[SSC21-XII-10] Satellite Constellations - 2021 Industry Survey and Trends

Erik Kulu NewSpace Index, Nanosats Database, Kepler Communications [email protected]

ABSTRACT

Large satellite constellations are becoming reality. Starlink has launched over 1600 spacecraft in 2 years since the launch of the first batch, Planet has launched over 450, OneWeb more than 200, and counting. Every month new constellation projects are announced, some for novel applications. First part of the paper focuses on the industry survey of 251 commercial satellite constellations. Statistical overview of applications, form factors, statuses, manufacturers, founding years is presented including early stage and cancelled projects. Large number of commercial entities have launched at least one demonstrator satellite, but operational constellations have been much slower to follow. One reason could be that funding is commonly raised in stages and the sustainability of most business models remains to be proven. Second half of the paper examines constellations by selected applications and discusses trends in appli- cations, satellite masses, orbits and manufacturers over the past 5 years. Earliest applications challenged by NewSpace were AIS, Earth Observation, Internet of Things (IoT) and Broadband Internet. Recent years have seen diversification into majority of applications that have been planned or performed by governmental or military satellites, and beyond.

INTRODUCTION but they are regarded to be fleets not constellations. There were much fewer Earth Observation com- NewSpace Index has tracked commercial satellite panies in 1990s and 2000s when compared to com- constellations since 2016. There are over 251 entries munications and unclear whether any large constel- as of May 2021, which likely makes it the largest lations were planned. 1992 Land Remote Sensing public database. The initial target was nanosatel- Policy Act allowed U.S. commercial companies to lites, but scope was expanded to all satellites when a enter the satellite imagery business.12, 13 Digital- large quantity of small satellite constellations started Globe (WorldView, EarthWatch)12 and Orbimage to be announced. For comparison, there were only (GeoEye)13 were founded in 1992 and were followed 60 projects listed at the end of 2017. by Space Imaging in 1994.14 First commercial EO Literature review was performed and surveys satellite launches were in 1997, EarlyBird-1 for Dig- about constellations seem to be relatively sparse. italGlobe and Orbivew-2 for Orbimage. The first Seraphim has published a Constellation Map1 along high-resolution commercial EO satellite IKONOS with many similar market maps from other entities. from Space Imaging launched in 1999. French Space One paper surveyed 90 missions and had basic statis- Agency established subsidiary Spot Imaging to sell tics.2 Selected papers have focused on Earth Obser- commercial imagery in 1982.15 Israel’s ImageSat vation3 or telecommunications.4–6 launched EROS-A in 2000 and planned a 4-satellite constellation, but it was cancelled.16 First Wave of Commercial Constellations Current Second Wave of Constellations Around 15 commercial LEO constellations for communications (broadband, messaging and voice) After the failures of satellite communications were announced in the late 1980s and during the companies, commercial Earth Observation satellites 1990s.7–11 Most of them had failed by the early continued to be launched about one every 2-3 years. 2000s including Teledesic. While Orbcomm, Irid- The market saw consolidation, because Orbimage ium and Globalstar were successful in launching acquired Space Imaging in 2006 to form GeoEye and their first-generation constellations, they also went DigitalGlobe bought GeoEye in 2013.17 bankrupt and were restructured by the early 2000s. Earliest entries in the database are AprizeSat Commercial geostationary communications satellites and SaudiComsat with the first launches in 2002 and have been launched since 1960s (Telstar, Intelsat), 2004, respectively. ExactEarth started its constella-

Kulu 1 35th Annual Small Satellite Conference tion in 2008 and some payloads were hosted aboard MARKET SURVEY AprizeSat. These were all smaller microsatellites with a mass of about 12 kg and offered Internet of Survey Criteria Things (IoT) and Machine-to-Machine (M2M) com- ˆ Satellite constellation - Number of similar munication services or AIS message collection. satellites, of a similar type and function, de- RapidEye was created in 1998, announced financ- signed to be in similar, complementary, or- ing in 2004 and launched 5 spacecraft in 2008.18 It bits for a shared purpose, under shared con- entered bankruptcy in 2011 and was sold when sales trol.37 Sometimes defined as a set of satellites of $18 million were not strong enough to pay back working together in order to provide a service debts.19 Planet was founded in 2010 as Cosmogia, or a group of satellites with a common pur- launched 3U CubeSats in 2013 and the first batch pose.38, 39 Different from satellite programs of 28 in early 2014.20 Skybox, founded in 2009, was and fleets. inspired by CubeSats,21 raised $91 million by April ˆ 3 or more spacecraft - Minimum required for a 2012,22 launched spacecraft in 2013,23 was bought continuous coverage in geostationary or Mol- by Google for $500 million in 2014 (renamed Terra niya orbits. Literature review revealed mini- Bella) and sold to Planet in early 2017.24 mum satellite counts of 240 and official defini- O3b (now SES) was founded in 2007 and tion has not been found, but 3 appropriately 25 launched first 4 MEO satellites in 2013. Orb- filters out many satellite pairs. comm, Iridium and Globalstar developed second- ˆ generation constellations. After finishing OG-1 in Commercial focus - Primarily owned, financed 1999, Orbcomm followed with replenishments in and managed by commercial entities for the 2008 and launched OG-2 in 2014 and 2015.26 Irid- purpose of providing a commercial service. Ex- ium contracted for the NEXT constellation in 2010 cluding government, military, academic, scien- and launches were between 2017-2019.27 Globalstar tific and non-profit constellations. Likely to be launched replenishments in 2007 and next genera- expanded in the future, because all constella- tion from 2010 to 2013.28 tions will have an impact to the space economy including launch and manufacturing services. News articles and speculations about OneWeb and SpaceX constellations appeared throughout ˆ Announced or launched after 2002 - Filtering 2014.29–31 OneWeb (WorldVu) was founded in out the first-generation constellations for Irid- 2012 and was publicly announced in 201532 as ium, Globalstar and Orbcomm as of now, in well as Starlink in early 2015.33 OneWeb entered addition to other projects from the 1990s. bankruptcy in 2020, was bought and has continued with satellite deployments.34 LeoSat was founded in Applications Classification 2013, raised over $10 million and shut down in 2019 There are constellations with multiple applica- after failing to attract further funding.35 tions, both by naming or the same satellite having First CubeSats started launching in 2003, popu- different payloads. The following list of keywords, larity started to increase around 2008 and launches applications and fields is not exhaustive. really took off in 2013, coinciding with the constel- lations from Planet and Spire. This partially cor- ˆ Optical Earth Observation (EO) - VNIR (Vis- relates with an increase in constellation companies ible and Near-Infrared), Multispectral - Earth from 2009 and with the jump in 2015 as seen on Observation and remote sensing are very broad Figure 7. ”The CubeSat Orbcomm” was proposed terms, but here defining optical EO as using in 2010. Sky and Space went public in 2016 and passive silicon CCD or CMOS image sensors partially kicked off the IoT/M2M field popularity. to collect imagery of Earth in visible and near- ITU first received half a dozen filings for large infra-red wavelengths.41 Most multispectral constellations in late 2014 and early 2015.36 Perhaps instruments have 3 to 20 spectral bands and largely inspired by Planet, Spire, Skybox, OneWeb, there is overlap with VNIR constellations.42 SpaceX and by the general NewSpace prevalence. ˆ Infrared EO - Divides into SWIR (Short-Wave Many constellations are raising more and more Infrared), MWIR (Mid-Wave Infrared) and funding even when business models have not been LWIR (Long-Wave Infrared) and often also proven. Latest trend is SPACs: Spire, BlackSky and known as thermal imaging. Commonly used Arqit with likely more upcoming. After the recent for weather and climate monitoring, and for SEC rule changes, this approach might slow down. forest fire detection for example.43

