WASHINGTON — The world’s biggest, best established satellite operators talk of broadband as an enormously lucrative opportunity.

But in truth, nothing is causing them more frustration.

Demand for ever-faster broadband internet connections is maxing out today’s satellites, setting off an industry-wide stampede toward increasingly powerful high-throughput satellites (HTS).

While that might sound like a good thing, the rush to HTS is driving down bandwidth prices so fast that some fairly low-mileage satellites are struggling to keep up. A study by Northern Sky Research found capacity prices dropped between 35 and 60 percent over the last two years and will continue to decline into next year, with a rebound unlikely.

Seasoned operators, determined to stay ahead of the curve, are thinking twice before investing $200 million or more in a geostationary broadband satellite designed to operate at least 15 years. Factoring in the two to four years it can take to build and launch a communications satellite, such a long-lived asset risks falling behind before it is in orbit.

Big, slow-moving geostationary satellite projects costing about as much as Paul Allen’s superyacht or a Marvel superhero blockbuster have long dominated the satellite communications business. Fifteen-year satellites continue to do the trick when it comes to the biggest source of revenue for operators — transmitting television broadcast signals.

But broadband is a faster animal.

The dominance of big geostationary satellites is about to be tested by the emerging megaconstellations — nimble fleets of hundreds or even thousands of relatively small satellites orbiting the planet at significantly lower altitudes.

Cheaper to build, less costly to lose one or two, the low- and medium-Earth-orbit broadband satellites being built by OneWeb, SpaceX, Telesat and others promise nearly fiber-optic speeds and global coverage without the small-but-annoying lag that radio signals suffer during their 72,000-kilometer round trip to geostationary orbit and back. Though measured in milliseconds, latency is the bane of automated stock trades, hardcore gaming and Skype video chats.

Megaconstellations are not without their own sets of challenges, including uncertainty about how the fast-moving broadband market will evolve and questions about how cheaply satellite receivers can be put into the hands of the world’s least served and close the digital divide (a popular raison d’être for many of these projects).

Planned for non-geostationary orbits, the megaconstellation satellites will orbit closer to Earth for faster connections. Those close orbits mean greater numbers are needed to cover the same surface area, driving the need for mass manufacturing and lots of launches. New types of constellations also means new regulatory approvals from nations where companies want to do business.

But these hurdles aren’t stopping companies new and old from taking a swing at LEO and MEO broadband.

According to Northern Sky Research, at least 10 companies are planning to build broadband constellations of 100 satellites or more in non-geostationary-orbits. Most of these ventures anticipate having their first-generation constellations in orbit in the next five years.

Analysts agree that the advent of huge fleets of small satellites in non-geostationary orbits have the potential to change the paradigm for satellite internet. But no one knows exactly how since the megaconstellations aren’t in orbit yet.

Some satellite operators are hedging their bets by taking stakes in the new constellations; others are buying large communications satellites equipped with features to make them more flexible and maximize throughput.

And some are doing both.

What’s an operator to do?

A Telesat LEO prototype satellite built by SSTL launched in January on an Indian PSLV rocket. Credit: SSTL/Kathryn Graham
A Telesat LEO prototype satellite built by SSTL launched in January on an Indian PSLV rocket. Credit: SSTL/Kathryn Graham

Chris Quilty of Quilty Analytics says the usual metrics for predicting how fast a satellite will pay back the upfront investment “are out the window nowadays.”

Armand Musey, president of the Summit Ridge Group telecom consultancy, agreed.

“The change is so rapid that five-, seven- or 10-year satellites for broadband are essentially all you’d want,” Musey said “You don’t want a 15- or 20-year broadband satellite any more than you want a 20-year-old laptop.”

Avanti Communications provides a recent telling example. In December, the struggling British fleet operator said two of its still-young satellites — one launched within the past five years, the other within the past seven — can barely compete with only slightly younger systems.

WildBlue is another case in point. Boasting roughly 400,000 subscribers and $200 million in revenue when Viasat bought the company for $570 million in 2004, those numbers had shrunk to 100,000 subscribers and maybe $80 million in revenue by 2014, according to Quilty. WildBlue was a “strategically important acquisition,” given ViaSat’s desire to be in the consumer broadband space, Quilty said, “but hard to see how they could have possibly generated a fair rate of return for the satellite.”

