knowledge centre

Low Earth orbit satellite broadband makes a comeback

Philip Bates Principal, Consulting

New entrants such as earth-observation companies Skybox Imaging and Planet Labs have shown that LEO satellites can be built at low cost.

It's 1994, the year the Channel Tunnel opens, Nelson Mandela is elected president and O. J. Simpson is arrested. A company called Teledesic, set up by cellular pioneer Craig McCaw and Bill Gates, files a plan with the FCC for the 'Internet-in-the-sky', a USD9 billion project to build and launch 840 Ka-band satellites into low Earth orbit (LEO) to provide global data connectivity. A demonstration satellite was launched in 1998 but the commercial failure of other LEO projects including Iridium and Globalstar meant that Teledesic was ultimately unfinanceable (as were other similar projects such as Celstri). Construction of Teledesic's satellites was officially suspended in October 2002.

Fast-forward to 2015, and the hot topic of conversation at the Satellite 2015 conference and trade show in Washington, DC was a series of rival proposals from OneWeb, SpaceX and LeoSat for new LEO constellations designed to provide global broadband connectivity (see Figure 1). What has changed in the industry to revive the concept of delivering broadband via LEO satellites and what are the key challenges that these new projects need to address?

Figure 1: Overview of selected LEO broadband projects [Source: Analysys Mason, 2015]

  OneWeb SpaceX   LeoSat
Champion  Greg Wyler, ex-CEO of O3b Networks (pioneer of MEO satcomms)  Elon Musk, founder of SpaceX  Vern Fotheringham, ex-CEO of Kymeta (solid-state satellite antenna company)
Partners and backers  Virgin Group
Qualcomm 
Google  None announced 
Number of satellites  650  4000  80–140 
Proposed operating band Ku band  Ku or Ka band  Ka band 
Satellite manufacturer  To be selected from Airbus, Lockheed Martin, OHB, SSL and Thales Alenia Space  SpaceX (in new factory)  Not yet announced 
Other details  Aiming to procure satellites for USD400 000 each; mass-market fixed terminals for USD250. Also planning specialised terminals for aeronautical and first responder markets  Appears to be aiming at mass market. Press reports indicate system cost of USD10 billion–USD15 billion  Aiming at high end users, 'top 3000 rather than other 3 billion'. Will use high-speed inter-satellite links to provide fixed point-to-point connections at up to 1.2Gbit/s. System cost stated to be USD2.5 billion–USD3 billion 

Firstly, it's worth pointing out that while the original investors in Iridium and Globalstar lost their money, these two constellations did actually get built and have been operating effectively for 15 years. Both companies have been able to raise money for second-generation constellations (Globalstar's was launched last year and Iridium should start launching its NEXT satellites in October). Consequently, there are now commercial precedents for LEO communication services.

Secondly, the broadband market is vastly more developed than it was in Teledesic's day. Back in 1994, less than 0.5% of the world's population was online and most connections were narrowband. By 2002, the year that Teledesic shut up shop, penetration exceeded 10% but even in the USA narrowband connections outnumbered broadband by three to one. Today over 40% of the world's population is online and almost all users enjoy some form of broadband.2

Thirdly, new entrants such as earth-observation companies Skybox Imaging and Planet Labs have shown that LEO satellites can be built at low cost: Skybox (bought by Google in 2014 for USD500 million) developed its dishwasher-sized satellite and built the prototype in house, although manufacture was subsequently outsourced to SSL. Planet Labs, whose system uses larger numbers of very small satellites (offering lower resolution but more frequent updates) built all of its satellites in house. Launch costs are also coming down, largely as a result of competition from new players such as SpaceX.

Finally, there have been some advances in the design of antennas for LEO satellite terminals, which need the ability to track a satellite as it moves across the sky and transfer seamlessly to a new one when the first satellite approaches the horizon. Most commentators agree that solid-state antennas will be essential to meet the cost and reliability needs of the new systems. Two companies, Kymeta and Phasor, were exhibiting prototype low-cost solid-state antennas at Satellite 2015, though at present both companies need to deploy separate antenna units for transmit and receive, and manufacturing costs are in the thousands of dollars rather than the low hundreds.

The cost of user terminals therefore remains a major challenge to the viability of LEO broadband. Similarly, it remains to be seen whether the satellites can really be built for a few hundred thousand dollars apiece. Co-ordinating the spectrum for global constellations of LEO satellites with the operators of existing geostationary satellites in the same band is also a complex undertaking. Lastly, the key to running a profitable telecommunications satellite business is the fill factor: all of these projects will therefore need to set up effective global distribution targeting customers on land, at sea and in the air. Such challenges should not be underestimated, but then neither should the drive and tenacity of the new LEO broadband pioneers.

Analysys Mason works extensively with satellite operators and service providers and their investors on market forecasting, business planning and transaction support projects. For further information please contact Philip Bates.


1 Ku-band systems are likely to use spectrum at around 12GHz for downlink and 14GHz for uplink while Ka-band systems are likely to use spectrum at around 20GHz for downlink and 30GHz for uplink

2 Source: Internetlivestats.com and Internetworldstats.com