Digital ConneCtivity anD low earth orbit Satellite ConStellationS opportunitieS for aSia anD the paCifiC ADB SuStAinABle Development working

Satellite communication plays an important role in the global connectivity ecosystem. It connects rural and remote populations, provides backhaul connectivity to mobile cellular networks, and enables rapid communications for emergency and disaster responses. Low Earth orbit constellations may prove to be transformational to the connectivity landscape based on their global coverage and their suitability for areas not served by fiber optic cable networks. The Asian Development Bank’s developing member countries are well placed to benefit from this expansion of internet connectivity. It will be particularly valuable for small island developing states and landlocked developing countries with limited international bandwidth internet.


eXeCUtIVe sUMMARY
Satellite communication plays a necessary role in the global connectivity ecosystem, connecting rural and remote populations, providing backhaul connectivity to mobile cellular networks, and rapidly establishing communication in emergency and disaster response scenarios. This Asian Development Bank (ADB) Sustainable Development Working Paper, the first in a series reviewing emerging innovations in connectivity technologies, focuses on low Earth orbit (LEO) satellites, which have been in deployment for decades and are again a subject of intensive investment as new large constellations are in early stages of deployment. These new LEO constellations, such as those being deployed by Starlink by SpaceX, Project Kuiper by Amazon, OneWeb, Lightspeed by Telesat, among others, may prove to be transformational to the connectivity landscape based on their global coverage and their suitability for areas not served by fiber optic cable networks.
ADB's developing member countries are well placed to leverage and benefit from this expansion of internet connectivity, particularly for underserved geographies and countries with limited international internet bandwidth, such as landlocked developing countries and small island developing states. With their global reach and coverage, LEO constellations are expected to dramatically expand the availability of high-speed broadband internet access with levels of service that rival fiber optic cables in terms of speed and latency, and at significantly reduced price levels compared to traditional geostationary satellites.
A proactive engagement with LEO solutions is likely to yield benefits as the relevant business models are still evolving. Well-informed early action by regulators and investors can ensure that developing member countries prepare for opportunities presented by the anticipated expansion of connectivity bandwidth.
This Emerging Connectivity Innovations Case Study on SpaceX Starlink and low Earth orbit (LEO) satellite constellations is intended to provide readers, particularly in developing countries in Asia and the Pacific, with a background understanding of the role of satellite communications in global internet connectivity and an exploration of the potential impact of the next generation of LEO constellation systems.
While the adoption of internet connectivity across the world has generally increased incrementally, some innovations have been transformational, dramatically expanding the geographic reach of connectivity and bandwidth capacity. For example, the introduction of basic mobile phones in the late 1990s and early 2000s led to rapid adoption of mobile telephony across low-and middle-income countries (a phenomenon known as the "mobile miracle"). Similarly, public and private investment in undersea fiber optic cables circling sub-Saharan Africa in the 2000s significantly reduced the cost of bandwidth in many countries in the region.
Satellites have used low Earth orbits since the beginning of space exploration; however, private investment in LEO constellations, consisting of hundreds or thousands of satellites, has been limited because significant up-front capital expenditure is required. While it remains to be seen how the next generation of LEO satellite constellations will evolve, LEOs are forecasted to significantly increase the available internet bandwidth in remote and rural geographies not currently served by fiber optic cables. This increased bandwidth could be leveraged to increase economic and social development opportunities for individuals, organizations, businesses, and government facilities (including public schools) located in these areas, provided that the private sector satellite companies investing in LEO constellations see market opportunities to extend service to these areas. This case study is intended to introduce to Asian Development Bank developing member countries how to start preparing for the expansion of LEO satellite communication services.

II. BACKGRoUnD: sAteLLIte ConneCtIVItY As A MeAns FoR BRoADBAnD InteRnet
Internet connectivity has become a necessary component of every country's critical infrastructure given the reliance of all aspects of economic activity, governance, and social development on internet communications.
The coronavirus disease  pandemic dramatically increased the importance of internet communications infrastructure. Trade, employment, learning, leisure, and communications quickly shifted into the digital sphere and countries with robust internet infrastructure and high adoption rates of internet-enabled devices were better able to adjust and adapt to the shift to digital activity. The United Nations estimates that 1.6 billion learners were affected by school closures in 2020, affecting 94% of the world's student population and up to 99% in low and lower middle-income countries. 1 1 United Nations. 2020. Policy Brief: Education during COVID- 10 and beyond. Access to distance learning opportunities varies greatly by country and income groups, with estimates of less than half of students in low-income countries able to access distance learning. 2 Internet access and adoption in the developing member countries (DMCs) of the Asian Development Bank (ADB) continues to grow, particularly as a result of public and private investment in telecommunications infrastructure, increased competition, and allocation of shared resources, such as spectrum auctions and assignment. Despite these efforts, large access gaps remain in Asia, where the most remote, difficult to reach, or sparsely populated districts remain disconnected, leaving more than half of the population without access to the internet. This lack of digital infrastructure represents a missed opportunity to accelerate economic and social development.
