywhere else.
In March, Huawei executives averred that European 5G rollouts would “certainly be delayed”. In its April 28 call with investors, Samsung management said, “investments in 5G networks will be reduced or delayed domestically and internationally as more effects of COVID-19 unfold”. Ericsson’s CEO Borje Ekholm said “COVID-19 and actions taken by governments to slow down the spread are making our service delivery and supply harder due to lockdowns and travel restrictions in many countries,”. Spectrum auctions in France, Spain, Portugal, Austria, and the Czech Republic have been indefinitely deferred.
While European 5G network buildouts are inevitable, the timing is malleable. For carriers that offer both wireless and wireline services, residential broadband, straining under heavy usage, is taking investment
Exh 12: America Movil Reduced Capex for FQ with plans to cut further in FY20
priority. Wireless businesses, suffering with scant roaming fees, new phone sales or overage fees, must conserve cash. 5G, a luxury when hotspot congestion has eased, can wait. Even in countries where carriers are eager to continue deploying the new standard, such as Germany, pandemic related work stoppages have put deployments months behind schedule. The same logic applies to much of the not China, South Korea, or Taiwan world. Brazil and India are delaying 5G spectrum auctions. America Movil, Mexico’s largest carrier and a major player in the rest of Latin America, announced a cut to capex after delivering poor 1Q results announcing that it would focus its spending on maintaining the quality of its existing networks (Exhibit 12). All three incumbent Japanese carriers have launched 5G service, but the coverage areas are extremely limited with the carriers actually posting specific location addresses where subscribers will be able to get signal. Broad coverage isn’t anticipated until 2022.
If You Don’t Build It, They Won’t Come
Beyond South Korea, China (in certain cities) and the US (again, in certain cities) it is not clear that any country will have broad coverage 5G networks in place before year end. That context alone is sufficient to raise doubts for 2020 5G subscriptions and device sales. Even in the US, the top two carriers are relying on very high frequency “millimeter wave” spectrum for most of their 5G buildouts, meaning blisteringly fast connection speeds from extremely small radius cells subject to significant interference from solid objects juxtaposed between device and cell antennas. Mobile users will be constantly moving in and out of 5G coverage. Given reports of disruptive 4G/5G handoffs, network performance is likely to be a considerable
Exh 13: US Carrier Capex Estimates could be Overly Aggressive for FY+1 and +2
Exh 14: US Carrier Capex consensus estimates have been aggressively revised to the upside YTD – post COVID reality may be different
issue that could poison word-of-mouth around the new standard. Couple that with the poorly understood facts that a jump from 100Mbps to 1Gbps will have absolutely no bearing on the quality of video streams and cannot resolve the congestion issues at the server running the applications, and a 5G backlash is a distinct possibility.
Now, add in the financial circumstances of the carriers. Revenues hit by sharply lower roaming fees, fewer data overage charges, higher churn, and consumer switching to lower priced plans. Costs burdened by high network fixed costs and retail shops that remain closed. We’ve already noted that many have pulled back on capital spending. How many will step up to co-promote new 5G smartphones in the fall? We think it is reasonable to expect less than vigorous promotion (Exhibit 13, 14). This will affect adoption.
Then consider that many anticipated flagship 5G devices will be either delayed or made available in limited quantities. With Chinese factories shuttered for 6 weeks, with other markets involved in smartphone supply chains still earlier in the COVID-19 curve, and with international travel severely curtailed, it will be hard to manage the logistics of launching new models and ramping to scale. The WSJ reported that Apple still intends to launch its 5G iPhone 12 on its typical end of September schedule, but that the product will not be widely available for weeks thereafter. Sounds about right.
Let’s layer on the effect of economic recession. The autocratic government of China can manipulate the news flow to induce greater confidence in the economy and promote the industrial policy goal of 5G leadership, but consumers in most countries are undoubtedly skittish about the economic future. Rampant unemployment, reduced wages, business failures and political fearmongering are hardly the ideal backdrop to stoke consumer demand for discretionary $1,000 smartphone upgrades (Exhibit 15). During the worst of the 2008-2009 financial crisis, the iPhone was a fledgling product mostly in the hands of economically comfortable fanboys trying to live life on the edge. Today, nearly one in 7 humans on earth has one. The propensity to upgrade has been falling for 5 years. It is not clear why we should expect an inflection point in the current environment.