Kulu 2 35th Annual Small Satellite Conference ˆ Hyperspectral EO - More than 20 spectral satellites. Aim is to connect space assets to bands and typically hundreds of narrow spec- mission control centres in real-time and to in- tral bands, which enables new applications in crease the amount of downlinked data.59 agriculture, vegetation, geology and water re- ˆ QKD (Quantum Key Distribution) - Satellite- 44 sources to identify specific organic matter. based QKD uses laser communications be- ˆ SAR (Synthetic Aperture Radar) - Emits mi- tween two users to produce a common random crowave radiation and collects the reflected ra- secret key at a distance, which is then used diation to form an image. Synthetic aperture to encrypt messages sent over a standard com- is used to synthesize a long antenna by com- munication channel. Any eavesdropper on the bining reflected signals as the satellite moves quantum channel will introduce disturbance to along orbit. Spaceborne radar satellites, for the system that is in superposition state, which example in X, C or L radio bands, see through can then be detected by the two users.60, 61 45, 46 clouds and are not dependent on sunlight. ˆ ADS-B (Automatic Dependent Surveil- ˆ VR/AR (Virtual Reality, Augmented Reality) lance–Broadcast) - Every aircraft transmits - Subset of EO with stereo or 360 degree cam- its identification, position, altitude, speed eras for 3D pictures and videos. and other information, which is relayed to air traffic controls for more accurate situational ˆ Video - Subset of EO for high-resolution video awareness.62 Space-based ADS-B is used for capture and possibly real-time transmission.47 tracking aircraft beyond the reach of land- ˆ Emissions Monitoring - Remote sensing, quan- based receivers, but it could be expanded to tifying and locating greenhouse gas emissions, space-based air traffic services to provide more 63 for example CO2 and methane, using space- information to pilots. borne imaging spectrometers.48, 49 ˆ AIS (Automatic Identification System) - ˆ Internet (Broadband) - Wideband high-speed Space-based AIS is used to track ships beyond satellite communications offering real-time In- the reach of land-based receivers by listening ternet or data backhaul service.50 to the VHF radio transmissions of ships iden- 64 ˆ IoT / M2M - Using satellites to extend the tity, position, speed and heading. coverage of IoT (Internet of Things) and M2M ˆ VDES (VHF Data Exchange System) - In (Machine-to-Machine) communications for re- some ways the next generation of AIS, which is mote areas. Often one-way or two-way mes- starting to enter operation as of 2021. VDES sages and low data rates in narrowband.51, 52 encompasses AIS and adds a higher rate two- way data communications system, which also ˆ Store-and-Forward - Subset of satellite com- foresees a special satellite link. At the same munications, both narrowband and wideband, time, new dedicated channels will solve AIS which is not real-time and data is stored on overloading problems in crowded waters.65 the spacecraft until a ground station pass.53 ˆ Asteroid Prospecting - Mapping composition ˆ 5G - Subset of satellite communications and of asteroids for asteroid mining purposes. Here IoT / M2M using 5G standards including NB- envisioned as rendezvous or flyby spacecraft IoT, that may be compatible with common ter- for fine measurements, sometimes with MWIR restrial equipment including smartphones.54, 55 sensors. Could be also asteroid discovery, char- ˆ Satellite-to-Cellphone - Subset of satellite acterization and cataloguing constellation in communications, which is capable of sending LEO to increase the known population.66 and/or receiving text messages, voice calls or ˆ On-Orbit Inspection - Satellites intended to even real-time broadband data from unmodi- maneuver nearby to other spacecraft for iden- fied mobile phones to spacecraft.56 tification and to physically inspect deployables ˆ Laser Communications - Constellations ca- and malfunctions in case of problems.67 Sub- pable of space-to-ground optical communica- field of On-Orbit Satellite Servicing. tions. Foreseen for downlink and uplink, but ˆ 57, 58 In-Orbit Computing - Dedicated orbital lasers are also used for inter-satellite links. servers for processing data from other satel- ˆ Orbital Data Relay (Internet in Space, Space lites in space, before sending to ground or back Data Relay) - Providing telemetry & telecom- to the satellites.68, 69 Can be related to In- mand and ground station service to other Orbit Computing and Orbital Data Relay. Not

Kulu 3 35th Annual Small Satellite Conference the same as orbital edge computing, which in- craft using microwave or laser power transmis- cludes payload data processing boards and ma- sion.81 Sub-field of Space-Based Solar Power. 70 chine learning software. ˆ SSA (Space Situational Awareness) - Space- ˆ In-Orbit Data Storage - Space-based cloud based SSA for detecting, monitoring and de- data storage service on-board satellites or in- termining the orbits other satellites and space side laser beams, which could be highly se- debris. To enable higher revisit and consec- cure when using narrow radio or laser beams. utive observations of objects.82, 83 Often per- Could be related to In-Orbit Computing and formed with infra-red space telescopes, which Orbital Data Relay applications.71 could also be used for asteroid astronomy. ˆ Constellation-As-A-Service - Full end-to-end ˆ Space Weather - Monitoring the effects of service to build and operate a constellation for space weather to the ionosphere and magneto- a customer. Sometimes called Satellite-Data- sphere by measuring energetic charged parti- As-A-Service and can include payload develop- cles to help improve space weather forecasting ment. Related to hosted payloads when shar- models.84, 85 Satellites sometimes conceived to ing the same spacecraft. Many more commer- be located in Earth-Sun Lagrange-points to in- cial entities are offering or planning to offer crease warning time for solar flares.86 this solution than shown in Figure 1. ˆ Hosted Payloads - Could be one-off or numer- Form Factors Classification ous in-orbit demonstration & verification mis- Constellations can consist of spacecraft with dif- sions for payloads and subsystems, or alter- ferent masses and form factors. The database aims natively scientific or commercial microgravity to capture and publicize all of the used or announced experiments. Piggybacked on individual satel- size classes. In this survey, only the latest known lites or on another constellation.72, 73 Some- and most likely form factors for the majority of the times related to Constellation-As-A-Service satellites in a constellation have been presented for and also called rent-a-satellite. clarity. No official standard for mass definitions ex- 87 ˆ GNSS (Global Navigation Satellite System) ists and the FAA mass classes are too exhaustive and PNT (Positioning, Navigation and Tim- so henceforth using the following reduced categories: ing) - Commercial satellite navigation systems. ˆ PocketQubes (Femtosatellites) - Sizes from 1p Often envisioned to be based in LEO to pro- 88 vide better performance, much stronger sig- to 3p, where 1p is ≈ 5 cm × 5 cm × 5 cm. nals, encryption and 3-D location when com- ˆ CubeSats (Nanosatellites) - Satellites follow- pared to free GPS, Galileo, GLONASS and ing the CubeSat Design Specification from BeiDou.74, 75 Possible that existing GNSS con- 0.25U to 16U, where 1U is 10 cm × 10 cm stellations signals will be used as inputs.76 × 11.35 cm.89 The general category is for constellations that will use CubeSats based on ˆ Orbital Display - Using formation flying satel- available information, but the exact type has lites to compose a message or advertisement not been made public. visible to the naked eye on Earth by using large deployable sails that reflect Sunlight.77 ˆ Hosted - Constellations hosting payloads on other multi-use satellites or constellations, ˆ RF Spectrum Monitoring and Geolocation - where spacecraft resources are shared and Listening for radio emissions with formation- spacecraft mass has lower correlation to the flying spacecraft to identify and triangulate payload performance and specifications. the sources. Common application is ships or vehicles that do not have AIS enabled.78 ˆ Microsats - 10 kg to 100 kg, except larger CubeSats which can be up to 25 kg, because ˆ Weather Monitoring - Measuring vertical at- those masses are rarely shared publicly. mosphere profiles of various parameters e.g. pressure, temperature, humidity and wind ˆ Smallsats - 100 kg to 500 kg. Small satellites speed for weather and climate forecasting. term is often used broadly for any satellite be- Common technologies are GNSS-RO,74 GNSS- low 500 kg. R,79 microwave radiometry and others.80 ˆ Satellites - 500 kg and above. ˆ Wireless Energy Supply - Providing satellites with extra electrical power from other space-

Kulu 4 35th Annual Small Satellite Conference Status Classification ˆ VLEO (Very Low Earth Orbit) - Here defined as sustained orbits with altitudes in the 250- Categories to indicate the current status of satel- 350 km range.90 Earth Observant’s Stringray lite constellations by consolidating various accessible spacecraft will be designed to fly at 250 km al- information and in some cases an educated guess has titude. Albedo Space91 is expected to use sim- been made: ilar orbit range. SpaceX V-band constellation between 336-346 km is also called VLEO.92 ˆ Launched - Constellation fully launched with Literature defines VLEO below 450-500 km no replenishments foreseen due to long ex- altitude and Super Low Earth Orbit (SLEO) pected lifetime and already existing orbital with perigee below 300 km, but both defini- spares. Iridium is a fitting example. tions seem to be rarely used in practice.93 ˆ Launched and replenishing - Most of the ˆ LEO (Low Earth Orbit) - Up to 2000 km planned constellation has been launched and from the Earth’s surface, excluding the pre- the expected size has been achieved. New it- vious VLEO definition. erations are being sent to orbit every 1-3 years ˆ as satellites retire or deorbit. Planet and Spire MEO (Medium Earth Orbit) - Altitudes be- fit well here. tween 2000 km and 35786 km. ˆ GEO (Geosynchronous Earth Orbit) - Alti- ˆ Launches ongoing - Constellation deployment tude of 35786 km. of identical or similar satellites actively in progress or ramping up. Starlink and OneWeb ˆ HEO (Highly Elliptical Orbit) - Low perigee are best examples. (1000 km) and high apogee (35786 km). In- cludes Geostationary Transfer Orbit (GTO), ˆ Prototype(s) launched - One or more first Tundra orbit and Molniya orbit, where the lat- satellites or hosted payloads launched, which ter two are special cases of Highly Inclined El- serve as prototypes or pathfinders. liptical Orbit (HIEO). ˆ Prototype development - First payload or ˆ Lunar - Constellations intended to orbit spacecraft is being actively worked on. Com- around the Moon. Sub-fields such as Low Lu- pany seems to be growing and likely has an- nar Orbit (LLO) will be created as applicable. nounced one or more rounds of funding. Likewise with other Solar System bodies. ˆ Early stage / Concept - A few announcements, presentations and pitches for a constellation Delay Classification idea, but likely in the stages of starting devel- opment, looking for funding and gaining trac- Comparing launches and upcoming manifests to tion with a very small team. statements about launch cadences and deployment intentions. Plans change and not all delays should ˆ Dormant / Unknown - Early signs towards to be taken as a bad sign, but nevertheless gives some the development has slowed down or stopped, indication about the company and industry. but in rare cases can be explained with stealth mode or temporary setbacks. ˆ Cancelled - Cancelled and dormant constella- tions where no more launches are expected. ˆ Cancelled - Company is bankrupt, website and social media channels have been quiet for more ˆ Generally on-time - Up to 6 months of delays, than 1-2 years, lack of funding announcements, some of which can be due to launch delays. team seems to have been disbanded and/or the ˆ Partially behind - Approximately 6-12 months idea never entered implementation phase. of delays. ˆ Retired - Constellation, which has been re- ˆ Year behind - More than 1 year of delays. moved from active service, but satellites could ˆ Years behind - More than 2-3 years of delays. still be in orbit and even operational. ˆ Unknown - No public announcements found and unable to make a comparison. Orbit Classification ITU and FCC primarily differentiate non- geostationary (NGSO) and geostationary (GSO) or- bits, but in the database large majority are NGSO.