Despite the rapid expansion of internet connectivity infrastructure across the world, significant gaps in internet adoption and infrastructure access remain. This highlights the importance of satellite communications that can bridge gaps, swiftly expand network coverage, and enhance existing infrastructure.
The latest estimates from the International Telecommunication Union (ITU) show that 3.7 billion people are still not participating online (49% of the global population), and 63% of rural households are without internet access ( Figure 1). 3 Also, 1.5 billion people reside in areas without high-speed mobile data coverage (fourth generation long-term evolution or 4G LTE), while 607 million people reside in areas with no mobile data coverage at all (at least 4G or third generation [3G] coverage). Furthermore, 313 million people reside in areas with only basic voice and short messaging service (SMS) coverage (second generation [2G]), and 220 million people reside in areas with no cellular coverage. The ITU estimates that nearly $428 billion is required to achieve universal access to broadband globally, $251 billion of which is required for Asia, with approximately 75% coming from the private sector and the remainder with support from the public sector. 4 The majority of the world's population, over 5 billion people, live more than 10 kilometers (km) away from any fiber optic cable infrastructure (3.6 billion reside more than 25 km away Satellite connectivity is predominantly used for backhaul connectivity for remote cellular base stations and as a last-mile connection for individual subscribers and enterprises.
Figure 2 provides an overview of the internet infrastructure network components, from international connectivity to the last mile. Because of the higher relative cost of bandwidth transmitted via satellite versus terrestrial technologies, satellite is currently primarily used in situations where fiber optic cables and other high-capacity technologies are not financially viable due to low population densities and large distances between high-capacity networks and last-mile networks. 6 However, in a few cases, satellite connectivity is relied upon for international internet gateway traffic or as part of a country's core network.
For landlocked developing countries that are dependent on terrestrial fiber connectivity, in some cases, satellite connectivity serves as a substitute to complex bilateral and multilateral negotiations to extend costly fiber connectivity to their country. 6 In the case of Pacific island countries, satellite connectivity is increasingly playing a complementary role to undersea fiber optic cable deployment, including as a redundancy option. UNESCAP. 2019. Satellite Communications in Pacific Island Countries. Satellite connectivity provides global coverage, with trade-offs in capacity and affordability.
Satellite communications coverage is already global, composed of geostationary orbit (GEO) satellites (also known as geostationary Earth orbit), medium Earth orbit (MEO), and low Earth orbit (LEO) constellations. 7 (See Chapter 1, Figure 5 for differences between LEO, MEO, and GEO.) There are currently at least 775 active satellites in orbit that serve primary communications functions (excluding the new LEO constellations). 8 However, the total satellite sellable capacity in 2020 of approximately 3 terabits per second (Tbps) is dwarfed by the roughly 2,000 Tbps of utilized terrestrial fiber capacity, 9 7 Low Earth orbit (LEO) and medium Earth orbit (MEO) satellites are often grouped together as non-geosynchronous orbit (NGSO). 8 ITU/UNESCO. 2020; and SatBeams. 2020. List of Satellites at Geostationary Orbit. 9 A. Mauldin. 2019. Will New Satellites End the Dominance of Submarine Cables. Telegeography Blog. 1 July. Digital Connectivity and Low Earth Orbit Satellite Constellations which still is only a fraction of the total potential capacity on existing cables (international internet capacity). As of early 2020, there were at least 406 submarine cables in service and 99% of total international internet data traffic is transmitted via fiber optic cables. 10 Wholesale prices also differ dramatically with international internet transit (IP transit) pricing as low as $1-$3 per megabit per second (Mbps) per month on major cross-country routes against wholesale prices for dedicated satellite capacity approaching $200-$400 Mbps per month. 11 Therefore, satellite connectivity is only costcompetitive for remote and dispersed populations where fiber deployments are challenging. The new generation of LEO and high-throughput GEO satellites are expected to lower the cost structure and make satellite connectivity more competitive (Figure 3).  Particularly in situations where a high degree of data throughput is required per site, such as satellite backhaul for broadband cellular networks, the data volumes as well as the distance to the nearest backbone node play a significant role in cost comparisons between satellite connectivity versus terrestrial 10 Telegeography Blog. 2020. Submarine Cable 101; The preference for submarine cable traffic is also due to latency requirements, such as for financial transactions. This may be expected to change as LEO demonstrates fiber-like latency (or better). network deployments (microwave backhaul, in particular). Figure 4 illustrates how higher data bandwidth requirements are more cost-effectively supplied by terrestrial ground networks; however, a crossover point occurs where satellite capacity may end up being more cost-competitive, depending on different price points of satellite bandwidth and total traffic demand per month. 12 Satellite connectivity is also wellsuited to deploy in emergency situations, such as in response to natural disasters or other external shocks, that require expeditious deployment of network connectivity where terrestrial infrastructure is either nonexistent or destroyed. For many rural and remote communities, satellites are the only connectivity option.
