Free to View / Equities Diversified Telecoms
October 2020 By: Global Telecoms team www.research.hsbc.com SPOTLIGHT
SPOTLIGHT 5G 5G Paths diverge
5G so far has underwhelmed, but the latest version brings new capabilities – and real opportunities to differentiate
Network ‘slicing’ is a potential revenue generator – but this will take time to develop
Asset // Subcategory // Asset Operators and suppliers are taking very different paths with 5G: this divergence in strategy will create Equities execution risk and reward | Diversified Telecoms
This is a redacted version of the report published on 07-Oct-20. Please contact your HSBC representative or email [email protected] for information. October 2020 October
Disclosures & Disclaimer: This report must be read with the disclosures and the analyst certifications in the Disclosure appendix, and with the Disclaimer, which forms part of it. Free to View ● Equities - Diversified Telecoms October 2020
Why read this report?
5G so far has underwhelmed, but the latest version brings real opportunities for operators – as well as execution risk
Network ‘slicing’ is a potential source of new revenue – but this will take time to develop
Operators and suppliers are taking very different paths with 5G: this divergence in strategy will create execution risk and reward
5G so far has been underwhelming – what’s coming is much more interesting The initial version of 5G has underwhelmed, with only minimal differences in services and performance for customers, and price and demand for investors. The latest version, Release 16, is being deployed now, and is materially different. Operators will be able to offer ‘slices’ of the 5G network to different groups of customers. 5G slices can be tailored to specific bandwidth, latency and Quality of Service parameters. This creates the potential to bill based on the quality of service provided, and may help grow revenue. Yet the upgrade decision isn’t simple. The migration to a ‘cloud-native’ StandAlone core is complex, with dispute in the industry as to the best architecture, and considerable execution risk. As a result, most operators (particularly in Europe) plan to deploy 5G radio alongside their 4G core − helpfully termed ‘Non StandAlone’. This will enable operators to get more return from their 4G investment in the absence of any clear 5G business case. After a period of relative homogeneity with 4G, we expect the 5G experience to differ between operators, creating winners and losers for investors.
Network ‘slicing’ is the service to watch Network slicing creates the ability to customise the network for corporate and consumer needs. This could help reclaim some of the value in the provision of connectivity – a service that is increasingly commoditised. The improved capabilities of 5G will make it relevant to new groups of corporate users. Needless to say, slicing brings its own complexities. Billing, provisioning, and after-sales support are challenges additional to making sure the technology works. Operators are likely to struggle with this – offering opportunity to those who execute well.
Disruption ahead Services have been relatively homogeneous in the 4G era. Equipment from the same ‘big four’ vendors is deployed to the same timelines. Operators offer the same services and handsets, and the same billing models. Connectivity has become commoditised, and the telecom industry investment case has suffered as a result. 5G promises dislocation, and with it opportunity and risk for investors. On the supply side, accelerated swap-outs of equipment offer opportunities for smaller vendors and the Open RAN movement. Operators are taking different approaches to the core upgrade, as well as to spectrum and Mobile Edge Compute. We explain why billing based on quality of service is a key indicator, why 5G mobile gaming is a test case for the enterprise, and why the Internet of Things and the connected car are (still) the services of the future.
This is a redacted version of the report published on 07-Oct-20. Please contact your HSBC representative or email [email protected] for information.
1 Free to View ● Equities - Diversified Telecoms October 2020
Contents
Why read this report? 1
Facts and figures 3
Executive summary 4
Release 16: the real deal 7
5G – new services 17
On the spectrum 25
Supply chain disruption 31
Disclosure appendix 35
Disclaimer 37
2 Free to View ● Equities - Diversified Telecoms October 2020
Facts and figures Release 16 3x
Finalised in July 2020, this is the Lower costs per unit of data on 5G next iteration of 5G – and very compared to 4G different from the current version
Network ‘slicing’ is one of the key features of Release 16: it enables precise Hundreds configuration of speed, latency, and Quality Of separate network ‘slices’ are of Service for a ‘slice’ of the network, which possible – although telcos are debating can be reserved for a certain group of users how many are practically feasible
5-10% 800MHz 200MHz
Increase in Average Quantity of mmWave 3.5GHz spectrum available Revenue Per User as spectrum held by each to each operator in China customers migrate to 5G operator in South Korea
Mobile radio technology typically uses 10-20+ vertically integrated kit from a single The number of vendors involved in vendor. Open RAN uses open interfaces Rakuten Mobile’s ‘Open RAN’ network, and advanced software to change this, compared to 1-2 for most operators bringing the flexibility and scalability of the cloud to mobile.
5-20ms 4Q20
Latency on a 5G network, compared to 5G launch using Open RAN technology by c50ms or more on 4G Rakuten in Japan, and Dish in the US
3 Free to View ● Equities - Diversified Telecoms October 2020
Executive summary
5G StandAlone network launch is starting now: this is the real deal, with potential for new services, and new billing models
The key in 5G is the potential for bespoke services – after a period of relative homogeneity in 4G, 5G offers the chance to differentiate
Differences in strategy are starting to emerge – network slicing could separate the 5G leaders from the others
Standalone deployments mark the beginning of the ‘real’ 5G
5G to date has been underwhelming. Users haven’t seen much performance improvement over 4G, and investors have seen little differentiation in tariffs or services. Yet in most markets capital intensity looks likely to rise due to 5G (and has done dramatically in Korea and China in particular). To restate the title of our 2018 report, What’s the use?
We think things will get (much) more interesting from here, as StandAlone 5G networks are deployed, alongside services and devices aligned to the latest 5G standard, Release 16. Key improvements are as follows:
A 5G connection from end-to-end, not just in the radio. Nearly all the 5G focus so far has been in the radio access network (RAN). Yet this is just one component of 5G. The true value of 5G will become apparent when operators and developers can work with a set of 5G processes and functions from deep in the core network to the device. The 5G core network is ‘cloud native’, enabling much improved flexibility and scalability. Processing can be pushed to the network edge, enabling lower latency.
Network ‘slicing’. This end-to-end control will allow ‘slicing’ of the network for different use cases and scenarios. These will be defined and set up according to quality of service
Exhibit 1: China − mobile revenue vs Exhibit 2: Next steps in 5G services mobile data
1,000 140,000 Enhanced mobile broadband (eMBB) 900 120,000 800 GB/s Communication 3D，UHD Video 700 100,000 600 Cloud Office/Gaming 80,000 Augment Reality 500 Smart Home 400 60,000 Industrial Automation 300 40,000 Voice 200 Highly reliable 20,000 Application, such as 100 Smart City Mobile Healthcare Autonomous - - Vehicles 2010 2012 2014 2016 2018 Massive machine type Ultra-reliable and low latency Revenue, RMB bn Traffic, GB m communications (mMTC) communications (uRLLC)
Source: Company data, HSBC estimates Source: ITU
4 Free to View ● Equities - Diversified Telecoms October 2020
parameters – an emergency service or smart robot facility would have a very different down-time and latency sensitivity compared to a logistics service tracking shipping containers. Operators could deploy tens and eventually hundreds of different slices.
Why care? Slicing will enable operators to create very different services, and potentially reclaim ‘Slicing’ is a 5G capability that could help reclaim some some of the value in the provision of connectivity. Currently, of course, Over The Top platform of the value in the network companies capture much of the value in the digital economy. We consume vastly more data than we did, and would be lost without our smartphones. Yet our mobile bills are – in many markets – lower than they were five years ago.
The shift in focus from connectivity to service could enable new billing structures. Currently, 5G tariffs have the same structure as 4G: while some operators have reported a 5-10% increase in spending as customers migrate, this is mostly a function of early adopters being willing to spend more – we expect this gap to narrow. The initial focus has been on enhanced Mobile BroadBand, a similar service to 4G. Release 16 and Standalone add the two remaining corners of the 5G triangle: uRLLC (Ultra Reliable Low Latency Communications) and enhanced Internet of Things capabilities, termed Massive Machine Type Communication (mMTC). Of course, this will take some time to develop – the near-term impact of 5G on stock prices will – mostly – be limited.
Supplier market is opening up, creating opportunity for operators
A key development in the past two years has been the opening up of the telecom equipment market:
Vendor lock-in has been the default mode. The complexity of mobile telecoms – particularly in the radio network – has meant that so far only vertically integrated hardware / software can handle the processing loads. And as competition has intensified in the suppliers, the supplier landscape for 4G was basically the ‘big four’ equipment vendors. Extensive use of proprietary equipment has created problems as governments start to limit the extent to which some suppliers are used. Operators with equipment from these suppliers in their network are having to replace prior kit in order to use new vendors for 5G.
Lack of innovation. Operators have become increasingly frustrated at the lack of innovation in the provision of mobile services. Arguably, the last round of innovation dates from 2005. This is when Huawei was introduced in Europe as a competing supplier. By contrast, Amazon launched Amazon Web Services (AWS) in 2006, and Netflix moved into streaming in 2007. The changes at both companies since then reflect the pace and speed of innovation the digital economy, changes that the telecom industry has largely missed out on.
The 5G era presents opportunities to resolve this. One of the most important is the Open RAN (Radio Access Network). Advances in processing have enabled vendors such as AltioStar and Mavenir to ‘virtualise’ the radio access network, with the software able to run (mostly) on off-the- shelf equipment. This is enabling many smaller vendors to get into the market, and Open RAN proponents claim it will accelerate innovation in the sector. We highlight that if swap-outs are accelerated due to the current embargo on using US chip technology, we believe that the Open RAN ecosystem will be the main beneficiary.
The move to the ‘cloud native’ version of 5G is also fraught with opportunity and risk. It allows for the opportunity to differentiate based on the ability to offer hundreds of new services or ‘slices’, each with bespoke quality of service characteristics; and risk of failure as the complexity proves too hard to handle. The core network is opaque and most operators have offered similar services so far in the evolution of telecoms – we believe 5G and StandAlone will change this.
5 Free to View ● Equities - Diversified Telecoms October 2020
Spectrum – some operators and markets are well ahead
Regulatory policy is important for spectrum allocations, and here North Asian markets such as China, Japan and Korea are ahead, awarding ample spectrum (without auction in Japan and China). The April 2019 US Defence Innovation Board report highlighted the advantage of Asia in the sub 6GHz market
In very select markets some carriers may have some competitive advantage through Views of mmWave vary – we believe this will create combinations or different strategies such as ‘layer cake’ spectrum strategies. mmWave – after a opportunities poor start – is improving, with recent tests suggesting much better range. The increased bandwidth and speed of mmWave creates enormous technical complexity – leading some operators and pundits to dismiss it. We argue that where it can be made to work (and broadly speaking US companies are ahead here) mmWave is genuinely game-changing. We also explore Dynamic Spectrum Sharing (DSS), noting that some operators will be able to provide much better 5G attachment rates than peers via this technology.
What to watch
North Asia, Australia and the US are the regions where 5G is most advanced. Some areas to watch include:
First services using network slicing. In 4Q20 operators will launch Microsoft xCloud mobile gaming services. These will utilise some of the capabilities of 5G, such as Mobile Edge Computing, and lower latency. Gaming is one example of a user group with similar network requirements – such services will be a useful test of both the technology and the demand.
StandAlone network rollout. In 4Q20 operators in China and South Korea will join first movers such as T-Mobile US and Telstra in rolling out StandAlone core networks. The test will be whether these operators can differentiate relative to peers using the Non StandAlone network architecture.
5G service launch using Open RAN technology. Rakuten will launch 5G services using its virtualised Open RAN (Radio Access Network) technology. Many observers believe the much higher processing requirements of 5G will be too challenging for Open RAN, making Rakuten’s launch a key test case.
6 Free to View ● Equities - Diversified Telecoms October 2020
Release 16: the real deal
The rush to market with the early (5G ‘New Radio’) version of 5G has negatively affected views of 5G
This is unfortunate, as the next wave of 5G – Release 16 with StandAlone – brings with it powerful improvements in performance
Yet the transition to a ‘cloud native’ core network is well outside the comfort zone of the industry – and brings risk as well as opportunity
5G so far has been incremental. Release 16 is a step-change, but it’s not without risk
The bull case: The July 2020 final version of Release 16 enables real changes in 5G capabilities, made possible by the end-to-end control of the Standalone network. This will enable operators to offer applications with different quality of service (latency, speed, reliability, etc.), enabling differential prices – essentially operators will reclaim the value in connectivity. At the same time, a move to a cloud native core and more efficient use of spectrum will lower operating costs.