Kulu 5 35th Annual Small Satellite Conference 2021 STATISTICAL OVERVIEW one or more prototypes while 29% are in develop- ment phase. About 25% have cancelled or dormant Statistical overview of the 251 entries included in status and this could be larger. The one retired con- this commercial satellite constellations survey. stellation is Planet’s RapidEye.19

Applications and Fields Launched 9 Launched & replenishing 2 Many constellations perform multiple functions Launches ongoing 24 with identical or similar satellites. Figure 1 counts Prototype(s) launched 45 services independent of spacecraft multi-usage. In Prototype development 83 other words, applications and fields have been counted Early stage / Concept 30 separately and the sum does not equal the number Dormant / Unknown 18 of entries in this survey. As seen on Figure 1, IoT / Cancelled 39 M2M is the most popular followed by Optical Earth Retired 1

Observation, Broadband Internet, SAR and AIS. 0 20 40 60 80

5G 5 Figure 2: Constellations Current Status ADS-B 8 AIS 18 Active Debris Removal 3 Asteroid Prospecting 3 Form Factors Asteroid Tracking 1 Astrophotography 1 The distribution of spacecraft size and type Blockchain 1 classes is visualized on Figure 3. CubeSats are the Constellation-As-A-Service 7 Data Storage 7 most popular segment with a combined 48%, includ- Earthquake (ELF) 2 ing the 15% which are likely to be CubeSats, but Emissions Monitoring 5 specific form factor is currently unknown. Followed GNSS / PNT 5 GNSS Radio Occultation 3 by microsatellites with 19% and small satellites with GNSS Reflectometry 1 14%. Mass classes for about 10% of the constella- Hosted Payloads 4 tions have not been made public yet. Hyperspectral Earth Observation 16 In-Orbit Computing 2 Among CubeSats, 3U continues to be the most In-Orbit Inspection 5 popular thanks to Planet and Spire. Astro Digital’s Infrared Earth Observation 15 Broadband Internet 36 predecessors launched two 6U pathfinders in 2014. IoT / M2M 61 Sky and Space, GeoOptics and Kepler were devel- Laser Communications 12 oping and started launching 6U CubeSats to gain Minerals Prospecting 1 On-Orbit Servicing 2 more capabilities. Kepler has 13 of 6U CubeSats in Optical Earth Observation 42 orbit, which is largest in that class. Many are now Orbital Data Relay 13 working with 12U-16U sizes. Swarm went into the Orbital Display (Advertising) 2 QKD (Quantum Key Distribution) 3 opposite direction by using 0.25U CubeSats. RF Monitoring / Geolocation 7 PocketQubes have been slower to become popu- SAR 21 lar, but now Alba Orbital, Fossa Systems and others SSA 9 Satellite-to-Cellphone 4 are developing constellations. Space Weather 1 Some entities are moving towards larger microsatel- VDES 3 Video 5 lites due to stacking, appendage, power and thermal Weather 10 limitations of CubeSats. Wireless Energy Supply 2 PocketQube 7 0 20 40 60 0.25U 2 1U 2 2U 2 Figure 1: Constellations Applications 3U 25 6U 29 8U 1 12U 6 16U 6 Current Status CubeSat 38 Hosted 4 Figure 2 shows the current activity and develop- Microsat 49 Smallsat 35 ment status of the commercial constellations. Only Satellite 23 about 4% are considered launched with 8% being ac- Unknown 22 tively launched. Approximately 20% have launched 0 10 20 30 40 50 Figure 3: Constellation Form Factors

Kulu 6 35th Annual Small Satellite Conference Manufacturers VLEO 2 Subcontracting the first or more spacecraft to a LEO 236 satellite manufacturer does not ensure a constella- tion order. Some have used the services of multiple MEO 4 integrators. Many have later switched to building HEO 1 the spacecraft in-house. The database lists multiple GEO 6 manufacturers when applicable, but in this survey only the latest used or announced satellite producer Moon 2 for a constellation has been presented on Figure 4. 0 50 100 150 200 250 About 37% are building spacecraft in-house and 33% of constellations are currently unknown. Rest Figure 5: Constellation Orbits divide between a large number of manufacturers. Few manufactures have full contracts and only a small number of constellations have been fully funded. Delays and Launch Cadences Going further, it is likely that most constellations Announced launch cadences were compared with will not be continued or completed. Any manufac- the current launch status to create Figure 6. Most turer being awarded a large contract is a major suc- companies have not announced their plans and de- cess, when compared to demonstration missions. lays are very common. Only a small number are on-

Unknown 83 time and delayed less than 6-12 months from plans. AAC Clyde Space 7 Airbus 1 Bellatrix 1 Generally on-time 8 Blue Canyon 2 Partially behind 11 DFH Satellite 1 Deep Space Industries 1 Year behind 17 German Orbital Systems 1 GomSpace 5 Years behind 13 ISISpace 4 In-house 95 Unknown 172 LeoStella 3 Loft Orbital 2 Cancelled 30 MDA 1 Maxar 1 0 40 80 120 160 200 NanoAvionics 11 Nexeya 1 Figure 6: Constellation Delays Open Cosmos 1 QinetiQ 2 SFL (Space Flight Laboratory) 3 SSL 1 Companies Founded SSTL 3 Sierra Nevada Corporation 1 Founding dates since 1991 are on Figure 7. Around Space Inventor 3 20 were left out with the earliest from 1895. There Spacety 2 Spire 2 is a clear increase after 2009 with major peaks in Thales Alenia 5 2015-2016. Decrease could be a sign of waiting to Tyvak 6 York Space Systems 2 see how existing entities will perform economically. 36 0 20 40 60 80 100 35 35

Figure 4: Constellation Subcontractors 30 27

25 23

Orbits 20 17 17 Figure 5 divides the NewSpace constellations by 15 12 orbit types. LEO orbit is by far the most popu- 10 10 8 7 7 lar with 94% of 251. Recently, very-high resolu- 5 4 5 3 3 3 tion Earth Observation constellations in VLEO have 2 2 2 2 2 2 1 been announced. MEO orbits are commonly used 0 for Space Data Relay services. GEO orbit will see 1991 1993 1994 1999 2000 2001 2002 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 satellites primarily for In-Orbit Inspection and SSA. Figure 7: Constellation Companies Founded

Kulu 7 35th Annual Small Satellite Conference Funding and Investments tional spacecraft. Peak can be seen between 2017- 2019, which is about 2-3 years after the correspond- Funding amounts, which have been made public, ing founding dates. 2020 saw the first operational or estimated internal investments, have been sum- batch launches for many constellations. Delays have marised on Figure 8. Great quantity of companies been very common and will likely continue to be so. have considerable funding. Capital does not always convert into results or the number of satellites. Many 50 47 other entities have extensive investments based on 40 37 activities, but the exact sums have not been an- 34 nounced. Approximately 17% have likely received 32 30 27 no funding. Announced SPAC rounds will be added 22 once the deals officially close. 20 12 a) Unknown 43 10 8 8 4 4 3 4 b) <$0.1m 3 1 1 2 2 1 1 1 c) $0.1-1m 19 0 d) $1-5m 18 2002 2004 2008 2010 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 e) $5-10m 17 Dormant Unknown f) $10-50m 42 Cancelled g) $50-100m 10 h) $100m+ 17 Figure 9: Constellation First Launches i) $1000m+ 5 j) Yes, amount ? 77 World Map

0 20 40 60 80 Distribution of the constellation companies by headquarters locations is on Figure 10. The most Figure 8: Constellations Funding notable case, where this does not match the manu- facturing location, is Spire, whose satellites are as- First Launches sembled and tested in Glasgow, UK. Generally, the US is far ahead with 40% of constellations being lo- First satellite launches of constellation compa- cated there, followed by China with 9%, UK with 7% nies are depicted on Figure 9. First launch is de- and Canada with 5%. While Russia and India with fined as the direct prototype or the first operational domestic launch capability do not stand out among satellite of a constellation. In many cases, the form Spain, Italy and Australia for example. factor, payload, performance or even application of a pathfinder can be very different from the opera-