For geographies without direct access to fiber optic cable infrastructure or at great distances from highcapacity bandwidth capacity, satellite connectivity is the only option available. Even where terrestrial network infrastructure that could be used for backhaul connectivity is available, satellite deployments may still be preferred because satellite terminals require only electrical power and a clear line of sight to the sky. However, an expansion of terrestrial infrastructure usually requires extensive civil works (underground fiber ducts, pole attachments, or tower construction for cellular base stations), which comes with challenges such as securing the rights-of-way, permits, and having to pay the related fees.
Satellite broadband is poised to become an even more important technology for addressing the growing digital divide.
As information and communication technologies play an increasingly important role in commerce, government services, health care, education, and other sectors, satellite connectivity allows communities to get connected swiftly, bypassing the infrastructure deployment challenges that come with terrestrial infrastructure deployments. The role of satellite connectivity in emergency telecommunications has also been vital where the communications satellites are heavily relied upon in disaster recovery efforts. 13 Satellite technology may also be complementary with traditional wired and mobile broadband, which are better suited for densely populated areas. Satellite service could become a default solution for remote areas, allowing terrestrial services to focus on improving access in their current coverage areas. Satellite connectivity is already being used for network redundancy at national levels for international internet capacity, as well as for backup in core and backhaul networks. 14 The recent $50 million loan to Kacific by ADB for the deployment of a broadband satellite, which covers large parts of Southeast Asia and the Pacific, demonstrates the relevance of satellite connectivity for unserved and underserved regions. 15 By deploying new satellite technology (in the Ka-band 16 ), Kacific's service offering is commercially viable despite the existing presence of other major competitors in Asia and the Pacific, including global entities such as Intelsat, SES, and Eutelsat, as well as more regional players such as AsiaSat, Thaicom, MEASAT, and SKY Perfect JSAT.

III. InnoVAtIon In LoW eARth oRBIt sAteLLIte ConsteLLAtIons
Low Earth orbit satellites have been a feature of space satellites since the beginning of space exploration.
Since the start of space exploration, low Earth orbits between 160-2,000 km above the Earth's surface have been used for satellite placement. For example, Earth observation, spy, and remote sensing satellites have been using low Earth orbits to obtain high-resolution measurements of the Earth's surface.  17 December. 16 Use of the Ka-band of radio spectrum frequencies, in the 26.5-40 gigahertz (GHz) range is an advancement over use of the Ku-band (12)(13)(14)(15)(16)(17)(18) as the higher frequency allows for greater throughput and higher data transfer (and corresponding lower cost per bit transferred). However, interference from water droplets (rain fade, humidity) is an issue in both spectrum bands.
The International Space Station (ISS) is in a LEO of 400-420 km and is in constant orbital decay, requiring re-boosting a few times a year. The Hubble Space Telescope is also in a LEO position at an altitude of 540 km. Communications satellites that provide satellite phone service use LEOs, for example, Iridium satellites orbit at about 780 km to limit latency (the round trip time for data to travel between communication points) delays in voice transmissions.
Before the 1990s, low Earth orbits were used only by government satellites for military and scientific missions, but have since become the focus of commercial deployments.
Large-scale commercial deployments first began in the 1990s, when several companies tried to provide global connectivity, for example, Globalstar, Iridium, Odyssey, and Teledesic. Because individual LEO satellites cover a limited geographic area, constellations need hundreds or thousands of satellites and ground stations to achieve robust global coverage depending on orbital altitude above the Earth's surface.
Unfortunately, previous large-scale plans were cancelled or reduced due to high costs and limited demand. Industry analysts have since been skeptical of LEO constellation viability (exemplified by the recent challenges of OneWeb and LeoSat). 17 Three communication satellite constellations have been operating in LEO since the early 2000s: Iridium, Globalstar, and Orbcomm (Table 1). 18 Low Earth orbit constellations provide fundamentally different value propositions compared to geostationary orbit.
GEO satellites are positioned at an altitude of 35,786 km and, as a result, each satellite has a very wide coverage area. This allows them to focus their bandwidth capacity on their coverage area and reduces the requirement for achieving global coverage to as few as three satellites. However, because of their distance from the Earth's surface, their minimum latency thresholds are high (roughly at least 0.477 seconds for round trip latency).