Some satellite operators are doubling down on geostationary broadband architectures.

Eutelsat is investing in a new satellite called Quantum that features beams that can change shape, size and power to better deliver capacity where and when it’s needed — a level of adaptability long sought in the satellite industry (Eutelsat ordered its first LEO smallsat last week, but the Tyvak-supplied nanosatellite is aimed at narrowband, Internet of Things applications, not broadband).

Global-IP, a recent startup led by a former Hughes Network Systems executive and advised by former AsiaSat CEO William Wade, ordered a 150-gigabits-per-second satellite from Boeing that’s launching in 2019 on a SpaceX Falcon 9 to connect users in Africa.

And dueling U.S. satellite broadband providers Hughes and ViaSat both have massive geostationary satellites launching in the early 2020s. — EchoStar 24/Jupiter-3 will deliver half a terabit per second of network capacity. ViaSat-3 expects to deliver at least twice that.

Other large operators, including Intelsat, SES and Telesat, are investing in constellations of non-geostationary-orbit satellite systems built for broadband connectivity.

The broadband bandwagon

OneWeb’s initial satellites are coming together at an Airbus factory in France before production shifts to the OneWeb Satellite factory in Florida. Credit: Dominique Eskenazi
OneWeb’s initial satellites are coming together at an Airbus factory in France before production shifts to the OneWeb Satellite factory in Florida. Credit: Dominique Eskenazi

OneWeb, founded in 2012 under the name WorldVu Satellites, took a big leap ahead in 2015 when it lined up $500 million from such heavyweights as Airbus, Coca-Cola and Virgin Group.

SpaceX grabbed its share of 2015’s spotlight when news spread it was opening a Seattle factory to build some 4,500 broadband satellites for a constellation since dubbed Starlink.

In 2016, OneWeb raised another $1.2 billion in an investment round led by SoftBank. SpaceX, focused on building and launching rockets, has refused to say much more about its constellation.

Meanwhile, OneWeb submitted its license application to U.S. regulators, touching off a spate of me-too filings from companies wanting the Federal Communications Commission to know that they also had plans for delivering broadband from non-geostationary orbits.

OneWeb’s initial satellites are coming together at an Airbus factory in France before production of the balance of 900 satellites shifts to the purpose-built OneWeb Satellite factory soon to open in Florida.

SpaceX and Telesat already have prototype LEO broadband satellites in orbit. Telesat’s Phase 1 LEO satellite launched Jan. 12 on India’s PSLV rocket. SpaceX, on Feb. 22, orbited two experimental satellites, Tintin A and Tintin B, on a Falcon 9 rocket.

OneWeb’s first satellites are slated to launch in May on an Arianespace-operated Soyuz rocket.

Come 2019, SpaceX and OneWeb expect to conduct regular launches to put their constellations in orbit.

Telesat plans to select a manufacturer for its constellation of 120 satellites later this year and start launching them in 2020.

LeoSat is working with Thales Alenia Space on a plan to produce a network of 78 to 108 satellites.

And more internet constellations are promised, such as China’s 300-satellite Hongyan constellation, and Boeing’s constellation of nearly 3,000 satellites.

Expectations of success

No one thinks every single proposed constellation is going to reach orbit, but most analysts agree one or more broadband constellation will ultimately pan out. “At least one, but less than five,” ventured Quilty. “Is there room for two? I think there is.”

People have different opinions on who the winners and losers will be.

Tim Farrar, president of the telecom analyst group TMF Associates, counts the medium-Earth-orbit O3b system now owned by SES as the first successful modern non-geostationary-orbit broadband system. O3b’s first satellites launched in 2013 ahead of the current rush.

OneWeb, whose May launch will be of operational satellites instead of prototypes, is in second place, he said.

“At the moment there is a potential scramble for who is going to be the third player between Telesat, LeoSat and SpaceX,” said Farrar. “It’s probably going to be clear in the next 12 months which of those is out ahead.”

The last four satellites in O3b’s first-generation constellation of 20 satellites are scheduled to launch in 2019 on an Arianespace Soyuz. O3b mPower, a second-generation system of seven satellites bringing some 10 terabits of throughput, are slated to start launching in 2021.

OneWeb anticipates rolling out its global broadband service in 2019, starting with Alaska. SpaceX aims to offer limited service as soon as 2020 and Telesat is shooting for 2021.