Figure 10: Constellations Headquarters Map

Kulu 8 35th Annual Small Satellite Conference 2021 SURVEY BY APPLICATIONS and Sen will be launching their first demonstrators in 2021. The rest have not started regular launches This section presents commercial satellite con- and most of them will likely not continue. stellations by selected applications in no specific or- BlackSky revenues were $22 million in 2020.94 der. Some trends discussions have been included. Planet has not announced funding rounds since early Each figure for an application includes the names 2019 and perhaps they are currently the only eco- of companies, status, launched and planned number nomically sustainable satellite constellation company. of satellites. Prototypes are added to the launched Comparing to late 2017, Planet had launched 180 count except in cases of established space compa- spacecraft including a single batch of 96 in early nies. Not all are independent satellites, some are 2017. Even 4 years later, no other CubeSat con- hosted payloads. Launch failures are included in the stellation is anywhere near by performance and only launched count, because the spacecraft were built Starlink has passed in size. BlackSky had launched and contracted to launch. The previously listed as- their first demonstration mission, while Skybox had pects are proof of activity and gives better insight. already launched 8 by the end of 2017. Astro Digi- tal had two 6U in orbit since 2014 and further pro- Application: Optical EO (VNIR) totypes were launched for promising constellation plans, which now seem to have been cancelled. Figure 11 gathers Earth Observation constella- Latest trend is very-high resolution satellites pur- tions in the visible and near-infrared (VNIR) range sued for example by Albedo Space91 and Earth Ob- and multispectral. Planet’s Dove (SuperDove) and servant,90 both aiming for approximately 10-cm per Skybox constellations together with BlackSky, Satel- pixel resolution, comparable to the capabilities of logic, Axelspace and Jilin-1 are being actively used best military spacecraft. and expanded. SatRevolution has multiple launches booked for 2021 and many others e.g. Aerospacelab

SatRevolution (REC) 1 / 1024 Prototype(s) launched Efir / Sfera 0 / 640 Unknown EarthNow 0 / 500 Cancelled Spacety 18 / 480 Prototype(s) launched Planet (Flock/ Dove/SuperDove) 475 / 150 Launched & replenishing Space JLTZ (Gemini) 0 / 200 Prototype development ADASPACE 4 / 192 Prototype(s) launched Chang Guang (Jilin-1) 24 / 138 Launches ongoing ZeroG Lab (Lingque, Magpie) 2 / 132 Prototype(s) launched Canon 1 / 100 Prototype(s) launched Satellogic 18 / 90 Launches ongoing Xoterra Aerospace 0 / 88 Cancelled Hera Systems 1 / 50 Prototype(s) launched Axelspace (GRUS) 9 / 50 Launches ongoing Astro Digital (Landmapper) 6 / 25 Cancelled?,Prototype(s) launched UrTheCast (UrtheDaily / OptiSar) 4 / 24 Cancelled,Bankrupt Planet (Terra Bella / Skybox) 15 / 24 Launches ongoing OHB Italia (EAGLET) 1 / 20 Prototype(s) launched SpaceWill 5 / 16 Launches ongoing BlackSky 10 / 16 (60) Launches ongoing Earth-i (Vivid-i) 1 / 15 Prototype(s) launched Karten Space 0 / 14 Unknown Tekever 0 / 12 Early stage Planetary Resources 3 / 10 Cancelled Dauria Aerospace 2 / 8 Cancelled Maxar (WorldView Legion) 0 / 6 Prototype development Planet (RapidEye) 5 / 5 Retired Airbus (Pleiades Neo) 1 / 4 Launches ongoing VEOWARE 0 / ? Cancelled? SpaceTime 0 / ? Cancelled?,Prototype development Space Systems Engineering 0 / ? Dormant?,Unknown Sen (EarthTV) 1 / ? Prototype development Scanway 0 / ? Prototype development SatSure 0 / ? Prototype development SATLANTIS 0 / ? Prototype development Rupercorp 0 / ? Cancelled Promethee 0 / ? Prototype development Planet (Pelican) 0 / ? Prototype development Live Earth 0 / 3 Unknown ExoTerra 0 / ? Early stage Alba Orbital 2 / ? Prototype development Planned Aerospacelab (VHR) 0 / ? Prototype development Launched

0 200 400 600 800 1000 1200

Figure 11: Optical (VNIR) Launched and Planned Spacecraft

Kulu 9 35th Annual Small Satellite Conference Application: Infrared EO lion in total and offering imagery with 30 m resolu- tion. They have launched 18 satellites with the first Figure 12 shows Earth Observation constellations constellation spacecraft reaching orbit in 2016 and in the infrared (IR) and thermal range. Here exclud- plans for 90 satellites in total.95 ing Near-Infrared (NIR), but including Mid-Wave Zhuhai-1 has 8 spacecraft in orbit focused on Infrared (MWIR) and Long-Wave Infrared (LWIR). capturing hyperspectral imagery launched in 2018- Planetary Resources (ConsenSys Space) had 2 2019.96 spacecraft in orbit by the end of 2017 and was plan- Pixxel, Orbital Sidekick and Wyvern are some of ning a MWIR Earth Observation constellation until the upcoming constellations entering the same mar- bankruptcy and shutdown in 2018. ket. Scanway, Scanworld and KP Labs should also Aistech Space has launched CubeSat prototypes launch their first satellites in the next 1-2 years. for ADS-B tracking, but none yet for thermal im- Comparing to the end of 2017, Satellogic had agery, after seemingly pivoting the constellation idea launched 3 CubeSats as technology demonstrators to high-resolution IR data, and are now planning a and Orbital Sidekick was listed in the database. crowdfunding campaign to accelerate development. OroraTech, ConstelIR and 4pi Lab are aiming to start launching constellations for forest fire detection in 2021-2022. Albedo Space spacecraft will also have a thermal imager in the LWIR range to capture 2-m resolution photos.91

Application: Hyperspectral EO Figure 13 plots Earth Observation constellations with hyperspectral imagers. Satellogic was founded in 2010 and is one of the most accomplished com- pany in the hyperspectral EO after raising $123 mil-

Aistech (DANU, HYDRA) 3 / 150 Prototype(s) launched OroraTech (Walker-Star) 0 / 100 Prototype development ConstelIR 0 / 30 Prototype development Albedo Space 0 / 24 Prototype development QianSheng 1 / 20 Prototype(s) launched EarthCube 0 / 10 Cancelled Planetary Resources 3 / 10 Cancelled Scanway 0 / ? Prototype development SatelliteVu 0 / ? Prototype development SaraniaSat 0 / ? Prototype development Live Earth 0 / 3 Unknown Koolock 0 / ? Dormant? ISISpace (AMS) 0 / ? Prototype development Hydrosat 0 / ? Prototype development Planned 4pi Lab (EPiC) 0 / ? Early stage Launched

0 20 40 60 80 100 120 140 160 180 200

Figure 12: Infrared EO Launched and Planned Spacecraft

Satellogic 18 / 90 Launches ongoing Reaktor Space Lab 1 / 36 Cancelled?,Prototype(s) launched Pixxel 0 / 24 Prototype development QianSheng 1 / 20 Prototype(s) launched Zhuhai Orbita (Zhuhai-1) 12 / 10 Launches ongoing Scanworld (SpaceBel) 0 / 10 Prototype development Planetary Resources 3 / 10 Cancelled Orbital Sidekick (GHOst) 1 / 6 Prototype(s) launched HyperSat 0 / 6 Prototype development Wyvern 0 / ? Prototype development Spacefab.US 0 / ? Prototype development Scanway 0 / ? Prototype development SaraniaSat 0 / ? Prototype development Kawa Space 0 / ? Cancelled? KP Labs 0 / ? Prototype development Planned HyperCubes 0 / ? Cancelled Launched

0 20 40 60 80 100 120

Figure 13: Hyperspectral EO Launched and Planned Spacecraft

Kulu 10 35th Annual Small Satellite Conference Application: SAR their first prototypes to orbit. In the last months, Tomorrow.io and Climavision97 have raised large rounds Figure 14 illustrates SAR (Synthetic Aperture and announced radar satellite constellations. Radar) constellations. ICEYE was one of the first to enter this market after being founded in 2014 and has launched 10 SAR payloads since 2018. Application: ADS-B Compared to the end of 2017, no commercial Figure 16 illustrates ADS-B constellations of which SAR satellites had been launched, but in addition to Spire and Aireon (hosted payloads on Iridium-NEXT) ICEYE intentions had been announced by Capella are currently offering service. Aerial & Maritime was Space, Umbra and XpressSAR. a partial spin-off of GomSpace, which raised $12.2 Years since have seen many new companies en- million, but never launched satellites and was shut tering this field. Capella has launched 5 spacecraft. down in 2020 after failing to raise more funding.98 During the same time iQPS, Synspective, Spacety, QianSheng have launched first prototypes. Recently Skycraft (Foundation) 0 / 210 Prototype development PredaSAR, Alpha Insights, Orbital Effects, MDA, Spire (Lemur / Minas) 143 / 150 Launches ongoing EOS and many others have announced plans. Aistech (DANU, HYDRA) 3 / 150 Prototype(s) launched Aerial Maritime 0 / 100 Cancelled,Prototype development Aireon 75 / 75 Launched Application: Weather HEAD Aerospace (Skywalker) 5 / 48 Prototype(s) launched Figure 15 shows constellations generating data Tekever 0 / 12 Early stage uniquely suited for weather prediction models. Startical 0 / ? Prototype development Spire has been the most active with over 140 0 100 200 300 400 500 600 Lemur-2’s launched since 2014 and is offering both Planned Launched GNSS-R and GNSS-RO data. GeoOptics has launched Figure 16: ADS-B Launched and Planned 7 satellites for GNSS-RO measurements. Spacecraft In the last year, PlanetiQ and Care Weather sent