The latest generation of GEO satellites, known as high-throughput satellites (HTSs), have a significantly increased capacity (at least 10 times the throughput) than previous generations of GEO satellites, while the high latency (which is a function of distance and the speed of light) remains the same. 19 In comparison, LEO constellations require a network of satellites to provide internet service because each LEO satellite is traversing the Earth's surface, orbiting the planet every 88-127 minutes (depending on their altitude, between 160-2,000 km). Their closer distance to the Earth's surface enables them to provide high-speed, low-latency internet. MEO satellites are positioned between LEO and GEO orbits, circling the planet by 2,000-35,786 km. These satellites handle high-speed, low-latency data traffic (particularly cellular backhaul  The sheer capacity of bandwidth that low Earth orbit constellations will be able to offer will dwarf existing high-throughput geostationary orbit satellites. Individual GEO satellites provide bandwidth capacity of 1-10 gigabits per second (Gbps), while the first-generation HTS range is 10-50 Gbps, and third-generation (HTS Class III) provides capacity of 150-350 Gbps (footnote 18). As an example, the Kacific 1 satellite includes 56 individual beams (Kaband), each having a capacity up to 1.25 Gbps (one of the highest signal powers achieved in the region). 21 With the next generation LEO constellations consisting of hundreds, if not thousands, of satellites each able to transmit tens (10s) of Gbps, the total capacity of new constellations is forecast to be in single digit to 10s of Tbps, dramatically surpassing existing sellable capacity of GEO ( Figure 6). One report forecasts satellite broadband capacity to increase from an estimated 2 Tbps at the end of 2020, to 20 Tbps by end of 2021, and 60 Tbps by the end of the decade. 22 In Asia, bandwidth demand supplied by non-GEO satellites is forecast to rise dramatically ( Figure 7).
Low Earth orbit versus geostationary orbit: differences in subscriber data pricing may be substantial once low Earth orbit commercial services begin.
While prices for wholesale bandwidth are generally not public information, the advent of HTS has led to a divergence in cost for satellite services, with rates for broadband connectivity falling faster than that for broadcast and distribution. 23    As LEO and MEO satellites become more ubiquitous, these technologies are expected to provide a cost advantage compared to GEO satellites. Since LEO constellations are expected to provide identical coverage for the service area under the constellation's footprint, they could potentially offer a uniform pricing model anywhere in the world (footnote 24). Combined with the high-capacity, shorter-term (and, therefore, more competitive) contracts, LEOs could rapidly decrease satellite bandwidth costs globally. There will also be significant opportunity for regional pricing due to the fact that the up-front investment in the constellation is the primary cost, and individual satellites are only able to serve the area under their current location.
One of the new LEO constellations in deployment is SpaceX's Starlink service, which is the most advanced LEO constellation in deployment in terms of number of satellites and current stage of internet service offered. Starlink began a public beta trial program in October 2020 for subscribers in the northern United States (US) and Canada between the latitudes of 45º and 52º focusing on rural locations (called the "Better Than Nothing Beta"). 25 The service pricing offered during the public beta is $99/month for speeds between 50-150 Mbps, latency between 20-40 milliseconds, plus a one-time equipment fee of $499. No data cap has yet been implemented. 26 Based on this data, a per Mbps price comparison would see Starlink at $0.50-$1.50/Mbps compared to a range of pricing from HTS GEO services shown in Table 2. It is important to note that this comparison should be revised when commercial pricing of LEO becomes available.
Even as prices have fallen over the years, satellite broadband remains expensive compared to connectivity from terrestrial technologies at the same speeds. This is particularly relevant for Asia and the Pacific, where providers have to contend with users who have less disposable income relative to users elsewhere in the world. 27  The retail satellite broadband market in parts of Asia and the Pacific is less mature than that of other regions, with retail subscriptions lagging in speed, price, and data allocation. This is illustrated in Table 2, which contains a summary of retail satellite broadband packages offered in various countries. 30 All of these options use GEO satellites, which have high latency and limited coverage areas. Comparing low Earth orbit constellations shows that Starlink is ahead of its competitors in terms of deployment and service position.
