PARIS — Satellite fleet operator SES on Jan. 22 said it would jump into the market for offering global broadband communications from a low-orbiting constellation if the latest efforts announced by SpaceX, Google, Virgin Group and Qualcomm — with others to come — show signs of proving themselves.
Luxembourg-based SES, which believes it surpassed rival Intelsat last year as the world’s largest fixed-satellite services operator by revenue, said that as the only big satellite operator with assets in both geostationary and medium Earth orbit, it is willing to accept that these well-financed companies may be on to something.
SES is 47 percent owner of O3b Networks, a company in which Google has about a 5 percent stake. O3b operates 12 satellites, with more to come, orbiting at 8,000 kilometers in altitude and focused on a global belt of customers located between 40-plus degrees north and south of the equator.
SES Chief Executive Karim Michel Sabbagh said Google has prevented a dilution of its O3b stake by taking part in each O3b capital raise. He characterized Google as a “very active, involved” O3b shareholder while reiterating that SES expects to become the majority owner by 2017.
O3b Chief Executive Steve Collar said the company, which began full commercial service only in September, has sold 20 gigabits per second of capacity to 36 customers, half of them already on the network, with the rest to be switched on by spring. The current O3b network of Ka-band satellites has 100 gigabits per second of throughput.
Collar said O3b is likely to reach $100 million in revenue in 2015 and would order more satellites, likely in groups of four or eight, by midyear. With each new satellite tranche, O3b adds throughput, and the satellites to be ordered in 2015 will be higher-powered than the 12 now operating.
Collar said O3b expects to be offering 1 terabit per second of throughput within three or four years.
In a press briefing here, Collar, Sabbagh and SES Chief Technology Officer Martin Halliwell outlined the company’s views on the several high-profile investors now turning toward satellites to provide global Internet coverage.
Sabbagh said many questions remain about the viability of what SpaceX, Google, Virgin, Qualcomm and the others are proposing. But he said the mere fact that high-profile companies are promoting satellites to boost connectivity is a boost for the industry.
“Look, five years ago we would not have believed in many of the new markets that O3b is now serving,” Sabbagh said. “If there is going to be a business there [where SpaceX, Google and other low-orbit global Internet initiatives propose to operate] then we are going to be active in the process. But we’re not there yet.”
Sabbagh said it is not just Google and SpaceX and OneWeb —a company founded by the same people that founded O3b — that are eyeing low-orbiting constellations for global connectivity.
“I can assure you that in my travels I hear similar concepts from people with similar [investment and engineering] capacity, and from nations with similar capacity, that are looking at this”, Sabbagh said. “What is new is that it’s now on page one of the media.”
While expressing optimism for their potential, Sabbagh said the new projects need to answer the same questions that, some 15 years ago, could not be answered by companies including Teledesic and SkyBridge, which proposed similar constellations to those now making news.
“The four questions we asked of O3b, and which need to be asked of the current low-orbit proposals, are whether the technical, operational, regulatory and economic/financial models all stand up. So far, all we have heard discussed are the technological issue: Can I produce 800 or a few thousand satellites? If so, what is the best industrial model? But that’s not the whole story.”
Collar said O3b’s orbit of 8,000 kilometers was selected after a careful look at the alternatives.
“From 8,000 kilometers you can see about 50 degrees” north and south of the equator, Collar said. Others have said O3b is limited to more like 40 degrees either way.
“But as you get closer to the Earth and stay equatorial, your visibility reduces dramatically. That forces you into a much broader constellation with inclined planes and polar orbits. That adds huge complexity to the network. From our orbit, every one of our satellites is providing service permanently, with the exception of one, which is always performing [signal] handovers.
“In [low-Earth orbit], your efficiency drops from, say, 95 percent to maybe around 20 percent. So more and more of your satellites are spending more time over the poles, and others are spending more time handing over the signal from one satellite to the next. This has dramatic consequences for the number of satellites that you need.”
Halliwell, who cautioned that he had not made a detailed study of the SpaceX/Google or OneWeb designs, said the low-orbit solution is “a very difficult one. First, if you are looking at 650 satellites in operation, then to support that you are going to have to construct about 900 satellites to include in-orbit spares and spares in the event of a launch failure.”
Halliwell said connecting with the ground terminals will be challenging too, because the terminal sophistication will mean a high cost unless production volume is huge.
“Imagine this,” Halliwell said, describing one possible low Earth orbit constellation design. “The antenna is 30 centimeters in diameter and it’s fixed and it has a field of view of about 15 degrees looking at the sky. The satellite passing over has a footprint, like a plate, and it has many different sub-beams inside its footprint. Each sub-beam footprint is in view of the terminal for only 45 seconds. So you have to change frequency and polarization every 45 seconds — bam, bam, bam — as the satellites pass over.
“And that’s just the start of the fun and games,” Halliwell said. “Getting them into orbit, into a multi-plane configuration, is not so simple. Let’s say it takes 18 months to get them all into operations. Let’s say each satellite costs just $500,000, which means a $450 million investment for 900 satellites. And in five years or so you have to start buying your second generation.”