Spacety (SAR) 1 / 56 Prototype(s) launched Trident Space 0 / 48 Unknown PredaSAR 0 / 48 Prototype development Capella Space (Whitney) 5 / 40 Launches ongoing iQPS 2 / 36 Prototype(s) launched Synspective (StriX) 1 / 30 Prototype(s) launched UrTheCast (UrtheDaily / OptiSar) 4 / 24 Cancelled,Bankrupt QianSheng 1 / 20 Prototype(s) launched ICEYE 10 / 18 Launches ongoing Umbra 0 / 12 Prototype development EOS (SAT / SAR) 0 / 12 Prototype development Orbital Effects (XR) 0 / 8 Early stage Channel Logistics (Space Eyes) 0 / 5 Early stage XpressSAR 0 / 4 Early stage Ursa Space Systems 0 / ? Cancelled Smart Satellite 1 / ? Prototype(s) launched NARA 0 / ? Cancelled,Concept Mino Space 0 / ? Prototype development MDA (SARnext) 0 / ? Prototype development D-Orbit (NOCTUA) 0 / ? Prototype development Planned Alpha Insights (SAR-XL) 0 / ? Prototype development Launched

0 10 20 30 40 50 60 70

Figure 14: SAR Launched and Planned Spacecraft

Spire (Lemur / Minas) 143 / 150 Launches ongoing OroraTech (Walker-Star) 0 / 100 Prototype development Climavision 0 / 50 Prototype development Care Weather 1 / 50 Prototype(s) launched Orbital Micro Systems 1 / 40 Prototype(s) launched Tomorrow.io 0 / 36 Prototype development PlanetiQ (GNOMES) 1 / 20 Prototype(s) launched Harris Corporation 1 / 12 Prototype(s) launched GeoOptics (CICERO) 7 / ? Launches ongoing Planned 406 Aerospace 0 / ? Cancelled Launched

0 20 40 60 80 100 120 140 160 180 200

Figure 15: Weather Launched and Planned Spacecraft

Kulu 11 35th Annual Small Satellite Conference Application: AIS data relay constellations. Airbus SpaceDataHigh- way GEO-based satellites have been omitted because Figure 17 displays AIS constellations, which was there are only two spacecraft and alternatively NASA one of the first commercial constellation applications. TDRS is not commercial. Many student CubeSat teams have performed simi- lar missions. Spire entered the market in 2014 and has the largest constellation. ExactEarth has 58 Application: VDES hosted AIS receiver payloads on Iridium-NEXT. Figure 19 gathers VDES constellations, which is Spire revenue in 2020 was $28 million99 and AIS a nascent field and where it will take time for the contribution to that was reportedly comparable to VDES transmitters to be rolled out, but currently Orbcomm and exactEarth revenues.100 Saab, Sternula and HEAD plan to be ready. It is un- Many countries have dedicated satellites to re- clear whether existing HEAD satellites have VDES ceive AIS messages, for example Norway has a se- payload onboard. ries of AISSat and NorSat and Luxembourg recently launched ESAIL with the same mission. Saab 0 / 100 Prototype development Sternula (MARIOT) 0 / 50 Prototype development HEAD Aerospace (Skywalker) 5 / 48 Prototype(s) launched Application: Orbital Data Relay 0 50 100 150 200 250 Figure 18 displays constellations that provide data Planned Launched communications service to other satellites or assets in space. Analytical Space, Kepler and Warpspace Figure 19: VDES Launched and Planned are some of the known examples, with the first one Spacecraft possibly having demonstrated service in orbit. This application will include future lunar space

Spire (Lemur / Minas) 143 / 150 Launches ongoing Aistech (DANU, HYDRA) 3 / 150 Prototype(s) launched Aerial Maritime 0 / 100 Cancelled,Prototype development Commsat (Ladybug / Ladybird) 8 / 72 Cancelled?,Prototype(s) launched ExactEarth 68 / 67 Launched Orbcomm (OG2) 50 / 52 Launched & replenishing HEAD Aerospace (Skywalker) 5 / 48 Prototype(s) launched Kleos Space 4 / 40 Launches ongoing OHB Italia (EAGLET) 1 / 20 Prototype(s) launched ISISpace (TRITON) 0 / 16 Cancelled Kineis 0 / 15 Prototype development Karten Space 0 / 14 Unknown Tekever 0 / 12 Early stage AprizeSat 12 / 12 Launched SRT Marine 0 / 6 Cancelled? Channel Logistics (Space Eyes) 0 / 5 Early stage SPUTNIX 0 / ? Prototype(s) launched Planned Elecnor Deimos (NEPTUNO) 0 / ? Prototype development Launched

0 20 40 60 80 100 120 140 160 180 200

Figure 17: AIS Launched and Planned Spacecraft

Kepler Communications 15 / 140 (360) Launches ongoing Stara Space 2 / 120 Prototype(s) launched Globalstar (Second-Generation) 84 / 24 Launched Iridium (NEXT) 75 / 75 Launched Analytical Space (Fast Pixel) 2 / 36 Prototype(s) launched EOS SpaceLink (EOSLink) 0 / 4 Prototype development WARPSPACE (WarpHub InterSat) 2 / ? Prototype development Audacy 1 / 3 Cancelled,Bankrupt,Prototype(s) launched Transcelestial 0 / ? Unknown OrbAstro (Guardian) 0 / ? Prototype development Astrogate Labs 0 / ? Prototype development Space Union 0 / ? Early stage Skyloom 0 / ? Early stage Celestial 0 / ? Early stage Planned Argotec (Andromeda) 0 / ? Early stage Launched

0 20 40 60 80 100 120 140 160 180

Figure 18: Orbital Data Relay Launched and Planned Spacecraft

Kulu 12 35th Annual Small Satellite Conference Application: Laser Communications Application: GNSS and PNT Figure 20 gathers constellations uniquely focused Figure 23 represents constellations offering com- on laser communications between Earth and space. mercial positioning, navigation and timing services. Many others are using optical communications for Satelles has hosted payloads on 66 Iridium-NEXT intersatellite-links. EOS Spacelink bought Audacy spacecraft.102 Xona Space Systems raised $1 mil- assets in 2020, but will not have lasers in the first- lion pre-seed in 2020 to develop GNSS/PNT services generation satellites.101 QKD could have been in- with improved accuracy and encryption using satel- cluded here also due to large overlap in technologies. lites based in LEO.103

LaserFleet 2 / 288 Cancelled? Future Navigation 2 / 120 Prototype(s) launched Analytical Space (Fast Pixel) 2 / 36 Prototype(s) launched Shanghai Lizheng Satellite 1 / 90 Prototype(s) launched Laser Light (HALO) 0 / 12 Early stage Satelles 66 / 66 Launched 0 / 4 Prototype development Xona Space Systems 0 / ? Prototype development EOS SpaceLink (EOSLink) Planned WARPSPACE (WarpHub InterSat) 2 / ? Prototype development Trustpoint 0 / ? Early stage Launched 0 / ? Unknown Transcelestial 0 50 100 150 200 250 Skyloom 0 / ? Early stage Golbriak Space 2 / ? Prototype(s) launched Planned Figure 23: GNSS and PNT Launched and 0 / ? Prototype development Astrogate Labs Launched Planned Spacecraft 0 200 400 600

Figure 20: Optical Communications Application: Constellation-As-A-Service and Launched and Planned Spacecraft Hosted Payloads Figure 24 indicates constellations that host mul- Application: RF Spectrum Monitoring tiple payloads concurrently or are outsourced as a Figure 21 displays constellations detecting radio full end-to-end service. In some cases called Satellite- emissions from objects and likely also performing ge- Data-As-A-Service. olocation. Kleos Space, Unseenlabs, HawkEye 360 Hosted payloads is a concept that emerged early and Aurora Insight all have active satellites in orbit on with CubeSats and microsatellites under the In- and many further launches planned. Orbit Demonstration (IOD) term with a large num- ber of publicly and internally funded projects and UnseenLabs 3 / 50 Prototype(s) launched services by now. Kleos Space 4 / 40 Launches ongoing Iridium-NEXT is the only operational hosted con- HawkEye 360 6 / 21 Launches ongoing stellation with Satelles, exactEarth and Aireon as Aurora Inisght 2 / 12 Prototype(s) launched Ningxia (Ningxia / Zhongi) 5 / ? Prototype(s) launched customers. New actors are Loft Orbital, AAC Clyde LatConnect 60 0 / 3 Prototype development Space, Spire, Satellogic, Modularity Space and many Planned 0 / ? Prototype development Horizon Technologies (Amber) Launched others. Most of them have first customers contracted 0 20 40 60 80 100 120 140 for single or even multiple satellites, but long-term viability of this approach has not been proven. Figure 21: RF Spectrum Monitoring