Four main companies take the spotlight in terms of next-generation LEO constellation deployment for broadband communications ( Canadian Telesat began in 1969 with decades of experience in satellite communications. Its LEO plans began in 2016 with an announcement to launch 120 satellites. In 2018, its first (and only) LEO satellite in orbit was launched as it began testing and in 2021, it announced details of its LEO constellation, named Lightspeed, consisting of 298 satellites. 39 Telesat is expected to start launching satellites in 2023 (on Amazon's Blue Origin launch vehicles) and the constellation is expected to cost $5 billion. 40 The newcomer of the group is Amazon with its Project Kuiper announced in 2019. In July 2020, it announced planned investment of more than $10 billion, supported in part by its Blue Origin company, focused on reusable launch rockets. 41 Project Kuiper has not yet launched any satellites, but has Federal Communications Commission (FCC) approval for 3,236 satellites with commercial service beginning once 578 satellites are in orbit. 42 Project Kuiper is expected to launch half of its satellites by 2026 and the remainder by 2029 (footnote 41). Table 3 shows a comparison of the different companies' deployments, constellations, and satellites. Other LEO satellite constellations have been announced, but have not reached similar levels of (planned) deployment and/or information about their plans is not publicly available.
Significant differences planned in constellation configurations.
The three constellations with current deployments (Starlink, OneWeb, and Telesat) have significant differences in configuration and satellites. Starlink satellites are configured to have lower orbit altitude than its competitors, its latest configuration at up to 550 km off the ground. 43 This will provide lower latencies for satellite-to-earth and earth-to-satellite communication. It originally planned to include optical inter-satellite links (OISL) which would reduce latency even further over long distances and provide communications that could theoretically even be faster than fiber optic cable (due to the vacuum in space). However, inter-satellite links have not yet been fully deployed, and only a few of the satellites launched to date include OISL technology. 44 In Telesat's constellation, each satellite will be an internet protocol node, with each satellite functioning as a router making decisions on where to send traffic. Telesat's more complex design helps to optimize network architecture and efficiency as its satellites orbit at a higher altitude and will be equipped with OISL.  Unlike Telesat's model, OneWeb will employ a traditional "bent-pipe" architecture. Signals coming up from the ground will be "repeated" to another beam going down, with no routing taking place, simplifying the networking component of satellite design and engineering, but without the potential benefit of overall network redundancies and efficiency in data packet routing. The differences in orbits and coverage are demonstrated in Figure 8, including differences in average and maximum data rates per satellite.   It is important to note these were estimates from over 2 years ago and what this research found was that the capacity and coverage of each constellation is highly dependent on assumptions of satellite fleet size, satellite efficiency, and the number of ground stations demonstrating that Starlink, in particular, would require significant resources for its ground infrastructure with hundreds of ground stations and thousands of gateway antennas. More recent updates from the LEO providers note that commercial service will begin for Starlink with 1,440 satellites, OneWeb with 648 satellites, and Telesat with 298 satellites. 46 Low Earth orbit user terminal pricing will be an area of intense competition on price and innovation.
Very small aperture terminals (VSATs) for HTS GEO service range around $1,000 each. 47 The current $499 price for the Starlink terminal (named "Dishy McFlatface") is a significant price reduction. However, it is not clear if this price level will be sustained during commercial service and even if so, the up-front cost may price some subscribers out of the market, unless some form of amortized financing is offered.
The level of subsidy that SpaceX may be absorbing per unit is not known, but industry reports suggest that phased array flat panel antennas cost above $1,000 per unit. 48 OneWeb's most affordable user terminal offering is priced at $1,200 per unit before freight, taxes, and other costs. However, Starlink should be able to achieve economies of scale with millions of terminals eventually required for its global customer base. 49 Even Amazon has been detailing its innovations in phased array antenna development, noting successes in designing and developing a small 12-inch antenna. 50 In-house satellite launch capacity and expertise may result in a competitive advantage for Starlink and, in the future, Project Kuiper.
Because Starlink benefits from the in-house launch capacity of SpaceX, its launch cost equates to the marginal launch cost (removing profit margin) of Space X, which currently gives them a significant cost and capacity advantage over other competitors. The Space X website lists the standard payment plan for a Falcon 9 launch at $62 million. 51 This price level has been steady since 2016, before the current development of reusable stage 1 rockets. At this pricing, it was already winning contracts over other launch vehicles. 52 46 See  19 November. 49 Once at scale deployment, consumer financing may also come into play such that installations will be $0 up-front for qualifying individuals. 50 Amazon. 2020 With the development of reusable rockets, starting with the first stage component of Falcon 9, the standard price for a Falcon 9 launch appears to have dropped to $50 million. 53 Estimates suggest this could drop further to $36 million with reusability across all launch costs, 54 and even SpaceX chief executive officer Elon Musk has suggested that with further optimization of launch operations and full reusability, the marginal cost could drop to $5 million-$6 million for a Falcon 9 launch. 55 This suggests, even at the $36 million price level, the launch price per payload is already one of the cheapest on the planet (footnote 52). Falcon Heavy's payload is roughly 3 times larger at 1.5 times the price and the company is currently developing an even larger rocket system, the Interplanetary Transport System, renamed StarShip, which is designed to dramatically increase space payload capacity up to 150,000 kilograms, or up to 400 Starlink satellites, per launch (Figures 9 and 10).