Who cares who’s first?

While OneWeb appears to be leading the pack, being first to market has its pros and cons, analysts said.

“Being the first to launch a new LEO broadband constellation has advantages, possibly to attract rural customers which don’t have any broadband connectivity or households with poor quality internet access,” said Khin Sandi Lynn, an industry analyst at market-foresight advisory firm ABI Research. “But it also has risk such as LTE coverage expansion or possibility of 5G networks in future.”

AT&T, T-Mobile and Sprint announced at the Mobile World Congress last month that they will start rolling out 5G, or fifth-generation wireless networks, to limited U.S. cities later this year. The long-awaited upgrade is expected to be 100 to 1,000 times faster than current 4G service.

Lynn noted that early this year the second-largest cable operator in the U.S., Charter Communication, announced it will be testing the delivery of 5G service to rural areas — one of the markets that LEO broadband constellations aim to serve.

Miscalculations about market size and equipment costs can be deadly, as ICO and Teledesic found out during the ill-fated satellite broadband gold rush of the 1990s.

In the race to equip corporate road warriors with go-anywhere satellite phones, Iridium beat rival Globalstar to orbit by about a year. Iridium also beat Globalstar to bankruptcy court, filing for Chapter 11 in 1999. Globalstar hung on until 2002.

“I don’t think it’s important to be first. It’s important to get it right,” said Musey.

Both Iridium and Globalstar eventually emerged from bankruptcy. Globalstar finished deploying its second-generation constellation in 2013. Iridium, which got off to a later start the second time around, is well on its way to having all of its new satellites in orbit by year’s end.

Teledesic and ICO weren’t so lucky. Both collapsed short of the finish line with no such phoenix-style resurrection.

“This whole idea of a first mover advantage is a bit overstated,” Musey said. “It works in certain consumer brand applications, like opening a McDonald’s in a neighborhood, but I don’t think that’s ever been an issue in the satellite industry.”

Rather than being an advantage, first movers could find competitors drafting behind them, taking advantage of their trailblazing in technology and regulatory development to speed their own systems, Quilty said.

Justifying megaconstellations

The greater question as to who will succeed lies in how much demand their constellations will actually generate.

“If you make a comparison with the 1990s, the biggest issue was that the market forecasts were dramatically wrong,” said Farrar. “It wasn’t that they couldn’t build the satellites or handsets. They made the technology work, but they didn’t find the market that they expected. That’s always the challenge for something big and new, and it doesn’t matter whether you are building LEOs or GEOs. When you are trying to get into a new market, it’s very difficult to come up with an accurate market forecast.”

OneWeb and SpaceX have both highlighted consumer broadband with an emphasis on connecting large, unreached populations to the internet as a central purpose of their constellations. Those populations will need low-cost, easy-to-install user terminals, otherwise prospective customers will find themselves priced out.

Iridium and Globalstar bet big on a large satellite-telephony market — a much larger market than actually materialized, thanks in part to a faster-than-expected expansion of cellular networks to previously underserved regions.

Sima Fishman, managing director of Euroconsult USA, said the current generation of LEO constellations face similar issues as some of their 1990s forebears, whose ambitious plans sparked a reusable launch vehicle boom that went bust (None of these ventures ever reached space, but many of them managed to raise millions and produce at some hardware all on the strength of good slide decks and non-binding letters of intent from Teledesic and the like).

“The most significant hurdle from our perspective is establishing credibility that demand will exist at the prices necessary to make the business case work,” she said.

A good ground game

Perhaps the biggest variable in calculating demand is figuring out what consumers will have to pay for the user terminal—the receiver and antennas customers will use to connect to the constellation.

In contrast to geostationary satellites that always appear to hover at the same point in the sky, LEO broadband constellations will require antennas capable of tracking multiple satellites as they orbit overhead. Such antennas exist. In use for more than a decade, phased-array antennas commonly have no moving parts, relying instead on electronically steered beams to communicate with satellites. The technology is proven, but the cost is sky high.

For cheap, abundant satellite broadband to succeed, consumers will need cheap but capable antennas. Companies such as Kymeta and Phasor Solutions say more affordable flat-panel antennas are coming soon, but analysts are quick to note that they are not here yet — at least not at price points acceptable to the average consumer.