The bear case: There is still no clear business case for the capabilities that Release 16 enables. Enterprise services development will be challenging, requiring bespoke account management and the ability to manage a multitude of new services. Migration to a cloud native core will be difficult, to say the least, for an industry built on proprietary integrations of software and hardware. Finally, the cloud-native quality of 5G – along with network slicing and private networks − creates the possibility that telcos will be disintermediated, with over the top providers capturing the value. In this scenario, telcos will once again be relegated to mere connectivity providers.
All initial 5G deployments have been based on the Release 15 version of the 5G standard. This was rushed to market, as operators and vendors (even countries!) wanted to notch up a world first in 5G. This has had several negative effects. The ‘New Radio’ version of 5G enabled by Release 15 focuses on enhanced Mobile Broadband only – just one of the three main innovation areas of 5G. Users were primed to expect faster download speeds but initial service availability in leading markets such as the US and South Korea was patchy, generating criticism from customers, and pressure from regulators. And even when users found a signal, no application really needed the extra speed – particularly in South Korea where 4G speeds routinely approach 40-60Mbps (OpenSignal, June 2019).
Release 16, finalised in July 2020, is the ‘real’ 5G. Particularly, this enables the broad Release 16, finalised in July 2020, enables the ‘real’ 5G deployment of ‘StandAlone’ 5G. This was previously available, but most operators have waited for the version freeze in Release 16 before rolling it out. The standard improves and refines the benefits of Release 15, and broadens the capabilities of 5G for Ultra Reliable Low Latency Communication (URLLC) services, as well as the Internet of Things.
7 Free to View ● Equities - Diversified Telecoms October 2020
StandAlone requires a cloud-native 5G core, rather than the Evolved Packet Core used in the Release 15 New Radio implementation. Core network migration is an arcane topic, but is critical to delivering 5G services rapidly and flexibly. Operators and vendors have taken different stances with regard to 5G StandAlone, creating opportunity for differentiation. Key here is the transition from a vertically integrated supplier / operator relationship, to a much flatter cloud- native structure. It’s possible that some telcos may struggle with this transition, although others have been working hard on the change for a long period of time. Certain companies – such as Rakuten Mobile – claim that the migration of mobile telecoms to a cloud-based, Internet-like model favours them. It also creates the risk that telco services could be ‘disintermediated’ by enterprise service and software providers.
Release 15: the rush to market has affected perception of 5G
Release 15, which was completed (or ‘frozen’) in mid-2018, primarily focused on enhanced Mobile Broadband, or eMBB. The name ‘New Radio’ (NR) reflected its focus on the radio interface, giving users higher speeds in new spectrum – mostly 3.5GHz in the first phase. This, and the fanfare around being first to launch in April 2019, was followed by a number of disappointed user reports:
Coverage weakness in Korea. South Korea launched 5G services in April 2020, with both operators (and the government) racing the US operators to be the first to launch globally. By end-June more than 1.6m customers had migrated to 5G, but customers were unhappy with both speeds and coverage (FT, 17 July 2019). Operators ended up accelerating their investment plans, resulting in a spike in investment that contrasted with guidance prior to service launch of flat or a ‘slight’ increase in capex.
Signal strength and vulnerability in the US. Initial tests and user experience for Verizon’s mmWave deployments forced a clarification from CEO Hans Vestberg in April 2019 that mmWave is ‘not a coverage spectrum’. The service in Chicago and Minneapolis was initially supported by one Motorola device. Both Verizon and AT&T have been careful since the launch to reframe expectations of mmWave. TMUS used a GIF to show mmWave speeds dramatically affected by a closing door, while journalists struggled to find a signal in central Chicago (ArsTechnica, 9 April 2019).
Generally slow speeds (AT&T, TMUS). TMUS, which had been vocally critical of its competitors, launched 5G with NSA at 600MHz, but initial speeds were not much faster than 4G.
mmWave is not a coverage spectrum Hans Vestberg, Verizon CEO
8 Free to View ● Equities - Diversified Telecoms October 2020
Exhibit 3: US carriers’ 5G speed (Mbps), Availability (% time 5G service availability)
600.0 25.0% 22.5% 494.7 500.0 20.0% 400.0 15.0% 300.0 10.3% 10.0% 200.0
61.0 5.0% 100.0 49.0 0.4% 0.0 0.0% AT&T TMUS (Standalone) Verizon Speed Availability
Source: Opensignal, “USA 5G User Experience”, June 2020. Note: Speed refers to average download speed; availability refers to proportion of time 5G device had 5G connection
Exhibit 4: Wireless revenue relative to wireless data traffic in China: enhanced Mobile BroadBand (eMBB) will offer faster speeds, but operators may struggle to monetise it
1,000 140,000 900 120,000 800 700 100,000 600 80,000 500 400 60,000 300 40,000 200 20,000 100 - - 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 Revenue, RMB bn Traffic, GB m
Source: Company data, HSBC estimates
Perhaps inevitably, the 5G standardisation process became subject to companies and governments pursuing commercial goals, particularly relating to being first to launch 5G in a market. Several entities at 3GPP disagreed with the focus on enhanced Mobile Broadband in the first wave – we believe their concerns have been vindicated by the underwhelming impact of 5G services so far. More broadly, this speaks to the need for the telecoms industry to move to more frequent standard ‘patches’ and updates, rather than the significant releases that were more workable in the early days of mobile technology. A system of regular updates is of course more compatible to the cloud computing paradigm being introduced with 5G:
The first set focused on the Non-StandAlone (NSA) New Radio configuration, giving operators who wanted to launch 5G early the fastest path to market. Critics argued that this resulted in a sub-optimal platform for initial 5G services.
The second wave concerned the Standalone network, and was frozen in September 2018. However, most operators and vendors have deferred SA until the finalisation of Release 16 in July 2020. Operators such as Telstra activated SA in May 2020, but this was rather moot given the absence of SA compatible terminals. Some issues with backwards compatibility were later found in this release.
The third wave of Release 15 was released in March 2019 – the so-called ‘late drop’. This mostly focused on dual connectivity.
9 Free to View ● Equities - Diversified Telecoms October 2020
In retrospect, the first phase of 5G has been underwhelming. The hype cycle was too large for what is (in many markets, in the first iteration) an incremental improvement in downlink speeds. Further, no applications or services need the enhanced bandwidth. The 3G-to-4G migration was rapid in large part because it made streaming seamless. Since then the main focus has been to add capacity: video service resolution (and thus bandwidth requirements) has not changed dramatically.
5G pricing – so far it shows there’s not much that is new
So far, most operators have priced 5G services similar to 4G. Free ‘introductory’ periods are common. Where 5G has been priced at a premium, this is relatively low at c10-15%.
Japanese telcos have offered 5G at no additional cost until later this year.
Telstra offered 5G at no extra cost for the first year, and then implemented price changes in July 2020.
Verizon in the US initially started the 5G service with an additional fee of USD10 over the 4G plan but then waived the surcharge.
Bell Canada also offers 5G access for CAD10 a month on any Bell Mobility postpaid plan and is currently offered as a free bonus until 31 March 2021.
Bouygues Telecom has increase price by EUR1 (+3.3%) for its 80GB package that will be compatible with 5G)
Proximus in Belgium has lifted its price by 16% for its unlimited data pack.
We believe one indicator of the arrival of genuinely different 5G services will be the launch of genuinely different pricing plans.
Release 16 and StandAlone 5G
Release 16, frozen in July 2020, is much more consequential, with both enhancements and completely new features. Release 16 strengthens the two other components of the 5G vision: Ultra Reliable Low Latency Communication (URLLC) and Massive Machine Type Communication (mMTC). Key to this is the StandAlone network which – as the name implies – enables 5G from end to end. StandAlone was substantially codified in Release 15, but operators have chosen to wait for the final version in Release 16 before launching – in part due to the lack of enabled handsets.
One of the real benefits of Release 16 being finalised is that operators will now start to move ahead with ‘StandAlone’ 5G deployments. As the name implies, StandAlone allows for an end- to-end 5G connection, unencumbered by older standards and protocols. Whereas 5G New Radio focused on the user plane and depended on the 4G Evolved Packet Core for the control plane, the StandAlone 5G network manages user and control plane on an end-to-end basis, allowing greater control.
Whereas 5G New Radio focused on the user plane and depended on the 4G Evolved Packet Core for the control plane, the StandAlone 5G network manages user and control plane on an end-to-end basis.
10 Free to View ● Equities - Diversified Telecoms October 2020
Launch timing While the 5G Standalone standard has been available since the second release of Release 15 (with operators such as Telstra announcing a ‘world-first’ launch in May 2020), most operators have chosen to wait until Release 16 was frozen in July 2020. TMUS was the first to launch commercial services on a StandAlone network on 4 August 2020. Only two handsets supported the service initially, and only then after a software update.
There is a now a clear division between operators moving quickly to StandAlone and those Operators have – already – taken very different positions planning to stick with NSA. Operators choosing to migrate rapidly to StandAlone do so to enable on the utility of StandAlone new services and more efficient use of spectrum: Enabling new services. Network slicing, Quality of Service guarantees, low latency, etc. are all better enabled on a StandAlone network, where ‘end-to-end’ 5G is possible. Consequently, operators that are more bullish on the new features offered by 5G will be among the first to migrate. Those who are more cautious, and aim to get the most return from their 4G investment, are most likely to stick with NSA. Most European operators are in this camp.
Coverage and spectrum considerations. StandAlone means operators are not restricted to mid-band LTE coverage (for control plane purposes). This is important for some operators, such as TMUS, as handsets using its 600MHz 5G spectrum no longer need to connect to mid-band LTE to operate. It estimates a c30% improvement in coverage and speed as a result. SA also allows for improvements in carrier aggregation (across 5G bandwidths) and a more scalable approach to DSS. These are significant benefits in larger markets and for those operators with a variety of different spectrum.
Other operators are taking a ‘wait and see’ approach with StandAlone, monitoring the market and looking to maximise the utility and therefore return on the 4G network. Reliance Jio, for example, is promoting a core network version known as Option 6, which is a cost-efficient migration to 5G with a focus on mobile broadband.
Etisalat (UAE), the incumbent in UAE, is taking a phased approach to 5G rollout to maximise the RoI on its investment. In the short term, it is focusing on NSA mode. In the mid-term, the company “will focus on high throughput and low latency for applications in the standalone mode when the new 5G core is ready and will be looking at industrial applications such as ports operations for operating cranes and vehicles wirelessly, immersive VR experience in gaming and training.” In the long term, the focus will be on “mission-critical services such as autonomous vehicles and remote surgery in the standalone mode.”
Exhibit 5: Main areas of service development in 5G
Enhanced mobile broadband (eMBB)
GB/s Communication 3D，UHD Video
Cloud Office/Gaming Smart Home Augment Reality
Industrial Automation Voice Highly reliable Application, such as Smart City Mobile Healthcare Autonomous Vehicles
Massive machine type Ultra-reliable and low latency communications (mMTC) communications (uRLLC)
Source: ITU
11 Free to View ● Equities - Diversified Telecoms October 2020
Exhibit 6: Technology characteristics across telco generations
Source: 3GPP
StandAlone rollout implications
StandAlone requires dedicated 5G core network functions. These are cloud-native in nature, and operators will likely end up with several ‘cores’ for different network functions relating to the Internet of Things (IoT), enterprise, consumer, etc.
The key step forward in 5G StandAlone is the move to a ‘cloud-native’ configuration. This may The StandAlone core is ‘cloud-native’: migrating create problems for telcos and vendors more used to a legacy architecture, and opportunities existing services will be for companies with software expertise. tough Separating the two isn’t easy. What goes on in the core network is by definition out of sight, and issues with cost overruns and poor execution are not immediately apparent. Clarity isn’t helped by the ‘cloudification’ of press releases: were these to be believed, every operator would by now be a cloud leader. We believe it is hard for third parties to verify and corroborate claims of efficiency from network function virtualisation and other cloud-based procedures.