In-house satellite production could also factor as a cost and management advantage for Starlink compared to other low Earth orbit constellations that outsource satellite manufacturing.
SpaceX itself manufactures Starlink satellites whereas other LEO providers rely on sourcing satellites from external parties, particularly from Boeing and Airbus. Telesat, for example, is working with Airbus, Thales Alenia Space, and Maxar for its satellites, while OneWeb has partnered with Airbus. 56 Amazon has announced that it will be building its own satellites at its facilities in Washington state in the US.
The vertical integration in both Amazon's and Starlink's production and launch will give them deployment advantages and cost efficiencies. While satellite production costs are proprietary information, disclosures by SpaceX suggest each satellite's capital expenditure cost is below $500,000. 57 It is worth noting that each Starlink satellite can transmit roughly 20 Gbps of capacity per satellite, while an HTS GEO has a capital expenditure range between $200 million for 10 Gbps on the low end, to $700 million for 1,000 Gbps on the high end, including launch costs. 58 This makes LEO significantly cheaper on a per Gbps basis, but each geographic service area requires many LEOs in orbit versus only one GEO. One should note that LEO satellites are designed for shorter life spans (see Table 3 showing LEO life spans ranging from 5-12 years) compared to GEO satellites with a minimum 15   Differences in target markets: consumer, enterprise, government, and telecommunications (backhaul).
Differences in target markets and customer base are difficult to determine prior to the launch of commercial services. However, Telesat, for example, has publicly expressed a focus on government clients and large enterprises requiring global connectivity. OneWeb, prior to restructuring, focused on connecting the last mile and unconnected communities in emerging markets as well as mobile cellular backhaul. Starlink has focused on unserved consumers with income levels that can afford its service offerings, as well as noting a potential strategy to deploy community wireless fidelity (Wi-Fi) models. Of note is that Starlink has partnered with Microsoft to connect its network directly to the latter's Azure cloud and data center infrastructure, suggesting a focus on enterprise and government clients. Similarly, Amazon's Project Kuiper will no doubt leverage the Amazon Web Services cloud as part of its service offering.

IV. In FoCUs: stARLInK's DepLoYMent, DIFFeRentIAtIon, AnD VIABILItY
Starlink's 1,445 satellites launched is by far the most advanced stage of deployment by any lower Earth orbit constellation. Coverage across Asia and the Pacific may not occur until mid-to late-2021 and into 2022.

SpaceX is an
While Starlink's current public beta trial is serving customers in northern US and Canada between latitudes of 45º and 52º, the expectation is that service will expand to the 33º latitude by the beginning of 2021, and then to full equatorial coverage the middle to end of 2021. Commercial operations have already begun in Australia and New Zealand, focused on establishing earth stations. 59 Note that commercial operations may begin once 1,440 satellites are in orbit, but publicly available calculations of broadband throughput are based on a constellation estimate of 4,425 satellites, making the time frame for coverage and throughput across Asia and the Pacific uncertain.

Financial Viability
Starlink's current public beta pricing structure ($99 per month) suggests strong revenue potential from a discrete base of subscribers. Even before the start of commercial service, Starlink was awarded an $885 million contract by the FCC in December 2020, to provide broadband internet service to 642,000 locations. 60  7-10 years of initiating commercial revenues. 61 At the current beta pricing level, such revenue projections would require roughly 25 million subscribers across the world where Starlink has service footprint. In comparison, US fiber cable provider Comcast alone had 26.9 million residential high-speed internet subscribers in the first quarter of 2020. 62 As such, Starlink's revenue is forecast to dwarf SpaceX's launch revenue by the mid-2020s. 63 The combined revenue streams (satellite broadband and space launch) are expected to grow SpaceX into a $100 billion+ company in terms of valuation. 64

V. oppoRtUnItIes AnD BARRIeRs to LeVeRAGInG LoW eARth oRBIt sAteLLItes In DeVeLopInG MeMBeR CoUntRIes
Satellite connectivity demand in Asia and the Pacific is forecast to grow rapidly.
Satellite broadband is a growing, and dominant, segment of the overall satellite sector. Overall, the sector's revenues have been growing almost 4% per year and have increased by $38 billion over the last 5 years. The space market is expected to reach between $1.1 trillion to $2.7 trillion in the next 30 years, with the provision of internet access via satellite predicted to account for 50%-70% as the primary driver of growth. 65 Figure 11 highlights the growth of GEO HTS bandwidth demand driven by broadband access and enterprise data traffic. Figure 12 forecasts that LEO and MEO satellites will capture half of the market for high-throughput satellite communications by 2027.
Low-cost, high-capacity coverage over an entire country gives governments an alternative to national satellite launches.