Exh 15: Snapshot of Flagship 5G Phones Announced To-Date with Price Points
The great 2020 5G delay will wreak havoc across many sectors of the TMT landscape.
Carriers – Wireless operators were counting on 5G as the way out of the plague of congestion created by the extraordinary growth rate of mobile data usage. Resolving network bottlenecks was like a game of Whack-a-Mole, fix one overwhelmed site and another pops up somewhere else. The pandemic ended the game in a different way with substantial unanticipated burdens. Thus, carriers may be able to save on capex, but will sufferer from sharply lower roaming charges (roughly 1-4% of carrier revenues), reduced overage fees for data use beyond a subscriber’s plan (3-5% of revenues), higher churn, and mix shifts to lower priced plans. These impacts will vary from market to market and carrier to carrier and we advise caution.
Carrier Suppliers – Obviously, if capital spending on 5G is pushed out, so will be the sales of 5G network equipment. Huawei, Ericsson, and Nokia have already called out pandemic-related delays to explain their slowing sales and cautious forward guidance. Test equipment, an area where we had been quite bullish, will likely get hit from carriers with less urgency AND from device makers facing delays of their own. The tower
Exh 16: Last 3 Months Cumulative Returns Relative to S&P 500 – US Carriers
Exh 17: Last 3 Months Cumulative Returns Relative to S&P 500 – Global Carriers
Exh 18: Last 3 Months Cumulative Returns Relative to S&P 500 – Carrier Suppliers
Exh 19: Last 3 Months Cumulative Returns Relative to S&P 500 – Smartphones
Exh 20: Last 3 Months Cumulative Returns Relative to S&P 500 – Mobile Suppliers
companies (American Tower, Crown Castle and SBA) primarily serve the US market, where the T-Mobile Sprint buildout appears to be full steam ahead, so those stocks may see less impact.
Smartphone Makers – The Chinese market might prove to be okay, but we are skeptical of government-influenced reports of a society back to normal in the wake of massive forced quarantines affecting hundreds of millions of people. We will see if the factories remain open and operating full tilt for the rest of the year and if Chinese consumers are willing to wait in line for an iPhone 12 come October. We have a bit more confidence in South Korea, home of Samsung. The rest of the world? No confidence at all. Combine poor 5G network coverage, delays and tight availability for flagship devices, a semi-permanent shift to a less mobile way of life, scant marketing support from the carriers, and economic hardship and uncertainty. What you will most likely get is a very weak market for premium 5G smartphones.
Components – If equipment orders are delayed and if device upgrade decisions slide, the makers of chips and other components will feel the pain as well. Square in the crosshairs is Qualcomm, which delivered upside in its 2QFY20 numbers but called out sharply lower device sales for the June quarter as well. We think the company’s projections for 175-225M 5G devices sold in calendar 2020 are wildly optimistic. Qorvo, Broadcom, Skyworks, Xilinx, Corning, MediaTek, and others with disproportionate exposure to either base stations or smartphones carry significant risk of disappointment.
Stocks in these four categories have hardly been brought to the woodshed (Exhibit 16 – 20). International carriers are mostly down relative to the broader market, but US carriers have held on despite the further risk of a buffed-up T-Mobile. Ericsson and Nokia have recovered back to their January 31st valuations despite shaking earnings reports, significantly outperforming the S&P500. Test equipment leader Keysight is also flat to its pre-pandemic levels and ahead of the broader market. American Tower, Crown Castle and SBA have outpaced the market as well, although with more justification (Exhibit 21).
Apple has been a modest outperformer, a bit of a headscratcher given the significant risks to its robust consensus expectations. Qualcomm has underperformed, but still carries aggressive estimates going forward. Other chipmakers have similar profiles, valuations have pulled back a bit but expectations seem too rich.
Exh 21: Key Financial and Valuation Metrics for 5G Companies by Category
Exh 22: Summary of SSR Thesis Post-COVID for 5G