Several operators have begun the virtualisation process – but this has often happened in silos with few of the benefits of a fully cloud native solution. This involves flexibility and elasticity with workloads (e.g., enabling them to happen at both the core and the edge), ‘self-healing’ automation, and other factors that should help reduce the cost and manage the complexity.
It’s also clear that StandAlone implementations will look very different according to the market – TMUS had an initial focus on improving coverage, whereas the next launches in China and South Korea (where sub-6GHz coverage is now well developed) are likely to focus on enterprise applications.
12 Free to View ● Equities - Diversified Telecoms October 2020
One way to think about StandAlone is that it allows for multiple core networks – which can be tailored to individual user requirements.
We believe the key points relating to the StandAlone configuration are:
Flexibility for Mobile Edge Compute (MEC). The ability to transition different workloads and functions through the network is critical to realising the promise of Mobile Edge Computing.
Flexibility for cost savings. Traditionally, mobile network workloads have been relatively inflexible. A cloud-native core will provide the option to allocate and reassign these workloads to the most cost-effective venue, reducing opex.
Scalability. The cloud-native architecture should enable operators to ‘spin’ services and functions up and down much faster than in previous implementations.
Possible risks The introduction of the cloud-computing paradigm to a highly complicated set of technologies with hundreds of legacy services is fraught, to say the least. Potential issues include:
Network outages. Ensuring that the new architecture supports legacy services is highly complex. Operators such as Telstra are solving for this by running parallel IT stacks, and decommissioning older platforms very slowly.
Vendor interoperability. The shift to a cloud-computing, virtualised landscape has opened the doors to many smaller equipment and software vendors, as well as large providers of common off-the-shelf (COTS) equipment. This compares to the relatively small vendor landscape in 4G, which was dominated by Huawei, Ericsson, Nokia and ZTE. This creates potential issues in managing many more suppliers than was the historical norm.
Security risks. Migrating to a new cloud architecture may create security risks for telcos, who have been in part protected by the highly customised, integrated software / hardware solutions provided by their vendors (although not obviously, if the vendor itself is considered a risk).
Some vendors are seeking to mitigate those risks by offering a 5G core alongside the evolved packet core. Ericsson for example offers a dual-mode core for 5G, potentially giving operators more flexibility to manage growth on the 4G business, alongside new 5G functions. It supports the 4G Evolved Packet Core (EPC) and 5G Core network functions on a common cloud-native platform. This reflects operator concern over cost, with operators looking to leverage the Evolved Packet Core and their attempts at virtualisation so far while transitioning to StandAlone. Ericsson claims that the dual-mode scenario gives operators the chance to freeze investment in the EPC, and migrate to a full StandAlone deployment over time.
Virtualisation – it’s hard
Operator core network virtualisation is hard to get a clear view on. Many operators talk about virtualisation, relatively few provide clear markers of progress. One that does is AT&T, which outlined its virtualisation plan in 2013 as part of its Domain 2.0 initiative. AT&T aimed to have 75% of its core network functions virtualised by 2020 – it had reached 30% by end 2016, and is on track to hit its 2020 target, ending 2019 at 66%.
13 Free to View ● Equities - Diversified Telecoms October 2020
It notes that the last 10% includes network functions whose complexity makes them difficult to virtualise. Part of the problem is that vendors have designed a host of virtual network functions that work on different infrastructures. There is relatively little standardisation, which means these can become proprietary to different vendors and even operators. Part of the solution is to work with open source frameworks such as OpenStack, or the Akraino Edge Stack offered by the Linux Foundation. AT&T is also working with other operators to develop standardised Network Function Virtualisation infrastructure (NFVi).
Rakuten case study: Rakuten Mobile’s initial core network was designed around virtual machines. In June 2020 Rakuten announced that it would work with NEC to develop a container-based 5G core for 5G, to be offered as part of the suite of services on Rakuten Communications Platform. A container is a standard unit of software that packages up code and all its dependencies so the application runs quickly and reliably from one computing environment to another.
Rakuten has confirmed a move away from virtual machines to containers in its 5G deployment. This is also something that has only become feasible recently (although we believe Rakuten is – once again – well ahead of the curve). When Rakuten started building its 4G core strategy in 2019, a virtual machine approach was really the only option. Rakuten outlines the difference between virtual machines and containers as follows.
The 5G core is designed from the start as an architecture based on micro-services, using a Kubernetes platform. So Rakuten is essentially developing a new platform: a change from its original claim that it was developing a 5G StandAlone-compliant core from the start of 4G deployment. Such a change is indicative of the benefit that Rakuten sees in moving to a container-based platform.
AltioStar argues that a container network function (CNF) approach brings better flexibility and scalability. In particular, the Control Plane User Plane Separation is enhanced, allowing operators to size deployments appropriate to the initial customer base, and then grow or change the service according to customer needs.
The partnership with NEC is based around source code sharing and joint development. Rakuten benefits from the source code developed by NEC, while NEC should benefit from being part of the RCP for licensing overseas, given its limited footprint in the telecom equipment market. The two companies will develop a converged 4G and 5G core, suggesting that the existing Rakuten core will be collapsed into the 5G core when ready.
We note that Dish in the US also plans a container-based solution, following its July 2020 confirmation of VMware as the vendor for its cloud platform. For incumbent operators, migrating the 4G core onto the 5G cloud-native platform is going to be a challenge. Even advanced operators like TMUS don’t plan on starting this until 2021.
Exhibit 7: Virtual machine vs Container approach Virtual Machines Containers Heavyweight Lightweight Limited performance Native performance Each VM runs in its own OS All containers share the host OS Hardware-level virtualisation OS virtualisation Startup time in minutes Startup time in milliseconds Allocates required memory Requires less memory space Fully isolated and hence more secure Process-level isolation, possibly less secure Source: Rakuten
14 Free to View ● Equities - Diversified Telecoms October 2020
Disintermediation risk – software (continues) to eat the world
In September 2020, Vodafone UK CTO Scott Petty announced that Vodafone was expanding “from being a network service provider to a company that is also a software and IT business” (TechRadar, 4 September 2020). This idea, while plausible, also raises the opposite risk – that telcos will be ‘disintermediated’ by software companies using cloud services over a dedicated network ‘slice’ or private spectrum. Effectively, software and IT companies may become connectivity providers. Risks to the traditional telecom operators include:
National, rather than global, scope of operations. In part due to onerous regulation, expansionary trends from telcos have been cut back. There are few operators with a multinational footprint, and none that can match the Internet companies.
Lower ability to invest due to stagnant revenue growth. Telco investment, while huge, does not match the growth rate of investments in R&D, infrastructure and innovation by the Internet giants. This is a function of the outlook for revenue growth: telcos are nearly always trying to reduce or control capex, whereas the big Internet companies see investment in R&D, hyperscale cloud facilities, etc. as a source of competitive advantage.
Cloud newcomers, not natives. As cloud RAN proponents point out, wireless network technology has been dominated by integrated hardware / software products from a small group of core companies for many years. Telcos are captive clients and – for the most part – not able to easily switch equipment. This creates execution risk as the sector introduces the first ‘cloud native’ technology platform in 5G: lack of cloud background coupled with a proliferation in services to manage could be difficult.
In many respects, telcos are behind some of their enterprise clients in making the move to cloud architectures. Those making a ‘cleaner’ transition to a cloud environment may well have an advantage in developing and implementing cloud services for enterprise clients. This is part of the promise of greenfield operators such as Rakuten Mobile and Dish. Rakuten has long argued that its historical background in software, rather than the traditional integrated hardware and software of mobile telecoms, will give it an advantage in 5G.
Either way, the complexity of the move should not be under-estimated. One problem here for Rakuten Mobile claims that analysts and investors is the difficulty in getting a real insight into an operator’s migration to the its software heritage will give cloud. In our view, many virtualised network functions have ended up being semi-proprietary to it an advantage in 5G one vendor, and therefore operating in isolation, bringing little of the scalable benefits originally promised.
15 Free to View ● Equities - Diversified Telecoms October 2020
Opportunities for operators to differentiate
Service differentiation:
Network slicing. Slicing allows operators to offer bespoke combinations of latency, throughput and priority to different user groups. Potentially, this enables pricing power, with operators reclaiming some of the value in the network. The risk to this thesis, especially in enterprise IoT, is that it is undermined by Systems Integrators or other third parties using spectrum reserved for enterprises to offer the same services. We also note the risk from execution in the core network. Operational system execution needs to be flawless to handle the adaptive, real-time resource allocation needed for 5G network slicing.
Mobile Edge Computing for gaming and the enterprise. Verizon claims that it has a sustainable MEC advantage relative to peers due to its end-to-end control of the network, including fibre backhaul and backbone. This means that developers can rapidly provision applications for the edge network without having to ‘re-architect’ their code. All the AWS initial partners have fixed line assets (albeit to varying degrees). Verizon is planning a phased move to StandAlone. In August 2020 Verizon claimed a global first fully virtualised 5G data session in a live network. Interestingly, it did this using its virtualised EPC, with Samsung providing virtualised Distributed Unit and Centralised Units. Microsoft has partnered with operators such as SK Telecom to launch its xCloud gaming service on an exclusive basis in the initial phase.
Spectrum and coverage differentiation.
Ability to use low-frequency spectrum. T-Mobile US is the clearest example of this. It activated its Standalone core in August 2020, enabling a substantial coverage boost on its 600MHz 5G network (which no longer needs to rely on mid-band LTE for control functions). AT&T and Verizon are likely to activate their SA cores in late 2020 or early 2021. TMUS has an advantage as Verizon has temporarily slowed down its 5G rollout as it expects to win CBRS spectrum in pending auctions. A pause would allow it to update sites for all frequencies later in the year or in early 2021.
Greater flexibility with Dynamic Spectrum Sharing (DSS). DSS (which we discuss on page 31) allows for the flexible use of 4G and 5G spectrum. This capability is enhanced with Release 16 StandAlone, allowing for additional flexibility in aggregating different spectrum carriers.
16 Free to View ● Equities - Diversified Telecoms October 2020
5G – new services
StandAlone to enable a host of new 5G services on bespoke network ‘slices’ – launches to gather momentum in 2021
With a differentiated service comes the ability for bespoke pricing – a potential game-changer compared to commoditised services today
The enterprise opportunity is the largest, but we look to developments such as cloud gaming as a proof of concept for 5G
Network ‘slices’ for different users are a principal benefit of 5G
The bull case: 5G will begin the move away from the relentless commoditisation of connectivity. This began with 3G and accelerated in the 4G era, with market structure and regulation meaning that price has been the main marketing tool. Network slicing will enable operators to offer services according to payment and preference. This will stabilise and grow revenue in the consumer sector, as well as opening up new target markets in the enterprise.
The bear case: 5G so far has seen no tariff innovation. Connectivity will remain the main focus of pricing, and market structure and regulation will result in continued downwards pressure. Where 5G enables new capabilities, such as Mobile Edge Computing, cloud providers such as AWS (which already have the customer relationships) will command the premium. Operators lack the international scale or the investment budget to compete, and now there is an additional risk from Low Earth Orbit (LEO) satellite services such as StarLink.
In this section we explore some of the services that are enabled by Release 16 and the migration to a StandAlone 5G architecture. One of the conceptual challenges for investors looking at 5G is there isn’t one single capability enhancement or driver of demand. In this respect it is quite different from 4G which delivered one service over and above all others: the 3G promise of ubiquitous streaming on the move.
One area we expect to see more of is Mobile Edge Computing (MEC). Rather than a service in itself this is an enabler, bringing the benefits of cloud computing (flexibility, scalability) to services for both consumers and the enterprise. Low latency is certainly enabled by having processing functions closer to the end-user, but this allows for thin-client computing, as well as cost savings. We expect cloud gaming offers to be a proof of concept for MEC, rather than a meaningful revenue generator in their own right. While telcos have plenty of infrastructure (such as local exchange buildings, tower sites) that is suitable for MEC, they may be upstaged by the hyperscale cloud providers, who have existing contracts with customers that they are seeking to extend to the network edge.
17 Free to View ● Equities - Diversified Telecoms October 2020
Fixed Wireless Access (FWA) is a potential 5G service that has been through its own hype cycle. It featured heavily in initial operator plans in the US and Australia, but was put on hold as the sensitivity of mmWave signals and the cost of installation become apparent. We see opportunities for FWA in certain settings, and highlight Australia, Hong Kong and the US as potential opportunities.