Launched and Planned Spacecraft Spire (Lemur / Minas) 143 / 150 Launches ongoing Modularity Space 0 / 150 Prototype development Satellogic 18 / 90 Launches ongoing Application: Emissions Monitoring Iridium (NEXT) 75 / 75 Launched NanoAvionics 3 / 72 Prototype development Figure 22 shows companies monitoring greenhouse Loft Orbital (YAM) 0 / 10 Prototype development AAC Clyde Space (xSPANCION) 0 / 10 Prototype development gas emissions. GHGSat now has 3 satellites in orbit. Planned 0 / ? Concept Bluefield for methane has been delayed for years. Alen Space Launched 0 100 200 300 400 Bluefield 0 / 22 Prototype development GHGSat 3 / 10 Prototype(s) launched Figure 24: Constellation-As-A-Service and Scepter 0 / ? Unknown Hosted Payloads Launched and Planned ISISpace (Spectrolite) 0 / ? Prototype development Planned Promethean Labs 0 / ? Cancelled Launched Spacecraft 0 10 20 30 40 50 Figure 22: Emissions Monitoring Launched and Planned Spacecraft

Kulu 13 35th Annual Small Satellite Conference Application: SSA (Space Situational Aware- Application: Broadband Internet ness) Figure 28 shows Broadband Internet constella- Figure 25 plots constellations aiming to detect tions. These include the largest planned constella- and track other satellites and objects in space for tions and the most launched satellites up to now. orbit determination. None have been launched, but None of them cover the globe yet, but SpaceX ex- NorthStar has raised over $43 million and expects pects to achieve uninterrupted service coverage by to launch its first satellites in 2022.104 autumn of 2021 and OneWeb by the end of 2022. SpaceX Starlink launched two test satellites in NorthStar (Skylark) 0 / 52 Prototype development 2018, and recently passed 1600 satellites in 2 years Chandah Space 0 / 20 Early stage,Dormant? HEO Robotics (ARGUS) 0 / 12 Prototype development from the launch of the first batch of 60 in May 2019. Vyoma Space 0 / ? Prototype development OneWeb has continued launches after emerging SpaceAble 0 / ? Unknown from bankruptcy in 2020 and the total number of Space Forms 0 / ? Early stage spacecraft in orbit is now over 200. SCOUT 0 / ? Early stage Launchspace Technologies 0 / ? Prototype development China has had many constellation plans, which Planned 0 / ? Unknown Inovor (Skyris / Hyperion) Launched as of 2021 May seem to have been consolidated into 0 50 100 GuoWang. It is still conceivable that Geely and some others will or aim to deploy their own. Figure 25: SSA (Space Situational Aware- Russia has announced a constellation, but cur- ness) Launched and Planned Spacecraft rent progress is unknown and likely not commercial. European Union is planning its own constellation too, but possibly not using a commercial approach. Application: In-Orbit Inspection Project Kuiper by Amazon is expected to be a Figure 26 shows constellations aiming to fly close major player in the future, but no satellites have to other satellites and observe their status. None been launched yet. Telesat is another very active have been launched, but Rogue Space is running a stakeholder, which has raised billions and contracted crowdfunding campaign and has established partner- Thales Alenia to build its constellation. Lynk has ship with NanoAvionics.105 There is overlap with also performed multiple flight demonstrations. AST SSA as both can use similar technologies. & Science has gone public via SPAC and aims to offer Internet directly to mobile phones, but no con- Rogue Space Systems (Orbot) 0 / 40 Prototype development stellation prototypes have yet been launched. SEOPS (Chameleon) 0 / 36 Prototype development Chandah Space 0 / 20 Early stage,Dormant? Space Forms 0 / ? Early stage Application: IoT / M2M Planned SCOUT 0 / ? Early stage Launched Figure 29 details IoT / M2M constellations and 0 20 40 60 80 100 is the largest table with 61 entries. This was one of Figure 26: In-Orbit Inspection Launched and the first commercial markets pursued by Globalstar, Planned Spacecraft Iridium, Orbcomm and AprizeSat before CubeSats. In recent years and specifically in 2020, many new constellations have launched their first batches, in- Application: QKD (Quantum Key Distribu- cluding Swarm, Kepler and Astrocast, all of them tion) using different form factors of CubeSats. Figure 27 details QKD constellations for unhack- By late 2017, Sky and Space had announced their able communications. Chinese test satellite Micius intention to launch 200 satellites, raised consider- was launched in 2016 and it is still the only one in able funding and was traded publicly until entering 108 orbit.106 Arqit recently announced plans to go pub- bankruptcy protection in 2020. Sky and Space lic via SPAC while raising $400 million, a deal which also had first 3 CubeSats in orbit already in 2017. is expected to close later in 2021.107 This partially started an IoT / M2M constellation boom with Hiber, Kepler, Fleet Space, Astrocast, SpeQtral 0 / ? Prototype development Aistech and dozens more being founded up to now. QEYnet 0 / ? Early stage Planned 3.5 years later Swarm has taken the lead by the num- ArQit 0 / ? Prototype development Launched ber of satellites launched, helped by the small size 0 1 2 3 4 5 6 7 8 of 0.25U CubeSats. Figure 27: QKD Launched and Planned Spacecraft

Kulu 14 35th Annual Small Satellite Conference Guo Wang 0 / 12992 Prototype development SpaceX (Starlink) 1675 / 4408 (41493) Launches ongoing Amazon (Project Kuiper) 0 / 3236 Prototype development Hanwha Systems 0 / 2000 Prototype development Galaxy Space (Yinhe) 1 / 1000 Unknown Mangata Networks 0 / 791 Prototype development OneWeb 216 / 648 (6372) Launches ongoing Efir / Sfera 0 / 640 Unknown Hongyan (CASC) 1 / 320 Prototype(s) launched Telesat (Lightspeed) 2 / 298 Prototype development LaserFleet 2 / 288 Cancelled? LinkSure (Swarm Constellation) 0 / 272 Cancelled? AST&Science (SpaceMobile) 1 / 243 Prototype development CurvaNet 0 / 240 Prototype development Outernet 0 / 200 Cancelled Astrome (SpaceNet) 0 / 198 Unknown Kepler Communications 15 / 140 (360) Launches ongoing LeoSat 0 / 108 Dormant Globalstar (Second-Generation) 84 / 24 Launched NSLComm 1 / 80 Prototype(s) launched Hongyun / Xingyun (Xingyun-2) 4 / 80 Launches ongoing Iridium (NEXT) 75 / 75 Launched Commsat (Ladybug / Ladybird) 8 / 72 Cancelled?,Prototype(s) launched SES (O3b / mPOWER) 20 / 70 Launched Vesta Space Technology 0 / 35 Cancelled?,Prototype development Viasat (MEO Constellation) 0 / 20 Prototype development Methera 0 / 16 Early stage Lynk 4 / 10 (5000) Prototype(s) launched Rocket Lab 0 / ? Prototype development Ragnarok Industries 0 / ? Cancelled ProXops (VariSat) 0 / ? Prototype development PointView (Athena) 1 / ? Prototype(s) launched Geely 0 / ? Prototype development Astranis (MicroGEO) 1 / ? Prototype development Apple 0 / ? Early stage Planned Apogee Networks 0 / ? Early stage Launched