A camera mounted on Falcon 9’s upper stage shows the deployment of the first of two Starlink demo satellites SpaceX delivered to orbit during the Feb. 22 launch of Spain’s PAZ radar-imaging satellite. Credit: SpaceX Video
A camera mounted on Falcon 9’s upper stage shows the deployment of the first of two Starlink demo satellites SpaceX delivered to orbit during the Feb. 22 launch of Spain’s PAZ radar-imaging satellite. Credit: SpaceX Video

“With the enormous amount of capacity coming online, the only plausible way that it can be absorbed is that there is a large consumer adoption,” said Musey. “The only way you can have a large consumer adoption is if you have reasonably priced ground equipment. That’s particularly true in developing countries, but it’s also true in the U.S. and most advanced countries. Consumers are not going to pay $50,000 or $100,000 or more for a tracking antenna. To hit a consumer price point you need to have flat panel antennas at very low price points.”

In 2015, when Elon Musk first unveiled SpaceX’s broadband constellation plan, he said the flat-panel antennas customers will need to link up with the system would be inexpensive enough — $100 to $300 —to be affordable the world over. Last month, SpaceX acknowledged to The Wall Street Journal that it doesn’t yet know how much the terminals will cost.

OneWeb has said it is working with chipmaker Qualcomm, a OneWeb investor, on low-cost antennas.

Telesat, in an interview with SpaceNews last October, revealed only that the success of the Telesat LEO constellation “is not dependent on a big leap in ground terminal performance and capabilities.”

Consumer confidence

Farrar is among analysts who doubt consumer broadband will be the biggest application for LEO constellations. Bridging the digital divide is a laudable goal, but backhaul — using satellites to help cellular networks to increase coverage and improve service — is an early market they can dominate with or without cheap antennas.

“If you are backhauling from a cell tower, then if a terminal costs a few thousand dollars, that’s not a fundamental barrier to deploying it because you may be spending $1,000 a month on capacity,” Farrar said. “If you are providing home broadband to someone for $70 a month, then a terminal that costs a few thousand dollars makes it completely untenable.”

Not every operator is counting on consumer broadband for success. Intelsat, which tried unsuccessfully to merge with OneWeb last year, is a reseller of OneWeb capacity for almost every market but consumer broadband — in-flight connectivity, maritime, government networks, driverless cars and more. LeoSat, likewise, is targeting corporations instead of consumers.

Changing tides

Regardless of whether the non-geostationary-orbit (NGSO) constellations succeed, they are already having an impact on the satellite industry.

From 2014 to 2017, geostationary satellite orders were well below average, dropping into the teens before falling last year to just seven orders worldwide as operators hedge their bets on what technology is best.

In addition to the investments Intelsat, SES and Telesat are making in NGSO ventures, Asia’s largest geostationary satellite operator, Sky Perfect JSAT took a stake in LeoSat last year. ViaSat, meanwhile, has filed FCC paperwork for a constellation of up to 24 satellites operating in medium Earth orbit.

“A lot of satellite operators are flummoxed by the correct choice, and they are taking steps to stay engaged in NGSO activity,” said Quilty. “There’s only a handful of players that seem to have adopted a wholesale wait and see approach.”

“[I]ndustry players seem to be recognizing that the megaconstellation approach to capacity expansion represents a sea-change in the economics of the satellite industry,” said Fishman. “We see this recognition across the traditional value chain, from manufacturing and launch through operators and service providers, as well as customers. We also see stakeholders such as insurance providers, regulatory agencies/spectrum administrators, and the financial community sorting out the implications.”

Manufacturers are angling for constellation construction contracts by promoting new smallsat platforms. Launch providers are designing adapters and deployers for constellations, or building new rockets specifically sized for dedicated smallsat missions. And operators of ground-based satellite gateways are installing new antennas around the world to provide turn-key solutions for constellation operators.

Yet even as the entire satellite industry pivots to LEO broadband as the next big thing, no one knows for sure what the stars have in store for today’s constellations.

“It’s not an easy business,” Farrar said. “All of this is complicated.”

This article originally appeared in the March 12, 2018 issue of SpaceNews magazine.

Caleb Henry is a former SpaceNews staff writer covering satellites, telecom and launch. He previously worked for Via Satellite and NewSpace Global.He earned a bachelor’s degree in political science along with a minor in astronomy from...