In this section we discuss a potential competitor to 5G: Low Earth Orbit (LEO) satellite services. After decades of bankruptcies, LEO finally looks set to succeed, with the SpaceX backed StarLink in the lead. Satellite broadband can take advantage of space vacuums to reduce latency – potentially competing with 5G in niche areas where this commands a premium. It could also address the rural broadband market in areas such as the US. This hasn’t been a focus so far for the telcos, but low frequency deployments such as TMUS at 600MHz make it more feasible.
We also explore the potential for operators in the connected vehicle market, known as Cellular Vehicle to Everything, or C-V2X. We focus on China, where authorities are reducing red tape and expanding test areas (20+ provinces) for private enterprises to continue intelligent connected vehicle (ICV) development. The early focus of C-V2X applications is in taxis, as numerous companies have received government approval to install C-V2X infrastructure on roads to conduct live robotaxi trials. While we expect this area to grow steadily, we expect most of the initial vehicle intelligence to focus on sensors and in-car entertainment, with relatively few opportunities for telcos.
Mobile Edge Computing and Network Slicing
Mobile Edge Compute (MEC) refers to the distribution of processing power to the edge of the network, allowing for lower latency and better, more efficient processing of certain tasks in either the enterprise or for consumers. It has yet to be deployed in a commercial setting, but the first launches of Microsoft’s xCloud gaming service are expected in 2H20. Network slicing is the separation of the network into different ‘slices’ offering bespoke levels of bandwidth, latency, and service priority. The services are in many instances complementary, and we expect them to be used together.
Exhibit 8: 5G use case evolution
Source: World Economic Forum
18 Free to View ● Equities - Diversified Telecoms October 2020
Mobile Edge Compute (MEC) Telcos (and tower companies) have a head-start here with the potential to re-use existing assets − kerbside units, local exchanges no longer needed because of copper to fibre migration, cabinets at towers.
However, some of these locations are likely too small for initial edge applications. Initial ‘edge’ is likely to be 10-20km from the customer (i.e., in the same city rather than in the one-or-two per country hyperscale server farm), not the 1-2km that telco infrastructure may enable. SK Telecom, for example, has eight edge centres for the whole of Korea. Rakuten Mobile plans 4,000 edge sites to power its virtualised radio network, although many of these may lack space for extensive edge deployments: their primary purpose will be to run the virtualised radio access network functions.
It’s likely that telcos can't match what the hyperscale providers (Alicloud, AWS, Azure, etc) offer in terms of R&D heft, flexibility and scalability. These operators typically use specialist data centre companies (such as Equinix, GDS, etc) due to their scale and the interconnection they offer. Telco assets often – but not always – tend to be more focused on the private cloud market, both smaller in scale and without the power efficiency and interconnection offered by the specialists (and required by the hyperscale providers).
We think this informed the decision of leading telcos to partner with Amazon Web Services (AWS): The telco tie-up with Amazon SKT in South Korea, KDDI in Japan, Verizon in the US, and Vodafone in Europe. Announced in Web Services creates disintermediation risk December 2019, the AWS Wavelength press release announces that it ‘embeds AWS compute and storage services at the edge of wireless telecommunication providers’ 5G networks, enabling developers to serve use-cases that require ultra-low latency like machine learning inference at the edge, autonomous industrial equipment, smart cars and cities, Internet of Things and Augmented and Virtual Reality’. These, of course, are many of the areas that the telcos are hoping to monetise in 5G…
For developers, many of whom will already be working with AWS, this provides a relatively pain- free route to market for edge applications. AWS is nearly global in scope, and using Wavelength means developers don’t have to work with telcos in an individual market to resolve many of the local differences relating to the transport layer and other factors.
MEC applications should start to hit the market in 2H20 (notably the Microsoft backed gaming initiatives in Korea, Australia, etc). This should provide proof of concept from a technology point of view, rather than significant revenue. It’s far from clear that customers will be willing to pay extra for a low latency gaming service given the wide availability of free services. We also note that in many of the initial launch markets (e.g., Korea) the basic 4G and 5G cellular experience is already extremely fast.
Exhibit 9: 5G network slicing
Internet Entertainment Broadband slice Automation Mission critical IoT slice Medical
Massive IoT slice Retail and logistics Voice slice Smart city
Voice Communication
Source: HSBC
19 Free to View ● Equities - Diversified Telecoms October 2020
Latency improvements The latency improvement with 5G is often discussed in terms of up to 10ms. However, it’s important to clarify what is being referred to. A latency of 1-5ms might be recorded in the radio network, but is of little use if the round-trip journey in the core network takes a long time. This is where Mobile Edge Computing and network slicing come in, enabling operators to apportion mission critical processing closer to the customer, in a network slice with parameters applied proportional to their needs and budget.
Network slicing – a key enabler of service differentiation and tariff innovation The ability to offer different latency, bandwidth and priority to dozens or more of different user groups is one of the main potential benefits of 5G. Potential use-cases include:
Emergency services, first responders;
Consumer gaming to reduce latency;
Live gaming to augment stadium experience with live stats, AR and real-time data;
Industries: Mining, Industrial and Electronics around safety, automation and quality control.
Much of this will depend on how robust the operator’s system and implementation is – managing services to defined Quality of Service (QoS) criteria across multiple different slices is vastly different from the ‘best efforts’ to all basis of most current network management. There is scope within 5G for multiple different slices. Examples include:
Infrastructure slices: These may focus on FWA, low latency, high privacy, etc.
Consumer slices: Potential focus areas include AR/VR, video streaming, online gaming, live gaming stadium.
Enterprise slices: these could be configured for various types of industries and companies. A bespoke slice for each enterprise client is a possibility with 5G, resulting in hundreds of potential slices.
Fixed Wireless Access – back on the agenda again
Initially, operators in Australia and the US were keen to use 5G as a fixed wireless access (FWA) service, given the high cost of installing fibre to detached residential premises in these markets.
Australia In Australia, progress has been relatively slow. Some providers have had to shift strategy after recent government bans in 5G network equipment suppliers. Other operators have been more cautious, but have started indicating they would start to offer FWA solutions to specific households.
Hong Kong – SmarTone SmarTone has dusted off its 3G playbook for its 5G Home Broadband service, announced in September 2020. It will use mobile spectrum to offer ‘fixed wireless access’ – essentially a 5G cellular connection to a Wi-Fi 6 router. This could offer speeds from tens to hundreds of Mbps. SmarTone will use its mid-band spectrum for the service initially, although mmWave may be an option later in 2021.
20 Free to View ● Equities - Diversified Telecoms October 2020
The United States − Verizon Verizon, in the coming months, is banking on adding a more powerful 5G Home customer premises equipment (CPE) outfitted with new chipsets that will support self-installs and help accelerate its customer acquisitions for home 5G. Verizon CEO, Hans Vestberg, at a recent investor conference, reiterated that Verizon plans to have 5G Home operating in more than ten markets by year’s end. Verizon’s current 5G Home markets include Houston, Detroit, Indianapolis, Los Angeles, Sacramento, Los Angeles and Chicago.
Verizon plans to deploy 5G Home to about 30 million US households in the next five to seven years. CEO Hans Vestberg anticipates that 5G Home – with speeds up to 300 Mbit/s, will likely compete with wireline rivals and that the fixed wireless offering will be “less complicated” than cable broadband. Nonetheless, we also note that Matt Ellis, Verizon’s CFO, in a recent investor conference has stated that it does not expect 5G Home to be a material top-line contributor in 2021.
Competitive risk from Low Earth Orbit (LEO) satellite deployments
In the past several months, the Low Earth Orbit (LEO) satellite proposition has become increasingly credible. SpaceX-backed Starlink seems to be furthest ahead, but other entities include Amazon (branded Kuiper) and even the UK government after its investment in OneWeb.
Starlink – the furthest ahead Starlink launched 60 LEOs from one Falcon in early 2019 − aiming for 1,800 by end-2020, and 12,000 in total. It could have 15% operational by end-2020. It claims that with 400 in orbit most of North America will be covered, while with 1,400 it can cover most of the global population. With 1,800 this includes the poles − additional coverage will be mostly to add capacity.
Bandwidth of 11-60 Mbps in initial trials. We note that the bandwidth per cell won't be high enough to serve an urban market, meaning that the core business proposition of telecom operators is unlikely to be affected. GEO / LEO could also be used to meet backhaul requirements of 5G networks.
Cost – claimed total investment of cUSD10bn. Cost has been a significant issue for other LEO operations, but Starlink claims a cost of USD300k per satellite. It also benefits from an ability to launch from its own Falcon rockets − something few other LEO companies have (although Amazon’s Kuiper will also have this capability). Starlink plans a complete investment of cUSD10bn. We note that the LEO satellite lifecycle is relatively short at seven years.
Latency advantage. Starlink’s 25-35ms latency is similar to current fixed broadband for up and downlink, but over longer distances it benefits from the speed of light being 47% faster in a vacuum (i.e., space) compared to glass (i.e., fibre optic cable). Starlink advocates note that New York to London latency is 76ms using fibre optic cable, or c60ms using the privately-owned Hibernia Express cable. This was built at a cost of USD300m to shave 3ms from the previous fastest latency of 63ms. The latency for Starlink for NY to London could be as low as 43ms. This compares to the latency lags that are a big issue for traditional satellite broadband, where high orbits create up and downlink latency of 240ms or more.
Coverage. Each satellite can cover 500km with an angle of 81 degrees.
LEO has potential latency and coverage advantages, creating opportunities for a ‘barbell’ Very low latency services offered by satellite might help approach − premium services for high frequency traders (for example) and basic broadband defray operational costs services to users in rural areas. There may be some cannibalisation of 5G, but neither of these is a growth market for telcos in the near term.
We see limited near-term impact to telcos, but potentially this takes away medium-term growth possibilities in rural areas. We also see some potential for revenue cannibalisation in urban areas.
21 Free to View ● Equities - Diversified Telecoms October 2020
IoT: risk of limited revenue, increased complexity
Release 16 brings substantial enhancements to the IoT market: 1 million IoT devices per square km can be supported, with improved power and connectivity controls. Similarly, the low latency and quality of service guarantees are potential enhancements to the Cellular Vehicle to Everything market, known in acronym parlance as C-V2X.
We remain cautious. There are now millions of IoT sensors deployed in markets such as China. However, these are low bandwidth sensors that don’t require much in the way of network smarts, and narrowband IoT working on older generations of technology is perfectly appropriate. Low value-added means low revenue – IoT revenue of RMB5bn accounted for just 1.4% of China Mobile revenue in 1H20, and had declined 3% YoY due to price competition.
While Release 16 brings improved IoT capabilities, there is also scope for network overload. An IoT network with tens of thousands of devices, each executing micro-transactions, will lead to high control plane traffic compared to the total user plane data, which could be minimal for most IoT solutions. Without the appropriate core network capability, MNOs might struggle to handle this traffic alongside users streaming video, which generates the opposite: very high user plane, but limited control plane traffic. Dealing with this complexity is one reason why Rakuten has moved away from Virtual Network Function platform for its core, towards a container-based approach, which it claims offers enhanced flexibility and scalability (see page 15).
Cellular Vehicle to Everything (C-V2X) – are we there yet?
Similarly, connectivity in vehicles has been growing: AT&T serves c.37 million cars out of an addressable market of 285 million total vehicles. Yet the problem is the same as for IoT: limited ability to differentiate. While silicon and semiconductor companies such as Qualcomm and Arm (not listed) are very bullish on the vehicle market, a glance through their projections reveals a focus on the cockpit, on in-vehicle entertainment, and other areas that don’t need or require cellular connectivity.
Representative services launched so far are illustrated below:
Exhibit 10: Select connected car pricing Company Company Type Connection Name Connection Type Lowest Monthly Pricing AT&T MNO Harman Spark Aftermarket Plug-In 10USD T-Mobile MNO SyncUP DRIVE Aftermarket Plug-In 10USD Verizon MNO Hum Aftermarket Plug-In 10USD NTT Docomo MNO In Car Connect Embedded 1500JPY (13.6USD) Telefonica MNO Movistar Car Aftermarket Plug-In 3EUR (3.6USD) Tesla OEM Premium Connectivity Embedded 9.99USD Hyundai OEM Blue Link Embedded 8.25USD Audi OEM Connect Embedded 10USD General Motors OEM OnStar Embedded 14.99USD Apple OTT CarPlay Aftermarket Tether Free Google OTT Android Auto Aftermarket Tether Free Deutsche Telekom MNO Car Connect Aftermarket Plug-In EUR14.95 Source: Company data
T-Mobile reported a USD100m increase in revenue during the first year of operating SyncUP DRIVE. Other entities are monetising the data collected by cars. General Motors has highlighted high margins at the OnStar business, reporting aftersales contributions of USD3bn to USD4bn in 1Q20.