0 2000 4000 6000 8000 10000 12000 14000 16000

Figure 28: Internet Launched and Planned Spacecraft

Kulu 15 35th Annual Small Satellite Conference KLEO 2 / 300 Prototype(s) launched Lacuna Space 5 / 240 Launches ongoing Space JLTZ (Gemini) 0 / 200 Prototype development Sky and Space 3 / 200 Cancelled?,Prototype(s) launched Omnispace 0 / 200 Prototype development Swarm Technologies 93 / 150 (600) Launches ongoing Aistech (DANU, HYDRA) 3 / 150 Prototype(s) launched Kepler Communications 15 / 140 (360) Launches ongoing Sateliot 1 / 100 Prototype(s) launched Innova Space (Libertadores de América) 0 / 100 Early stage Fleet Space 5 / 100 Launches ongoing Shanghai Lizheng Satellite 1 / 90 Prototype(s) launched Globalstar (Second-Generation) 84 / 24 Launched NSLComm 1 / 80 Prototype(s) launched Hongyun / Xingyun (Xingyun-2) 4 / 80 Launches ongoing Astrocast 7 / 80 Launches ongoing Iridium (NEXT) 75 / 75 Launched NanoAvionics 3 / 72 Prototype development Commsat (Ladybug / Ladybird) 8 / 72 Cancelled?,Prototype(s) launched Orbcomm (OG2) 50 / 52 Launched & replenishing Sternula (MARIOT) 0 / 50 Prototype development Myriota 2 / 50 Prototype(s) launched Hiber 4 / 48 Launches ongoing HEAD Aerospace (Skywalker) 5 / 48 Prototype(s) launched Guodian Gaoke (Apocalypse) 10 / 38 Launches ongoing Vesta Space Technology 0 / 35 Cancelled?,Prototype development Helios Wire (EchoStar) 4 / 30 Prototype(s) launched Totum Labs 0 / 24 Prototype development SaudiComsat 7 / 24 Launched Dunvegan Space (BitSat) 0 / 24 Cancelled QianSheng 1 / 20 Prototype(s) launched Kineis 0 / 15 Prototype development Tekever 0 / 12 Early stage AprizeSat 12 / 12 Launched Lynk 4 / 10 (5000) Prototype(s) launched GomSpace 0 / 10 Cancelled German Orbital Systems 4 / ? Prototype(s) launched,Unknown UTOWAVE (SpaceDust) 0 / ? Cancelled? Tyvak 0 / ? Cancelled Telnet 1 / ? Prototype(s) launched Ragnarok Industries 0 / ? Cancelled Promethee 0 / ? Prototype development ProXops (VariSat) 0 / ? Prototype development Orbitare (Spaceloop) 0 / ? Prototype development OQ Technology 0 / ? Prototype development NuSpace 0 / ? Prototype development Ligado (SkyTerra (MSV)) 1 / 3 Launched Gravity Space 0 / ? Prototype development Galaktika (Pulsar) 0 / ? Cancelled GP Advanced Projects (FEES Cluster) 1 / ? Prototype(s) launched Fossa Systems 1 / ? Prototype development Eutelsat (CubeSats) 1 / ? Prototype(s) launched Elbit Systems (NANOVA) 1 / ? Prototype(s) launched Blink Astro 1 / ? Dormant?,Prototype(s) launched Bitlinq Space 0 / ? Unknown BHDynamics 0 / ? Early stage Artemis Space (BEACON) 0 / ? Dormant?,Concept Alya Nanosatellites 0 / ? Early stage AT&T 0 / ? Prototype development Planned 4Skies 0 / ? Cancelled Launched

0 50 100 150 200 250 300 350 400

Figure 29: IoT / M2M Launched and Planned Spacecraft

Kulu 16 35th Annual Small Satellite Conference CONSTELLATION TRENDS Latest trends are RF Spectrum Monitoring with Geolocation, Orbital Data Relays, Laser Communi- Most of the market and technology trend discus- cations and Emissions Monitoring. Very recent and sions have been integrated into the previous sections without prior governmental or commercial satellites under Survey by Applications, Introduction and Sta- are QKD (Quantum Key Distribution), SSA (Space tistical Overview, to avoid repeated thoughts. Hereby Situational Awareness), commercial GNSS / PNT, follows the methodology. VDES, In-orbit Inspection and Constellation-As-A- Service among others. Trends Methodology Future work will foresee studying application spe- cific trends and development timelines in much more Primarily comparing constellations in this sur- detail. Many new applications do not have exist- vey to a database version from the end of 2017, 3.5 ing markets, which will take many years to emerge, years ago, but also discussing developments before and/or the revenue potential seems to be small. and afterwards. Four approaches to discover trends Major open question is the economic sustainability have been used: of most or possibly all of the constellations and this 1. Viewing the database entry dates, because ma- will be researched in a future paper. jority have been discovered from announcements. References 2. Filtering by the founding dates, while acknowl- edging this will only give a partial picture. [1] Seraphim Capital. Announcing Seraphim Smallsat Constellation Market Map. 3. Funding news by dates and application types, to better understand growth and execution. [2] Giacomo Curzi, Dario Modenini, and Paolo 4. Looking at the first launch years, both in the Tortora. Large Constellations of Small Satel- past and upcoming ones, and what progress lites: A Survey of Near Future Challenges and has been made and at which pace. Missions. September 2020.

All of the four methods combined should give [3] Ibrahim Sanad, Zahra Vali, and David G. an acceptable view into historic and present trends, Michelson. Statistical classification of remote because constellation development times are still rel- sensing satellite constellations. atively long and measured in years. [4] Tom Butash, Peter Garland, and Barry Evans. Non-geostationary satellite orbit communica- CONCLUSIONS tions satellite constellations history. 2021.

Statistical overview of 251 commercial satellite [5] O. Kodheli et al. Satellite Communications constellations has been presented. In addition, the in the New Space Era: A Survey and Future most popular applications were shown in detail and Challenges. March 2020. relevant trends were discussed. Only about 4% of the constellations have been [6] Israel Leyva-Mayorga et al. LEO Small- fully launched and approximately 8% are currently Satellite Constellations for 5G and Beyond-5G being launched. Notable amount of additional enti- Communications. 2020. ties have raised significant funding and are in active [7] Lloyd Wood. Lloyd’s Satellite Constellations, development. Approximately 1/3 of the constella- September 2006. tions have already become dormant or cancelled for various technical and non-technical reasons and their [8] John E. Hatlelid and David E. Sterling. A Sur- quantity is expected to increase. vey of Small Spacecraft in Commercial Con- Broadband Internet constellations are the largest stellations. In AIAA/Utah State University planned satellite networks by size and also have the Conference on Small Satellites, August 1991. most spacecraft launched, lead by Starlink. Next in popularity are Optical Earth Observation, SAR, [9] E Lutz, Markus Werner, and Axel Jahn. Satel- Infrared EO and IoT/M2M. The historically proven lite Systems for Personal and Broadband Com- applications of Earth Observation, SAR and com- munications. 2000. munications are favoured thanks to existing value [10] William H. Mott and Robert B. Sheldon. Laser chains. AIS has also been a commercial market since Satellite Communication: The Third Genera- early 2000s and now followed by ADS-B. tion. Quorum, Westport, Conn, 2000.

Kulu 17 35th Annual Small Satellite Conference [11] Commercial Space Transport Study Final Re- [30] Peter B. de Selding. WorldVu Seeks Builder port. Technical report, May 1994. for Smallsat Constellation, Wants To Co-own Production Facility. November 2014. [12] Bob Tripp. EarlyBird Satellite Expected to Sharpen Focus of Commercial Remote Sensing [31] Rolfe Winkler and Andy Pasztor. Elon Musk’s Industry. October 1995. Next Mission: Internet Satellites. 2014. [13] DigitalGlobe. Commercial Remote Sensing: [32] Mike Carrozzo. OneWeb Announces Plans An Historical Chronology, September 2010. to Launch a New Satellite Constellation to Bring High-Speed Internet to Underserved Ar- [14] Lawrence W. Fritz. The Era of Commercial eas Around the World. January 2015. Earth Observation Satellites. January 1996. [33] Peter B. de Selding. SpaceX To Build 4,000 [15] ESA. 50 years of Earth Observation. Broadband Satellites in Seattle. January 2015. [16] ImageSat International. Company History. [34] Jeff Foust. OneWeb emerges from Chapter 11 [17] Warren Ferster. DigitalGlobe Closes GeoEye with new CEO. November 2020. Acquisition. January 2013. [35] C. Daehnick et al. Large LEO satellite constel- [18] RapidEye and MDA Enter Partnership Agree- lations: Will it be different this time?, 2020. ment, October 2002. [36] Peter B. de Selding. Signs of a Satellite Inter- [19] Peter B. de Selding. Canadian Earth Obser- net Gold Rush in Burst of ITU Filings. 2015. vation Firm Buys Bankrupt RapidEye. 2011. [37] Lloyd Wood. Satellite Constellation Networks. [20] Ryan Lawler. Planet Labs Nabs $95 Million In Internetworking and Computing Over Satel- And A New COO To Cover The Earth With lite Networks, pages 13–34. 2003. Flocks Of Tiny Satellites. January 2015. [38] Pauline C. M. Jakob et al. Optimal Satel- [21] Jessie Scanlon. There’s more to the small Sky- lite Constellation Spare Strategy Using Multi- box satellite than meets the eye. Much, much Echelon Inventory Control. September 2019. more, February 2014. [39] Joseph Howard, Dipak Oza, and Danford S. [22] Matt Burns. Skybox Imaging Raises $70M Smith. Best Practices for Operations of Satel- To Launch Two High-Res Imaging Microsatel- lite Constellations. 2006. lites. April 2012. [40] Klaus Brieß. Handbook of Space Technology. [23] Skybox Imaging Selects SSL to Build 13 Low Wiley, Chichester, U.K, 2009. Earth Orbit Imaging Satellites. October 2014. [41] Cubesat Handbook. Elsevier, 2021. [24] Will Marshall. Planet To Acquire Terra Bella From Google, Sign Multi-Year Data Contract, [42] Susan L. Ustin and Elizabeth M. Middleton. February 2017. Current and near-term advances in Earth ob- servation for ecological applications. Ecological [25] Richard Barnett. O3b – A different approach Processes, 10(1):1, December 2021. to Ka-band satellite system design and spec- trum sharing, September 2012. [43] Tom George et al. Comparison of Mid Wave Infrared (MWIR) and Long Wave Infrared [26] ORBCOMM Announces Commercial Service (LWIR) Imagery for Precision Agriculture Ap- for Its Final 11 OG2 Satellites, March 2016. plications. March 2019.