A number of governments in Asia are directly investing in launching their own national satellites (Table 4) with capital expenditures (CAPEX) and launch costs in the hundreds of millions of dollars. Starlink and other LEO constellations could serve as an alternative in supporting universal access objectives of national satellites. Figure 13 compares generations of GEO satellites, including HTS. While a direct comparison to LEO is challenging due to the differences in constellation design, assuming Starlink's total constellation capacity of 20 Tbps and $10 billion in forecast CAPEX investment, the equivalent CAPEX per Gbps ratio would be $0.5 (for comparison purposes only).  average pricing than the legacy markets, such as Western Europe and Japan (and to a lesser extent, North America).
The ARPU for HTS capacity stood at around $232/Mbps/month in 2017. After years of strong declines due to the impact of bulk leases and vertically integrated broadband services, the average HTS capacity ARPU will more progressively erode to about $165/month in 2027 (CAGR 2017(CAGR -2027. This global average combines a wide range of schemes, with pricing ranging from less than $100/Mbps/month for certain fully leased payloads to more than $1,000/Mbps/month for certain niche services, such as premium video services. Significant uncertainty remains about the anticipated price levels for the VHTS systems, which could have a significant impact on average ARPU at the end of the forecast period.

Affordability: price discrimination and community Wi-Fi deployments
At the current public beta pricing level, Starlink's $99 monthly plan is not affordable for many consumers in developing Asia. However, variable pricing by market (at different purchasing price levels) could result in a more affordable service offering. Similarly, community Wi-Fi models could be deployed, such as those being implemented by Hughes/Express Wi-Fi in Indonesia and Latin America where an individual subscription supports time-or data-bound service to potentially hundreds of users consuming small data bundles (in the megabytes) through a publicly accessible Wi-Fi access point. Areas of limited subscriber base may be opportunities for direct or subsidized partnerships.
Complex global networks require various regulatory approvals across countries, technologies, and business segments.
In the space segment, US-based entities such as Starlink and Project Kuiper require regulatory approval from the FCC as well the ITU. At the 2019 World Radiocommunication Conference, the ITU put into place new rules for non-geostationary orbit (NGSO) constellations to retain spectrum rights. As the ITU has noted: "Under the newly adopted regulatory approach these systems will be required to deploy 10% of their constellations within 2 years from the end of the current period for bringing into use, 50% within 5 years, and complete the deployment within 7 years." 66 In the ground segment, Earth stations or gateways will require technical and business licensing, and service provision to customers will require regulatory approval in every country of operation. How expeditious this process will be depends on the ease of doing business between countries, and the resistance that LEO companies may encounter as they try to enter markets dominated by incumbent operators.
For example, in India, telecom operators are already challenging the expected market entry of NGSO services. In many Asian markets, internet service licensing is restricted by regulators and can be a lengthy process. Interference issues will also have to be addressed if NGSOs interfere with the radio frequency propagation of GEOs. ITU rules are in place to limit power management of radio signals to prevent interference between transmission to and from satellites, as well as other adjustments such as changing frequency bands and reporting beams to avoid interference with GEO beams. The onus is on NGSOs to ensure they are not interfering; otherwise, they will have to shut down problematic assets. 67 Differences in licensing burdens between countries will impact LEO companies' ability to go to market with a uniform approach. The rules and requirements for internet service providers can be very different from those applying to supply backhaul to local internet service providers, potentially causing a delayed market entry or even a decision to not cover certain countries (

Other Risks and Challenges
Interference with astronomical observation has been cited as a concern with regard to LEO satellites. For example, after initial launches of Starlink satellites in June 2019, the altitude and design of the first satellites resulted in the satellite being visible from Earth with the naked eye. As a result, Starlink adjusted the design in subsequent launches. These are known as "DarkSats," with a darkening coating to make the satellites less visible to stargazers and ground-based observatories. However, issues remain for wide-field sky surveys, particularly at twilight or sunset (with reflection from the sun). For example, 30%-40% of exposures made at some ground observatories around twilight and dawn could be impacted. 68 Further design and operational adjustments are in progress.
Other concerns have been raised regarding the increase in man-made space objects (space debris), particularly in LEO, which could potentially result in cascading collisions. However, many of the LEO constellations in development include satellites designed for a useful life span of less than half of the 15+ years of GEO satellites (

VI. ReCoMMenDAtIons: WhAt DeVeLopInG MeMBeR CoUntRIes CAn Do to LeVeRAGe the oppoRtUnItY pResenteD BY LoW eARth oRBIt sAteLLItes
As the new LEO constellations begin offering commercial service in 2021, actions can be taken today to ensure that low-and middle-income countries are well positioned to take advantage of potential costsavings and increased coverage. These include: Ensure flexible and streamlined licenses procedures for domestic internet service providers and satellite broadband providers.