How will 5G affect the car? The advent of 5G will specifically help and enable the development of C-V2X. Cellular Vehicle- to-Everything (C-V2X) is the network over which cars will communicate with everything (pedestrians, vehicles, infrastructure, and network), which will enable intelligent transport. An
22 Free to View ● Equities - Diversified Telecoms October 2020
example use-case is identifying dangerous objects that are not in the line of sight, where either other cars or a nearby road-side unit (RSU) with sight on the object communicates the information to the endangered vehicle, allowing for a prompt response like braking immediately.
5GAA (5G Automotive Association) has identified in total 12 C-V2X potential use case examples, with a majority of cases requiring network capabilities that the 5G network brings. Others are also identifying and testing unique 5G-enabled C-V2X use cases. Huawei, one of the leading OEMs in developing C-V2X technology, highlighted 27 use cases in its city-level C-V2X project in Wuxi China.
Exhibit 11: Select C-V2X use cases that see benefits from 5G Use cases Range (m) Amount of data Latency (ms) Reliability (%) required Cross-Traffic Left-Turn Assist 350 300B per message 100 90.00% Cross-Traffic Left-Turn Assist 350 1000B per message 10 99.90% Intersection Movement Assist 100 300B 100 99.99% Emergency Brake Warning 360 200-400B 120 99.99% Emergency Brake Warning 290 200-400B 120 99.99% Software Update 100 1.5GB 30s 99.00% Remote Health Vehicle Monitoring N/A 1KB 30s 99.99% Hazardous Location Warning 300 300B 100 99.00% High Definition Sensor Sharing 80 1000B 10 99.90% See-Through for Pass Manoeuvre 100 15Mbps 50 99.00% Lane Change Warning 83 300B per message 400 99.90% Lane Change Warning 28 300B per message 400 99.90% Lane Change Warning 51 300B per message 400 99.90% Vulnerable Road User 80/150/300 Depends 20 99.90% Vulnerable Road User 80/150/300 Depends 20 99.90% Source: 5GAA
A key obstacle to the development of a C-V2X market is in the business model. There is no doubt that 5G could dramatically improve safety and convenience on the roads – but how to pay for the network installation on roads?
In terms of infrastructure, roadside units (RSUs) are important in the C-V2X ecosystem, as they can relay information beyond line of sight. With an effective transmission range of 500m for a RSU (WCTR Shanghai, 2016), a single RSU can cover 1km of road. For units handling transmissions in the mmWave, more RSUs to cover for a lower effective range may be needed.
It’s also possible that this model of connectivity and safety may be replaced by an on-board Tesla’s cars currently have 8 cameras, 12 ultrasonic system, such as that being built by Tesla. However, this requires considerable processing sensors, and 1 front-facing power in the vehicle itself – an extension of the thin-client debate also common in other areas radar that all feed into its with regard to 5G. WeRide, one of China’s leading AV companies, has highlighted that 5G’s low autopilot system latency capabilities can allow for some computing capabilities to be handled off-vehicle (like in the cloud), which can reduce hardware costs for OEMs.
China’s ‘New Infrastructure’ plan
The lack of a clear business case means that top-down markets like China are likely to press ahead first. The Chinese government has released a development plan which aims to prioritise the buildout of smart infrastructure. The “New Infrastructure” plan highlights a cRMB10tn national investment into seven digital infrastructure areas within the next five years: 5G base stations, AI, City rail systems, Data centres, Industrial IoT, NEV charging stations, and UHV power transmissions. These investments will be made in conjunction with investments from local governments with the aims of stimulating the economy in the short term, as well as the country positioning itself to become a leader in technology and innovation. Network coverage, NEV charging stations, and industrial applications (such as IoT platforms and smart factories) seem to be the early focus, with numerical targets set by authorities.
23 Free to View ● Equities - Diversified Telecoms October 2020
Exhibit 12: Select local authorities’ 3-year action plan targets Beijing Guangzhou Shanghai New 5G stations 13k (in 2020) 80k 34k New EV charging stations 50k 50k 100k Industrial IoT/AI applications 20 smart factories 10 AI industrial parks 100 unmanned factories 300km of smart road for Smart city platform 150k enterprises on cloud platform connected cars Source: Xinhua
Corporates are also following suit, with the major internet companies (Baidu, Alibaba and Tencent) having announced plans to invest in the areas the government has highlighted. Alibaba and Tencent so far have announced a combined RMB700bn investment into the space, with a common theme of improving data centres, cloud capabilities and AI.
Having connections to five of the seven areas of focus within the government’s strategic plan, the growth of intelligent connected vehicles (ICVs) (which includes autonomous vehicles) looks to increase. Authorities are focusing on reducing red tape and expanding test areas (20+ provinces) for private enterprises to continue intelligent connected vehicle (ICV) development. The early focus of C-V2X applications looks to be in the taxi space, as numerous companies have received government approval to install C-V2X infrastructure on roads to conduct live robotaxi trials.
Ride-hailing company, Didi Chuxing, deployed C-V2X infrastructure and robotaxi services in Shanghai in June 2020.
WeRide has also deployed C-V2X infrastructure and robotaxi services in Guangzhou in November 2019. WeRide has also reportedly partnered with China Unicom to work on 5G communications technology.
Baidu has similarly deployed a pilot program in Changsha in September 2019.
Whilst not an early focus within C-V2X, industrial applications for smart mobility are also being seen. Heavy smart truck (Level 4) trials in shipping ports have been reported to be conducted by Quectel, SAIC Motor, Shanghai International Port, and China Mobile. Given that potential cost savings and productivity enhancements from automating port processes can amount to 55% and 35% respectively (McKinsey, 2018), developing and integrating C-V2X technology is a possible option for port owners.
In our view, the technology looks more likely to be deployed in areas where travel routes are defined, and where public traffic is not present. This reduces complexity and may help with the automation of routine tasks. Airports and ports are potential candidates for this.
24 Free to View ● Equities - Diversified Telecoms October 2020
On the spectrum
Divergence in spectrum strategy – a barbell approach of broad coverage and zones of high speed is the most common
Growth of mmWave is expected to accelerate in 2021
Difference in vendor support for Dynamic Spectrum Sharing (DSS) – this could lead to substantial differences in 5G attachment rates
Spectrum portfolio and strategy differences will be more apparent
Bull case: Spectrum allocation has been rapid and plentiful in most markets. The majority of operators have access to both ‘sub-6GHz’ and ‘mmWave’ spectrum (near or at c3.5GHz and 26/28GHz respectively). 3.5GHz has been the initial focus for coverage, and some operators have deployed at lower frequencies. Going into 2021 we expect more focus on mmWave. This delivers a step-change in speed and latency and opens up the potential for new services. It is also more complex, favouring market leaders such as Qualcomm. In certain markets, DSS will improve 5G attachment rates and services.
Bear case: Initial launches of 5G at 3.5GHz in markets such as South Korea has proven relatively similar to 4G. Conversely, mmWave has been too fragile to deliver good performance except in the most protected environments. DSS uses up resources from both 4G and 5G networks, delivering a weaker user experience in both.
Spectrum: sub 6GHz outlook
Most initial rollouts outside the US have focused on sub-6GHz deployments. Given the option, operators have chosen to establish a broad base of coverage before mmWave or other ‘hotspot’ deployments:
US – opposing strategies. The US presents opposing strategies in a single market, with TMUS switching on 5G in its 600MHz spectrum in late 2019, and activating StandAlone 5G (which it claimed as a world first) in August 2020, while peers such as Verizon focus on mmWave connectivity.
Korea – decision to focus on 3.5GHz. 5G spectrum was awarded early in Korea, at both mmWave and 3.5GHz. Operators have thus had the choice in what spectrum to deploy, and have focused almost entirely on sub 6GHz deployments to date. This is partly a function of government scrutiny, with operators heavily criticised early after launch in April 2019 for offering poor coverage. Capital expenditure was brought forward as a result, and mmWave deployment has been delayed.
In most markets, spectrum has been distributed equally – scrupulously so, in the case of Japan. However, we highlight certain markets where spectrum shortfalls or disparities could impact 5G deployment:
25 Free to View ● Equities - Diversified Telecoms October 2020
Shortfall of spectrum:
Australia. Here operators have a maximum of 60MHz in urban areas, compared to standard allocations of 100MHz in most markets.
Hong Kong. Partly because of the risk of satellite interference, Hong Kong’s sub-6GHz spectrum allocation is lower than peers. China Mobile Hong Kong emerges slightly better off, with 80MHz of contiguous spectrum from 3.38-4.46GHz. All operators have smaller allocations at regional peers at different bandwidths. With early handsets not able to support 5G carrier aggregation, Hong Kong telcos are at a slight disadvantage, while having similar spectrum to regional peers in aggregate.
I wake up in the morning, there’s a smile on my face because I’ve finally got a great set of spectrum assets Neville Ray, T-Mobile US CTO
At the other extreme, certain operators and markets stand out for their abundant spectrum allocation:
China. China Unicom and China Telecom are able to share 200MHz of contiguous 3.5GHz spectrum (with an additional 100MHz for indoor use) via their ‘co-build, co-share’ network. China Mobile has 160MHz, but has recently confirmed a sharing agreement, enabling it to access the 2x40MHz at 700MHz frequency owned by China Broadcast Network (CBN).
Japan. In the sub 6GHz range, the two largest operators, NTT DoCoMo and KDDI, have 200MHz of spectrum each, while Softbank and new entrant Rakuten Mobile have 100MHz each. Each operator has 400MHz at 28GHz.
US. TMUS CTO Neville Ray claims that “I wake up in the morning, there’s a smile on my face because I’ve finally got a great set of spectrum assets and I’ve been battling away on this network one way or another for 25 years and we’ve never had the size and scale of this opportunity” (SDX Central, 13 August 2020). The operator has licenses for 319MHz in sub- 6GHz spectrum, and more mmWave spectrum than AT&T. Verizon has the fastest 5G while TMUS offers better coverage. According to OpenSignal’s June 2020 report, Verizon offers the fastest 5G with an average download speed of 494.7 Mbps. However, Verizon has the lowest 5G coverage of 0.4% vs 22.5% for TMobile (standalone) and 10.3% for AT&T. TMUS has been critical of peers such as Verizon relying on DSS to add boost 5G availability, but we note that speeds on the 600MHz network are relatively low (for 5G) at 40-80Mbps. TMUS aims to progressively switch over the Sprint 2.5GHz spectrum, realising speeds of c400Mbps.
Spectrum – mmWave outlook
5G markets thus far have been bifurcated, with only the US – out of necessity – deploying mmWave services at any scale. As noted, these services have been poorly received, due mainly to coverage issues. Nevertheless, we expect mmWave to grow in importance through 2021, overcoming some of the initial scepticism relating to the technology.
mmWave (millimetreWave) refers to high frequency spectrum that has been allocated for 5G. Basically there are two frequency ‘zones’ for 5G:
26 Free to View ● Equities - Diversified Telecoms October 2020
Frequency Range 1 refers to all ‘sub 6GHz’ frequencies. The majority of 5G launches to date have been in this area, and the majority at the 3.5GHz spectrum (used by Unicom and Telecom, the Korean telcos, etc).
Frequency Range 2 refers to the mmWave range between 24-100GHz. So this spectrum is vastly different. Allocations are typically very high (1GHz compared to the 100MHz typical at 3.5GHz). Radio waves in this bandwidth have wavelengths from 1-10 millimetre, hence the name. Because of the extremely high frequency, signal propagation is much weaker: the range is tens or hundreds of metres (like Wi-Fi) but unlike Wi-Fi the signal can be severely affected by rain, leaves, or a user turning his / her back on the antenna. So it will be most suitable for traffic hotspots in public places, in-building solutions, and also line of sight wireless backhaul. Approximately 85% of mmWave spectrum is globally harmonised.