[27] Iridium Declares Victory; $3 Billion Satellite [44] Julie Transon et al. Survey of Hyperspectral Constellation Upgrade Complete, 2019. Earth Observation Applications from Space in [28] Tai Nichols. Globalstar’s Second Generation the Sentinel-2 Context. January 2018. Satellites in Full Service, August 2013. [45] John H Steele, S. A Thorpe, and Karl K [29] Peter B. de Selding. Google-backed Global Turekian. Encyclopedia of Ocean Sciences. El- Broadband Venture Secures Spectrum for sevier, London; Burlington, MA, 2009. Satellite Network. May 2014.

Kulu 18 35th Annual Small Satellite Conference [46] Africa Ixmucane Flores-Anderson. Synthetic [59] W. Holmes. NASA’s tracking and data relay Aperture Radar (SAR) Handbook: Compre- satellite system. September 1978. hensive Methodologies for Forest Monitoring and Biomass Estimation. 2019. [60] T. Scheidl, J. Handsteiner, D. Rauch, and R. Ursin. Space-to-ground quantum key dis- [47] Veronica Magan. EarthNow Aims to Deliver tribution. SPIE, July 2019. Real-Time Video of Earth, April 2018. [61] Sheng-Kai Liao et al. Satellite-to-ground [48] Dylan Jervis et al. The GHGSat-D imag- quantum key distribution. 2017. ing spectrometer. Atmospheric Measurement Techniques, March 2021. [62] ICAO. Study on the convenience and feasibil- ity of space-based ADS-B for regional imple- [49] Daniel H. Cusworth et al. Potential of mentation. Technical report, 2018. next-generation imaging spectrometers to de- tect and quantify methane point sources from [63] Karl Baker. Space-Based ADS-B: Perfor- space. Technical report, May 2019. mance, Architecture and Market. April 2019. [50] Inigo del Portillo et al. A technical comparison [64] Torkild Eriksen et al. Tracking ship traffic with of three low earth orbit satellite constellation Space-Based AIS: Experience gained in first systems to provide global broadband. Acta As- months of operations. November 2010. tronautica, June 2019. [65] Francisco L´azaroet al. VHF Data Exchange [51] Ian F. Akyildiz and Ahan Kak. The Inter- System (VDES): An enabling technology for net of Space Things/CubeSats: A ubiquitous maritime communications. March 2019. cyber-physical system for the connected world. [66] Hannah R. Goldberg et al. Asteroids to Agri- February 2019. culture: Carving a Niche in Earth Observation [52] Vicente Almonacid and Laurent Franck. Ex- Using Asteroid Prospecting Instruments on an tending the coverage of the internet of things Earth-Orbiting CubeSat Constellation. 2016. with low-cost nanosatellite networks. Acta As- [67] David C. Woffinden. On-Orbit Satellite Inspec- tronautica, 138:95–101, September 2017. tion : Navigation and [Delta]v Analysis. PhD [53] Jeffrey Ward. Store-and-Forward Message Re- thesis, MIT, 2004. lay using Microsatellites: The UoSAT-3 PAC- [68] Tobias Pfandzelter, Jonathan Hasenburg, and SAT Communications Payload. In 4th An- David Bermbach. Towards a Computing Plat- nual AIAA/USU Conference on Small Satel- form for the LEO Edge. April 2021. lites, 1990. [69] Debopam Bhattacherjee et al. In-orbit Com- [54] Xingqin Lin, Stefan Rommer, Sebastian Eu- puting: An Outlandish thought Experiment? ler, Emre A. Yavuz, and Robert S. Karlsson. 5G from Space: An Overview of 3GPP Non- [70] Bradley Denby and Brandon Lucia. Or- Terrestrial Networks. March 2021. bital Edge Computing: Nanosatellite Constel- lations. March 2020. [55] Romain Barbau et al. NB-IoT over GEO Satel- lite: Performance Analysis. October 2020. [71] S. Briatore, N. Garzaniti, and A. Golkar. To- wards the Internet for Space: Bringing cloud [56] Tyghe Speidel et al. Results from Rapid Test- computing to space systems. 2018. ing of Space-based Mobile Network Technol- ogy. 2020. [72] Joseph N. Pelton and Scott Madry. Innova- tions in Hosted Payload Satellite Services. In [57] Mark Dmytryszyn, Matthew Crook, and Tim- Handbook of Satellite Applications. 2017. othy Sands. Lasers for Satellite Uplinks and Downlinks. January 2021. [73] Debra Werner. What happened to the promise of hosted payloads? It’s complicated. Space- [58] Stefan Seel. Space to Ground bidirectional News, August 2019. optical communication link at 5.6 Gbps and EDRS connectivity outlook. 2011. [74] Jade Y. T Morton et al. Position, Navigation, and Timing Technologies in the 21st Century.

Kulu 19 35th Annual Small Satellite Conference [75] Tyler G. R. Reid et al. Leveraging Broadband [90] Eric Tegler. How Low Can Satellites Go? Air LEO Constellations for Navigation. 2016. Force Bets Very Low Earth Orbit Will Give It More Capabilities. Forbes, August 2020. [76] Xona Space Systems. Commercial LEO PNT for Autonomy, October 2020. [91] Albedo Wraps Up Y Combinator with a $10M Seed Round, April 2021. [77] Troy Farah. This Russian startup wants to put billboards in space. Astronomers aren’t [92] SpaceX V-band Non-Geostationary Satellite impressed. January 2019. System - Attachment A. Technical report. [78] K. Sarda et al. Making the Invisible Visible: [93] N. H. Crisp et al. The Benefits of Very Low Precision RF-Emitter Geolocation from Space Earth Orbit for Earth Observation Missions. by the HawkEye 360 Pathfinder Mission. 2018. [94] Rachel Jewett. BlackSky Looks to Go Public [79] Jos´eDarrozes et al. The Reflected Global Nav- in SPAC Merger. February 2021. igation Satellite System (GNSS-R): From The- [95] Jeff Foust. Satellogic raises $50 million to build ory to Practice. 2016. out imaging constellation. December 2019. [80] William J. Emery and Adriano Camps. In- [96] G. Krebs (Gunter’s Space Page). Zhuhai-1. troduction to Satellite Remote Sensing: Atmo- sphere, Ocean, Land and Cryosphere Applica- [97] Jason Rainbow. Climavision emerges from tions. 2017. stealth mode with $100 million for weather in- telligence network. June 2021. [81] Corey Bergsrud and Jeremy Straub. A space- to-space microwave wireless power transmis- [98] Space in Africa. Aerial & Maritime Shuts sion experiential mission using small satellites. Down By Investor. June 2020. October 2014. [99] Rachel Jewett. Spire to Go Public through $ [82] Robert Scott and Stefan Thorsteinson. Key SPAC at 1.6B Valuation. March 2021. findings from the NEOSSat space-based SSA [100] Peter B. de Selding. ExactEarth, Orbcomm & microsatellite mission. Technical report, 2018. Spire posted similar maritime AIS revenue in [83] Mark R. Ackermann et al. A Systematic Ex- 2020. Now the fun starts. March 2021. amination of Ground-Based and Space-Based [101] Debra Werner. EOS establishes SpaceLink Approaches to Optical Detection and Tracking with Dave Bettinger at the helm. 2020. of Satellites. In 31st Space Symposium, 2015. [102] Gregory Gutt and Michael O’ Connor. Satelles [84] V. Lappas et al. NACON - a nano-satellite - Trusted Time and Location, Everywhere, constellation for space weather monitoring. May 2016. January 2005. [103] Darrell Etherington. Xona Space Systems [85] Debra Werner. Are small satellites the solution raises $1 million to improve satellite-based for space weather monitoring? March 2019. navigation services. May 2020. [86] Paulett C. Liewer et al. A Fractionated Space [104] Isabelle Kirkwood. Montreal Startups North- Weather Base at L5 using CubeSats and Solar Star Begins Construction of Satellites to Mon- Sails. 2014. itor Debris in Space. October 2020. [87] Publication produced for FAA AST by Bryce [105] TruCrowd - Rogue Space Systems. Space and Technology. The Annual Com- [106] Karen Kwon. China Reaches New Milestone pendium of Commercial Space Transporta- in Space-Based Quantum Communications. tion: 2018. Technical report, 2018. [107] Jason Rainbow. Arqit raising $400 million [88] S. Radu et al. The PocketQube Standard. with a SPAC to launch quantum encryption June 2018. satellites in 2023. May 2021. [89] The CubeSat Program, Cal Poly SLO. Cube- [108] Jeff Foust. Virgin Orbit to take stake in Sky Sat Design Specification Rev. 14. 2020. and Space Global. October 2020.

Kulu 20 35th Annual Small Satellite Conference