Licensing regimes of internet service providers differ widely between countries. Good practice licensing policy includes simple registration with administrative bodies, rather than complex and onerous procedures that require legislative approvals. Regimes that require congressionally awarded franchise licenses, for example, create a significant barrier to entry and allow for rent seeking by firms that are able to navigate the cumbersome processes.
Allow for satellite provision of international internet capacity without a requirement for domestic ground stations to route traffic to and from satellite transponders.
So-called data localization regulations, that require internet traffic within a country to pass through nationally mandated infrastructure points of presence, create barriers to deployment and increase cost and complexity. National requirements, which stipulate that satellite transponder data traffic must transit through a domestic ground station (rather than be allowed to transmit from a ground station located in another country), create burdensome infrastructure deployment requirements. LEO constellation firms may decide that these represent too large a cost for markets with small potential revenue opportunities. It is common for internet traffic to be transmitted via satellite transponders that originate in a country where the satellite's ground station is located, particularly for smaller markets whereby providers need to aggregate total traffic being transmitted to achieve economies of scale.
If future versions of LEO satellites are able to deploy optical inter-satellite links, optical laser beams will directly transmit traffic between satellites in space. This will create further efficiencies in space network segments and could allow satellite constellations to reduce the number of ground stations needed. Regulations that impose a domestic ground station requirement would negate these efficiencies. In cases where the requirements to transmit data locally are driven by sovereignty and cybersecurity justifications, it may be possible to put in place policies that preserve sovereignty and cybersecurity principles without requiring data localization.
Reduce or remove import tariffs, quotas, or local manufacturing requirements for satellite user terminals.
Hardware, software, and licensing costs for user terminals (consumer equipment, including satellite antennas, modems, and Wi-Fi access points) contribute to the total cost of satellite internet service, and user terminals are already priced significantly above other broadband internet service hardware such as cable modems or stand-alone Wi-Fi access points. Reducing the production cost of user satellite terminals requires mass manufacturing and economies of scale; the benefits of which can best be passed to consumers if trade barriers do not further increase the cost of devices. Engage in regional discussion and cooperation both in terms of regulatory convergence to improve the "ease of doing business" for LEO satellite connectivity as well as for potential demand aggregation between markets.
The expansion of satellite coverage and increase in available capacity will particularly benefit countries that have limited international internet bandwidth being supplied by undersea and terrestrial fiber optic cables, such as small island developing states and landlocked developing countries. Regional integration between countries and their neighbors could accelerate deployment timelines by easing the more cumbersome regulatory issues and forging convergence of licensing issues. Similarly, aggregation of demand between markets could present more attractive deployment timeline prioritization and may also generate bargaining power for favorable terms of service.
Deploy universal access funding to support public access through community Wi-Fi deployments.
Individual subscriptions to LEO service may still be commercially priced at levels beyond the affordability thresholds for low-and middle-income consumers in developing countries. However, public access could be subsidized by national and local governments partially or in full, either one-time or recurring, for the initial hardware and/or trial service, or to provide free public Wi-Fi. A number of countries in Asia and the Pacific fund public access by subsidizing free Wi-Fi services, of which LEO could provide connectivity in remote areas. Public access services via satellite could support connectivity to unconnected and underconnected government offices, schools, hospitals, and other public facilities. Public-private partnerships may be helpful in this regard, both in terms of consumer access as well as upstream infrastructure development (such as ground stations).
Invest in developing accurate, publicly available, mobile coverage and network infrastructure availability maps to better identify geographic areas that are unserved and underserved by current service providers.
Comprehensive, up-to-date, and publicly available (open data) mapping of network infrastructure (mobile cellular base stations, fiber optic cable routes, microwave backhaul lines, points of presence, internet exchange points, and others) as well as service coverage areas will better inform public and private decisions on where to allocate resources and investment to expand service availability. A tool made available by ITU in the regional and national planning is available at https://www.itu.int/itu-d/tndmap-public/.
Consider supporting consumer financing for user terminals.
To increase the affordability of user terminals for individual subscriptions, governments could work with consumer finance institutions to develop programs that amortize the cost of user terminals over a service period, reducing the up-front cost barrier to access LEO service.

Support demand generation activities.
These include enhancing the digital skills of the general public with emphasis on the youth. In addition, technical skill development, such as VSAT deployment and maintenance for remote areas can help not only in connectivity to communities, but also assists in recovery of post-disaster telecommunication. In addition, demand generation could be considered to sustain connectivity. For instance, pacific connectivity and rural connectivity could go with digital applications and services, such as e-commerce, trade, and financial services, so that the community and households will be able to pay.