The commercial implications of this are magnified by spectrum allocation in the US. Until The US is behind in sub- 6GHz spectrum, but ahead in recently, the only spectrum available for 5G in the US was mmWave, as the US military uses mmWave most of the sub 6GHz spectrum that has been cleared for 5G use in other markets (clearly expressed in the Defence Board report). This has prompted enormous concern that the US is and will fall behind China in 5G deployment. The allocation of ex-broadcast 600MHz spectrum is helping, particularly for TMUS.
This means that a lot of the Chinese R&D effort has focused on sub 6GHz, while more of the US R&D effort has focused on mmWave. With increased technology trade restrictions now in place, it is possible that the US could develop a lead at mmWave. However, this also depends on deployments. So far, operators with mmWave in Japan and Korea have not prioritised it, preferring to focus on the coverage that is easier to build at sub 6G spectrum. In Hong Kong, Hutchison declined to apply for free mmWave spectrum because it didn’t want to be subject to the coverage obligations.
Infrastructure considerations mmWave will be deployed at a high density, making backhaul and site location very important. This suits certain markets more than others:
Japanese PHS sites are as little as 100m apart with fibre backhaul. These are mostly owned by Softbank, but so far this operator has taken a ‘fast follower’ approach to 5G.
Korea also has a large number of repeaters which could be repurposed for mmWave.
This said, recent developments also indicate broader potential for mmWave:
Integrated Access Backhaul (IAB) deployment. IAB was finalised in Release 16, and gives operators the potential to use spectrum for backhaul as well as access. mmWave is highly suited to this, as it can be deployed between locations with line of sight access. This will enable operators to get 5G sites up and running while fibre deployments are pending, and even replace fibre in some instances.
Extended range. In June 2020, Qualcomm and Ericsson and Casa systems completed a test data call over mmWave in Australia, at a range of 3.8km. While this would be much less in a real-network deployment, it nonetheless indicates that range improvements are feasible, improving the commercial outlook for mmWave.
US companies lead in mmWave We expect mmWave deployments to increase in the next 12-18 months. mmWave is enormously complex – it requires multiple modules in the device, and generates large quantities of real-time signals from the baseband unit to the handset to deal with variables such as coverage and device position. This complexity favours early leaders such as Qualcomm, and the US supply system is broadly ahead on mmWave due to the focus of the US operators.
27 Free to View ● Equities - Diversified Telecoms October 2020
Qualcomm claims the most complete modem to antenna end-to-end solution, and most of its competitors in RF Front-End solutions are also US companies, such as Broadcom and SkyWorks. The complexity of adding mmWave is likely to test the capabilities of competitors.
This is also a potential issue for infrastructure vendors, who so far have different architectures for sub 6GHz and mmWave. So far none has a converged strategy because of the complexity of dealing with the higher frequencies alongside the sub 6GHz spectrum. This may create further opportunities for differentiation on the infrastructure side.
As per the June 2020 Open Signal report Verizon has the fastest 5G speeds of any carrier with an average download speed of 494.7Mbps. But it also has the lowest 5G availability — a paltry 0.4% to 22.5% for T-Mobile’s (standalone) nationwide network. The initial deployment would be concentrated in dense urban areas since even at peak power the mmWave signal can currently only travel a 100mtrs and between 300-500 when used with directional antennas. Buildings, structures and street corners – even rain and leaves – are a problem. Verizon has been experimenting with certain beam forming repeater technologies along with its partners like Movandi, Pivotal, etc. to help resolve these issues.
MTS to focus on mmWave MTS, the leading Russian mobile operator, recently received a license for mmWave spectrum. The company, in its press release, highlighted certain 5G solutions, for corporates, it is currently exploring, including manufacturing (machine vision, predictive analytics, and remote equipment monitoring and control), agriculture, pulp & paper, and oil & gas (drone-based infrastructure inspection and monitoring), healthcare, retail (digitised smart stores and biometric payment systems) and logistics (automated warehouse operations and autonomous vehicles).
Exhibit 13: The ‘layer cake’ approach to spectrum
Internet Entertainment Broadband slice Automation Mission critical IoT slice Medical
Massive IoT slice Retail and logistics Voice slice Smart city
Voice Communication
Source: HSBC estimates
Dynamic Spectrum Sharing – important in the early stage of coverage
What is Dynamic Spectrum Sharing (DSS)? Dynamic Spectrum Sharing enables the flexible use of both 4G and 5G spectrum to be used to enhance either a 4G or 5G signal. Previously, operators had to ‘re-farm’ parts of spectrum to use with another technology. This is a semi-permanent measure, and involves tough choices between users of legacy and new technologies. DSS allows spectrum usage to be adjusted
28 Free to View ● Equities - Diversified Telecoms October 2020
dynamically, according to demand, and at intervals of just 1ms. This is a significant step forward in network economics. Previously, CTOs had to favour investment in one generation over another. DSS means that 4G radios (since 2015, in the case of Ericsson) can be reused for 5G. We note that DSS favours single-vendor deployments, as it requires either a shared baseband card for 4G and 5G (i.e., deployed in new equipment) or adding a 5G card compatible with the 4G card (normally via interfaces that aren’t open).
Most 5G spectrum has been allocated at higher frequencies, with inferior propagation characteristics compared to 2G/3G/4G frequencies. DSS thus allows operators to benefit from existing spectrum allocations, giving customers greater 5G availability. Ericsson has backed DSS the most aggressively, creating a potential opportunity in markets where one operator is most aligned with Ericsson to differentiate in the early phase of 5G rollout. Potential examples include Telstra in Australia, SmarTone in Hong Kong, and Far EasTone in Taiwan.
However, critics of DSS note relatively high overhead, which impairs performance, and broad DSS offers much higher 5G attachment rates, but at some availability from all operators going into 2021, when StandAlone networks and second performance cost generation 5G phones will enable enhanced performance from DSS. DSS will improve dramatically with the third-generation of 5G chipsets, enabling the use of carrier aggregation on 5G on the Standalone network. Third generation chipset handsets will be available in volume by the second half of 2021, and this is when vendors such as Nokia are targeting the deployment of their DSS solution.
How does it work? DSS depends on an operator using compatible 4G and 5G equipment – typically from the same vendor. All Ericsson radios shipped since 2015 are ESS (Ericsson Spectrum Sharing) compatible, with the title of Ericsson’s technology indicative of what it feels is a substantial lead in this area. After a software upgrade, base stations are able to use DSS to assign time- frequency resources on a per millisecond basis to the devices that need them.
Implications from the move to next phase DSS and StandAlone 5G networks The initial phases of DSS linked to New Radio (NR) has a few restrictions, which should be eased with enhancements arriving in late 2020 and early 2021:
Capability to support asymmetric bandwidths – e.g. 20 in 4G, 50 in 5G. Current implementations of DSS support only symmetrical allocations, a drawback given (typically) much larger allocations of spectrum for 5G.
Ability to support different bandwidths compared to current 10, 15, 20 MHz. Again, this will provide additional capacity and flexibility.
3 Radio Access Technology (RAT) supports 3 technologies as opposed to just 2. This would add access to 3G or 2G spectrum.
Nokia believes that DSS with Carrier Aggregation releases the full potential of the technology, especially when combined with StandAlone (SA) 5G architecture.
DSS combined with SA enables access to low frequency spectrum, enabling much enhanced geographic coverage. This is because NSA requires a 4G signal in the mid-band (rather than low-band) to anchor the 5G signal, limiting the coverage improvement.
Carrier Aggregation support in the devices also enables an optimum experience, allowing more efficient use of bandwidth. This should become more prevalent from the second generation of 5G terminals, shipping in 2H20 and 2021.
StandAlone doesn’t use Dual Connectivity in the radio by definition, as there is no connection to the 4G radio. It therefore can’t take advantage of low-band 4G, so the solution is use 5G with DSS to boost 5G coverage
29 Free to View ● Equities - Diversified Telecoms October 2020
DSS could be crucial to bringing the benefits of StandAlone (better network slicing, lower latencies, etc) to a broader area of the population.
What are the potential disadvantages? The signalling overhead results in some impact on DSS. Nokia notes that the minimum would be 5% of LTE capacity, as 5G broadcast takes up 1ms of 20ms in total. Other controls can increase the impact, resulting in lower performance. Keysite Nemo found a 21% overhead compared to LTE only, and a 41% overhead compared to 5G alone. This was based on 20MHz of spectrum with 256QAM and 2x2 MIMO, delivering 195Mbps for 4G and 226Mbps for 56 on a standalone basis, and c160Mbps for both technologies using DSS.
Some industry figures appear to be quite sceptical regarding DSS as a result. TMUS CTO Neville Ray says that DSS ‘kind of eats away on the net capacity of the shared radio’. However, we also note the context, where TMUS is highlighting its dedicated low-band 5G spectrum at 600MHz, compared to peers that are using mmWave and then exploring DSS to expand coverage. We’d agree that in this context DSS isn’t likely to be able to match what TMUS can provide.
Vendor solutions – significant disparity DSS was provided for in outline form at the 3GPP, but vendors have taken different approaches, with Ericsson the most aggressive. The gap between different solutions is likely to be most apparent in the next 6-18 months, after which we expect solutions to converge:
It’s important to note that DSS works as part of a single-vendor solution: the 4G and 4G basebands of different vendors aren’t compatible. This gives single-RAN vendor operators a potential advantage over peers, most evident in markets such as Australia.
Private and unlicensed spectrum implications
The availability and use of private and unlicensed spectrum is expected to increase with 5G, bringing both opportunity and risk for telecom operators. So far more regulators in more than forty countries have separated or allocated spectrum for Enterprise or enterprise users.
Germany. BNetzA has awarded 67 local licences, each covering a single site. They make for a 10-month running total since Germany opened applications for local licences in the 3.7–3.8 GHz range, in a substantial change in approach. Companies such as Bosch, Siemens, Lufthansa, and BMW have applied for spectrum.
The US. The US has allocated CBRS spectrum to private users, but CBRS activity has slowed because of the complexity of the spectrum access system (SAS) in play, to manage overlap and interference between incumbent users.
France. In France, frequencies in the 2.6 MHz band have been offered to metropolitan businesses by regulator ARCEP. Paris airport operator ADP has a 10-year licence, as do Air France and EDF.
UK. Ofcom in the UK has released a tranche of spectrum at 3.8-4.2 GHz for local deployments, requiring national operators to hand over unused licensed spectrum to enterprises; plus, the lower 26 GHz band will be reserved for private and shared access as well.
The Netherlands. The Netherlands has set aside spectrum at 3.4-3.45 GHz and 3.75-3.8 GHz; Sweden has done the same at 3.72-3.8 GHz.
Asia. Asia is behind Europe in this regard, but Japan, Australia and Hong Kong will allocate spectrum for localised 5G businesses in the 3.7 GHz, 26 GHz, and 28 GHz frequency bands.
30 Free to View ● Equities - Diversified Telecoms October 2020
Supply chain disruption
Initial deployments have mostly used 3.5GHz spectrum, where Asia is ahead – we expect increased focus on mmWave in 2021
Dynamic Spectrum Sharing (DSS) will result in very different 5G availability between operators, and help boost return on investment
Private spectrum allocations are growing, creating opportunities for operators, but also risk as enterprises set up their own networks
Tech-tonic shifts
Bull-case: 5G requires a denser network, but operators can reduce capital intensity by network sharing outside urban areas. The growth of Open RAN will usher in a new era of innovation in the industry, with operators able to use open interfaces to put together bespoke networks from different specialists. The era of dominance by a select few integrated hardware / software vendors is over, and any requirement to replace equipment from certain suppliers will hasten that trend.
Bear-case: Network sharing reduces the ability for operators to differentiate from their competitors. This capability has already been affected by supplier consolidation, with the top four vendors essentially delivering on the same technology roadmap. Equipment swap-outs will be funded by cash that could be used to invest in 5G: operators that thought they had 8 years to swap out banned equipment may need to do it in a hurry. A mobile ‘splinternet’ may emerge, with China and the US backing different 5G standards. Chinese supply chain entities may struggle to access and deploy mmWave IP, where US companies have an advantage.
Network sharing: logical in certain markets
Much of the 5G supply chain was positioned for a large increase in telecom spending for 5G. This has materialised in some markets, but not in others. Part of the reason has been accelerated network sharing between operators looking to defray the cost of 5G given an uncertain revenue outlook. European telcos have long been ahead with network sharing, but 5G has prompted sharing in markets where telcos had been against it:
China Unicom and China Telecom. This co-build, co-share network is one of the largest of its kind, and was announced in September 2019. The two telcos have divided rollout between them (easier than working jointly) and are currently working on their revenue sharing arrangement.
China Mobile and China Broadcast Network (CBN). These two entities confirmed plans to co-build a network using the 700MHz spectrum allocated to CBN.
South Korean telcos in rural areas.
KDDI and Softbank Corp (and now NTT DoCoMo) in rural Japan.
31 Free to View ● Equities - Diversified Telecoms October 2020
We believe that network sharing is appropriate in certain markets – the decision of Unicom and Telecom to co-invest makes complete sense given the scale of China Mobile and the terrain they need to cover. This said, complete sharing reduces the ability to differentiate based on network quality. Consequently, we believe the Korean and Japanese telcos are correct to build their own network coverage in urban areas. In larger markets, network quality and coverage can be a durable competitive advantage.
Supplier swap-out: a growing trend
The past six months have seen an increase in uncertainty relating to certain vendors.
Despite the push for technology independence, China is still some years behind and – like all other countries – depends on US technology for both design and fabrication. Four companies – either US-owned or with strong US ties – control 90% of the world’s chip design tool market: Cadence Design Systems, Synopsys, Ansys and Mentor Graphics. Similarly, companies such as TSMC rely on US-made chip making and testing equipment for fabrication.
Possible implications include the following:
Swap-out. This implies an accelerated process already underway in markets such as Japan, Australia, the UK and France. However, we note that in many of these markets the time horizon is long – 2028 in the UK and France. This would need to be dramatically accelerated, resulting in a spike in cost and a likely delay in commercial launch due to the disruption.
Mobile ‘splinternet’. One potential outcome is a separation in standards, with a Chinese standard evolving in line with the progress of its technology independence. Operators and governments would essentially end up choosing a US-backed or China-backed communications protocol.
We believe the impact on telecoms equipment should be relatively contained as suppliers have likely built up stockpiles of chips. Suppliers that have faced bans may also have some leverage in the need to support the equipment currently deployed in telecoms networks globally.
Australia demonstrates the possible impact Australia provides an example of the impact of replacing a radio access network, after barring a supplier from participating in 5G network rollouts.
Wireless providers that focused on suppliers that were not subject to government bans have benefited in terms of time to market versus their peers who have had to slow down their 5G plans, delaying their time to market. This illustrates the potential risk to markets and operators.
Open RAN – implications and developments
Single supplier risk has increased due to increased trade tensions. This builds on operator dissatisfaction with the extent of ‘lock-in’ to single vendor solutions.
While we highlight the benefits of a single-vendor in areas such as Dynamic Spectrum Sharing, we also believe that recent trade tensions will increase the opportunity for new equipment suppliers.
Key benefits of Open RAN include:
Selection of vendors based on technical merits. Open RAN is a set of standards that govern interoperability, enabling equipment from different vendors to work together using open interfaces.
32 Free to View ● Equities - Diversified Telecoms October 2020
Cost control. Open RAN reduces the exposure to any one single entity, and allows for a competitive bidding process between different entities. Operators such as Rakuten are mostly using common off-the-shelf (COTS) hardware – the real value increasingly lies in the software.
Greater network customisation. The disaggregation of the baseband (into Distributed Unit and Centralised Unit functions) gives operators the choice on how they configure their network. This should lead to greater revenue opportunities.
Key operator developments relating to Open RAN since our September 2019 report are:
Rakuten commercial launch in April 2020. While delayed, Rakuten launched commercial services in April 2020, and announced in August that it had over 1m subscribers on its network, supporting usage of c15GB per month. It has set up its Rakuten Communications Platform to license its services overseas, and has established a US presence.
Dish has committed to an Open RAN network, and has signed up vendors including Mavenir and AltioStar. Mavenir has also worked with Rakuten, and has teamed up with AltioStar to create a portfolio of radios with Open RAN interfaces for the US market.
Vodafone has issued a request for quotations for parts of its European network. Vodafone has been one of the leading operators supporting Open RAN, in particular with Telefonica at the Telecom Infra Project (TIP). In late 2019 it issued a request for quotations for Open RAN technology across its European footprint, encompassing c100k sites.
Risk of fragmentation in Cloud RAN deployments and standards Open RAN’s new-found popularity and momentum carries with it certain risks. Some entities believe that greater vendor involvement is resulting in some dilution of the goal of open standards. Media reports (SDX Central, 25 August 2020) reported that the Open Networking Foundation (ONF) has assembled a new working group to focus on the Radio Access Network, with ONF marketing head Timon Sloane noting that larger vendors that are now part of the ORAN Alliance are starting to ‘advocate for standards that are beneficial to their businesses’.
We note that both Ericsson and Nokia now have products that are billed as Open RAN, but it’s not clear that either these products or the contributions to the ORAN Alliance are based on open source code. Operators, not vendors, are now founding members of the ONF’s Software Defined RAN (SD-RAN) project, alongside Facebook, Google, Intel, Radisys and Sercomm.
The ONF points to lacks in the O-RAN specifications, such as the need for a microservices- based RAN Intelligent Controller (RIC) and the development of applications and interfaces for open source xApps to control the RAN. The initial focus of the ONF project is an open-source Near Real-Time RIC that is compatible with the ORAN Alliance architecture. This would be used to control several multi-vendor elements, instead of current configurations that still end up favouring a mostly integrated solution from one vendor. The ONF claims established vendors would prefer to keep the Self-Organising Network (SON) and Resource Radio Management (RRM) integrated with their RAN solutions. The ORAN Alliance now has more than 150 vendor members, and this leads to inevitable differences in preferences and motivation. The ONF by contrast is mostly operator-led, and has its own team of developers.
In May 2020, the Open RAN Policy Coalition was set up to help reduce confusion among policy makers so that shared goals can be advanced by politicians and regulators. Thirty-one members joined at the outset, including both operators and vendors such as Nokia.
Open RAN operator developments Reliance Jio – 5G platform ambition. Radisys is one of the vendors involved in the ONF project. This vendor is owned by Reliance Jio (Not Listed), which has a stated aim of
33 Free to View ● Equities - Diversified Telecoms October 2020
developing 5G capability in-house and then offering that as a platform, like Rakuten Mobile with its Rakuten Communications Platform.
Nokia and AT&T – RAN Intelligent Controller. Nokia and AT&T trialled a jointly developed RIC in June 2020 in New York City. The software will be made available to the ORAN community. The RIC allows external applications (xApps) to control parts of the 4G and 5G network much faster than previous methods. xApps could use machine learning to more efficiently control and allocate traffic.
Samsung Electronics. In July 2020 Samsung announced the addition of a virtualised Distributed Unit (DU) to go with the virtualised Central Unit (CU) announced in April 2019. Samsung notes improved Total Cost of Ownership (TCO) for Open RAN solutions, but not in every scenario. Total processing requirements based on spectrum, bandwidth requirements and other factors such as backhaul need to be taken into account.
Fujitsu. Fujitsu has parlayed its work with Rakuten into a contract with Dish, which has committed to launching an SA-based Open RAN compliant service in one US market by end 2020. Fujitsu will supply low and mid band ORAN compliant radios.
Of the existing 5G vendors, Samsung is the most committed to supporting Open RAN in our view. Nokia and Ericsson have started to join more Open RAN forums, but both are likely to defend their existing integrated solutions.
Open RAN likely to be less efficient for mmWave and even urban 4G Even Open RAN proponents acknowledge shortcomings in certain areas.
Samsung has pointed out that the high processing requirements for mmWave mean that virtualised RAN is a better fit for low and mid-band 5G deployments. Rakuten’s 5G launch later in 2020 will be an interesting test case in this regard.
Open RAN outlook The risk here is that new groups outside the main alliance slow down the momentum of the project – adding to the inherent difficulty of maintaining momentum in an open source project. This slower timeline is likely one of the reasons Rakuten chose to develop its own Cloud RAN platform – it is loosely aligned with O-RAN, but not fully involved.
Delays in developing agreed open standards remain a risk for Open RAN solutions in general – global leaders in 5G have so far been those using established vendors – particularly Ericsson and Huawei (Telstra, T-Mobile US). Those managing multiple vendors or with Open RAN requirements (Rakuten, NTT DoCoMo, AT&T, Dish) are behind.
34 Free to View ● Equities - Diversified Telecoms October 2020
Disclosure appendix
Analyst Certification The following analyst(s), economist(s), or strategist(s) who is(are) primarily responsible for this report, including any analyst(s) whose name(s) appear(s) as author of an individual section or sections of the report and any analyst(s) named as the covering analyst(s) of a subsidiary company in a sum-of-the-parts valuation certifies(y) that the opinion(s) on the subject security(ies) or issuer(s), any views or forecasts expressed in the section(s) of which such individual(s) is(are) named as author(s), and any other views or forecasts expressed herein, including any views expressed on the back page of the research report, accurately reflect their personal view(s) and that no part of their compensation was, is or will be directly or indirectly related to the specific recommendation(s) or views contained in this research report: Neale Anderson, Piyush Choudhary, CFA, Adam Fox-Rumley, CFA, Nicolas Cote-Colisson, Christian Fangmann, Luigi Minerva, Sunil Rajgopal and Christopher Recouso
Important disclosures Equities: Stock ratings and basis for financial analysis HSBC and its affiliates, including the issuer of this report (“HSBC”) believes an investor's decision to buy or sell a stock should depend on individual circumstances such as the investor's existing holdings, risk tolerance and other considerations and that investors utilise various disciplines and investment horizons when making investment decisions. Ratings should not be used or relied on in isolation as investment advice. Different securities firms use a variety of ratings terms as well as different rating systems to describe their recommendations and therefore investors should carefully read the definitions of the ratings used in each research report. Further, investors should carefully read the entire research report and not infer its contents from the rating because research reports contain more complete information concerning the analysts' views and the basis for the rating.
From 23rd March 2015 HSBC has assigned ratings on the following basis: The target price is based on the analyst’s assessment of the stock’s actual current value, although we expect it to take six to 12 months for the market price to reflect this. When the target price is more than 20% above the current share price, the stock will be classified as a Buy; when it is between 5% and 20% above the current share price, the stock may be classified as a Buy or a Hold; when it is between 5% below and 5% above the current share price, the stock will be classified as a Hold; when it is between 5% and 20% below the current share price, the stock may be classified as a Hold or a Reduce; and when it is more than 20% below the current share price, the stock will be classified as a Reduce.
Our ratings are re-calibrated against these bands at the time of any ‘material change' (initiation or resumption of coverage, change in target price or estimates).
Upside/Downside is the percentage difference between the target price and the share price.
Prior to this date, HSBC’s rating structure was applied on the following basis: For each stock we set a required rate of return calculated from the cost of equity for that stock’s domestic or, as appropriate, regional market established by our strategy team. The target price for a stock represented the value the analyst expected the stock to reach over our performance horizon. The performance horizon was 12 months. For a stock to be classified as Overweight, the potential return, which equals the percentage difference between the current share price and the target price, including the forecast dividend yield when indicated, had to exceed the required return by at least 5 percentage points over the succeeding 12 months (or 10 percentage points for a stock classified as Volatile*). For a stock to be classified as Underweight, the stock was expected to underperform its required return by at least 5 percentage points over the succeeding 12 months (or 10 percentage points for a stock classified as Volatile*). Stocks between these bands were classified as Neutral.
*A stock was classified as volatile if its historical volatility had exceeded 40%, if the stock had been listed for less than 12 months (unless it was in an industry or sector where volatility is low) or if the analyst expected significant volatility. However, stocks which we did not consider volatile may in fact also have behaved in such a way. Historical volatility was defined as the past month's average of the daily 365-day moving average volatilities. In order to avoid misleadingly frequent changes in rating, however, volatility had to move 2.5 percentage points past the 40% benchmark in either direction for a stock's status to change.
35 Free to View ● Equities - Diversified Telecoms October 2020
Rating distribution for long-term investment opportunities As of 07 October 2020, the distribution of all independent ratings published by HSBC is as follows: Buy 55% ( 32% of these provided with Investment Banking Services ) Hold 36% ( 32% of these provided with Investment Banking Services ) Sell 9% ( 25% of these provided with Investment Banking Services )
For the purposes of the distribution above the following mapping structure is used during the transition from the previous to current rating models: under our previous model, Overweight = Buy, Neutral = Hold and Underweight = Sell; under our current model Buy = Buy, Hold = Hold and Reduce = Sell. For rating definitions under both models, please see “Stock ratings and basis for financial analysis” above. For the distribution of non-independent ratings published by HSBC, please see the disclosure page available at http://www.hsbcnet.com/gbm/financial-regulation/investment-recommendations-disclosures. To view a list of all the independent fundamental ratings disseminated by HSBC during the preceding 12-month period, please use the following links to access the disclosure page: Clients of Global Research and Global Banking and Markets: www.research.hsbc.com/A/Disclosures Clients of HSBC Private Banking: www.research.privatebank.hsbc.com/Disclosures HSBC and its affiliates will from time to time sell to and buy from customers the securities/instruments, both equity and debt (including derivatives) of companies covered in HSBC Research on a principal or agency basis or act as a market maker or liquidity provider in the securities/instruments mentioned in this report. Analysts, economists, and strategists are paid in part by reference to the profitability of HSBC which includes investment banking, sales & trading, and principal trading revenue. Whether, or in what time frame, an update of this analysis will be published is not determined in advance. Non-U.S. analysts may not be associated persons of HSBC Securities (USA) Inc, and therefore may not be subject to FINRA Rule 2241 or FINRA Rule 2242 restrictions on communications with the subject company, public appearances and trading securities held by the analysts. Economic sanctions imposed by the EU and OFAC prohibit transacting or dealing in new debt or equity of Russian SSI entities. This report does not constitute advice in relation to any securities issued by Russian SSI entities on or after July 16 2014 and as such, this report should not be construed as an inducement to transact in any sanctioned securities. For disclosures in respect of any company mentioned in this report, please see the most recently published report on that company available at www.hsbcnet.com/research. HSBC Private Banking clients should contact their Relationship Manager for queries regarding other research reports. In order to find out more about the proprietary models used to produce this report, please contact the authoring analyst.
Additional disclosures 1 This report is dated as at 07 October 2020. 2 All market data included in this report are dated as at close 05 October 2020, unless a different date and/or a specific time of day is indicated in the report. 3 HSBC has procedures in place to identify and manage any potential conflicts of interest that arise in connection with its Research business. HSBC's analysts and its other staff who are involved in the preparation and dissemination of Research operate and have a management reporting line independent of HSBC's Investment Banking business. Information Barrier procedures are in place between the Investment Banking, Principal Trading, and Research businesses to ensure that any confidential and/or price sensitive information is handled in an appropriate manner. 4 You are not permitted to use, for reference, any data in this document for the purpose of (i) determining the interest payable, or other sums due, under loan agreements or under other financial contracts or instruments, (ii) determining the price at which a financial instrument may be bought or sold or traded or redeemed, or the value of a financial instrument, and/or (iii) measuring the performance of a financial instrument or of an investment fund.
36 Free to View ● Equities - Diversified Telecoms October 2020
Disclaimer
Legal entities as at 1 September 2020 Issuer of report ‘UAE' HSBC Bank Middle East Limited, DIFC; HSBC Bank Middle East Limited, Dubai; ‘HK’ The Hongkong and Shanghai The Hongkong and Shanghai Banking Corporation Banking Corporation Limited, Hong Kong; ‘TW’ HSBC Securities (Taiwan) Corporation Limited; ‘CA' HSBC Securities Limited (Canada) Inc.; HSBC France, S.A., Madrid, Milan, Stockholm; ‘DE’ HSBC Trinkaus & Burkhardt AG, Düsseldorf; 000 HSBC Level 19, 1 Queen’s Road Central Bank (RR), Moscow; ‘IN’ HSBC Securities and Capital Markets (India) Private Limited, Mumbai; ‘JP’ HSBC Securities (Japan) Hong Kong SAR Limited, Tokyo; ‘EG’ HSBC Securities Egypt SAE, Cairo; ‘CN’ HSBC Investment Bank Asia Limited, Beijing Representative Telephone: +852 2843 9111 Office; The Hongkong and Shanghai Banking Corporation Limited, Singapore Branch; The Hongkong and Shanghai Banking Fax: +852 2596 0200 Corporation Limited, Seoul Securities Branch; The Hongkong and Shanghai Banking Corporation Limited, Seoul Branch; Website: www.research.hsbc.com HSBC Securities (South Africa) (Pty) Ltd, Johannesburg; HSBC Bank plc, London, Tel Aviv; ‘US’ HSBC Securities (USA) Inc., New York; HSBC Yatirim Menkul Degerler AS, Istanbul; HSBC México, SA, Institución de Banca Múltiple, Grupo Financiero HSBC; HSBC Bank Australia Limited; HSBC Bank Argentina SA; HSBC Saudi Arabia Limited; The Hongkong and Shanghai Banking Corporation Limited, New Zealand Branch incorporated in Hong Kong SAR; The Hongkong and Shanghai Banking Corporation Limited, Bangkok Branch; PT Bank HSBC Indonesia; HSBC Qianhai Securities Limited; Banco HSBC S.A. This document has been issued by The Hongkong and Shanghai Banking Corporation Limited (“HSBC”) in the conduct of its Hong Kong regulated business for the information of its institutional and professional investor (as defined by Securities and Future Ordinance (Chapter 571)) customers; it is not intended for and should not be distributed to retail customers in Hong Kong. The Hongkong and Shanghai Banking Corporation Limited is regulated by the Hong Kong Monetary Authority. All enquires by recipients in Hong Kong must be directed to your HSBC contact in Hong Kong. If it is received by a customer of an affiliate of HSBC, its provision to the recipient is subject to the terms of business in place between the recipient and such affiliate. This document is not and should not be construed as an offer to sell or the solicitation of an offer to purchase or subscribe for any investment. HSBC has based this document on information obtained from sources it believes to be reliable but which it has not independently verified; HSBC makes no guarantee, representation or warranty and accepts no responsibility or liability as to its accuracy or completeness. Expressions of opinion are those of the Research Division of HSBC only and are subject to change without notice. From time to time research analysts conduct site visits of covered issuers. HSBC policies prohibit research analysts from accepting payment or reimbursement for travel expenses from the issuer for such visits. HSBC and its affiliates and/or their officers, directors and employees may have positions in any securities mentioned in this document (or in any related investment) and may from time to time add to or dispose of any such securities (or investment). HSBC and its affiliates may act as market maker or have assumed an underwriting commitment in the securities of companies discussed in this document (or in related investments), may sell them to or buy them from customers on a principal basis and may also perform or seek to perform investment banking or underwriting services for or relating to those companies. HSBC Securities (USA) Inc. accepts responsibility for the content of this research report prepared by its non-US foreign affiliate. The information contained herein is under no circumstances to be construed as investment advice and is not tailored to the needs of the recipient. All U.S. persons receiving and/or accessing this report and wishing to effect transactions in any security discussed herein should do so with HSBC Securities (USA) Inc. in the United States and not with its non-US foreign affiliate, the issuer of this report. In the UK, this publication is distributed by HSBC Bank plc for the information of its Clients (as defined in the Rules of FCA) and those of its affiliates only. It is not intended for Retail Clients in the UK. Nothing herein excludes or restricts any duty or liability to a customer which HSBC Bank plc has under the Financial Services and Markets Act 2000 or under the Rules of FCA and PRA. A recipient who chooses to deal with any person who is not a representative of HSBC Bank plc in the UK will not enjoy the protections afforded by the UK regulatory regime. HSBC Bank plc is regulated by the Financial Conduct Authority and the Prudential Regulation Authority. In Singapore, this publication is distributed by The Hongkong and Shanghai Banking Corporation Limited, Singapore Branch for the general information of institutional investors or other persons specified in Sections 274 and 304 of the Securities and Futures Act (Chapter 289) (“SFA”) and accredited investors and other persons in accordance with the conditions specified in Sections 275 and 305 of the SFA. Only Economics or Currencies reports are intended for distribution to a person who is not an Accredited Investor, Expert Investor or Institutional Investor as defined in SFA. The Hongkong and Shanghai Banking Corporation Limited, Singapore Branch accepts legal responsibility for the contents of reports pursuant to Regulation 32C(1)(d) of the Financial Advisers Regulations. This publication is not a prospectus as defined in the SFA. This publication is not a prospectus as defined in the SFA. It may not be further distributed in whole or in part for any purpose. The Hongkong and Shanghai Banking Corporation Limited Singapore Branch is regulated by the Monetary Authority of Singapore. Recipients in Singapore should contact a “Hongkong and Shanghai Banking Corporation Limited, Singapore Branch” representative in respect of any matters arising from, or in connection with this report. Please refer to The Hongkong and Shanghai Banking Corporation Limited Singapore Branch’s website at www.business.hsbc.com.sg for contact details. In Australia, this publication has been distributed by The Hongkong and Shanghai Banking Corporation Limited (ABN 65 117 925 970, AFSL 301737) for the general information of its “wholesale” customers (as defined in the Corporations Act 2001). Where distributed to retail customers, this research is distributed by HSBC Bank Australia Limited (ABN 48 006 434 162, AFSL No. 232595). These respective entities make no representations that the products or services mentioned in this document are available to persons in Australia or are necessarily suitable for any particular person or appropriate in accordance with local law. No consideration has been given to the particular investment objectives, financial situation or particular needs of any recipient. This publication is distributed in New Zealand by The Hongkong and Shanghai Banking Corporation Limited, New Zealand Branch incorporated in Hong Kong SAR. In Japan, this publication has been distributed by HSBC Securities (Japan) Limited. It may not be further distributed in whole or in part for any purpose. In Korea, this publication is distributed by The Hongkong and Shanghai Banking Corporation Limited, Seoul Securities Branch ("HBAP SLS”) for the general information of professional investors specified in Article 9 of the Financial Investment Services and Capital Markets Act (“FSCMA”). This publication is not a prospectus as defined in the FSCMA. It may not be further distributed in whole or in part for any purpose. HBAP SLS is regulated by the Financial Services Commission and the Financial Supervisory Service of Korea. In Canada, this document has been distributed by HSBC Securities (Canada) Inc. (member IIROC), and/or its affiliates. The information contained herein is under no circumstances to be construed as investment advice in any province or territory of Canada and is not tailored to the needs of the recipient. No securities commission or similar regulatory authority in Canada has reviewed or in any way passed judgment upon these materials, the information contained herein or the merits of the securities described herein, and any representation to the contrary is an offense. In Brazil, this document has been distributed by Banco HSBC S.A. ("HSBC Brazil”), and/or its affiliates. As required by Instruction No. 598/18 of the Securities and Exchange Commission of Brazil (Comissão de Valores Mobiliários), potential conflicts of interest concerning (i) HSBC Brazil and/or its affiliates; and (ii) the analyst(s) responsible for authoring this report are stated on the chart above labelled “HSBC & Analyst Disclosures”. If you are an HSBC Private Banking (“PB”) customer with approval for receipt of relevant research publications by an applicable HSBC legal entity, you are eligible to receive this publication. To be eligible to receive such publications, you must have agreed to the applicable HSBC entity’s terms and conditions (“KRC Terms”) for access to the KRC, and the terms and conditions of any other internet banking service offered by that HSBC entity through which you will access research publications using the KRC. Distribution of this publication is the sole responsibility of the HSBC entity with whom you have agreed the KRC Terms. If you do not meet the aforementioned eligibility requirements please disregard this publication and, if you are a customer of PB, please notify your Relationship Manager. Receipt of research publications is strictly subject to the KRC Terms, which can be found at https://research.privatebank.hsbc.com/ – we draw your attention also to the provisions contained in the Important Notes section therein. © Copyright 2020, The Hongkong and Shanghai Banking Corporation Limited, ALL RIGHTS RESERVED. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, on any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior written permission of The Hongkong and Shanghai Banking Corporation Limited. MCI (P) 077/12/2019, MCI (P) 016/02/2020
[1156331]
37