Line-of-sight to the sky? Soon it won’t be a barrier
Today, Starlink has the limitation that it requires line-of-sight to the sky, while people do, of course, often use their mobile phones inside their homes, offices and vehicles.
It was this limitation that prompted the Commerce Commission to lean on One NZ to withdraw its original “100% coverage” Starlink marketing.
But Gartner senior analyst Khurram Shahzad – a specialist in the satellite market – notes that one of Starlink’s emerging rivals, AST SpaceMobile, plans to use low-band frequencies for a satellite-to-mobile signal that will penetrate indoors.
You can think of each of its “BlueBird” satellites as “a giant cell tower in the sky,” Shahzad told the Herald.
Each of AST SpaceMobile’s BlueBird satellites has an array the size of a tennis court. One side has solar panels, the other antenna for transmitting an unmodified 4G/5G signal to standard mobile phones on the ground.
“Giant” is the operative word.
While Starlink has thousands of satellites in its swarm, AST SpaceMobile will use just a few dozen, each with an array the size of a tennis court, the better to give it a huge coverage footprint.
One side of the array holds solar panels. The other, antennae to beam an unmodified 4G or 5G to stock-standard mobile phones on the ground.
This isn’t whiteboard stuff, either. AST – initially bankrolled by AT&T, Verizon, Google, Samsung, Vodafone and others, and now listed on the Nasdaq – recently launched its eighth, ninth and 10th “BlueBird Block 2” satellites, with its 11th, 12th and 13th launching next month.
2degrees, which will be AST SpaceMobile’s partner for its New Zealand service, has already secured the necessary spectrum from the Ministry of Business, Innovation and Employment (MBIE) for a satellite-to-mobile launch later this year.
And just outside Marton, in the lower North Island, 2degrees is also constructing a ground station for AST (its merger with Vocus NZ meant the telco inherited an infrastructure business that, according to industry scuttlebutt – and visual evidence – is also building ground stations for Starlink and Amazon Leo).
The site near Marton in the Manawatū where 2degrees is building a satellite station for newcomer AST SpaceMobile. Photo / 2degrees
Shahzad said that beyond its potential to penetrate indoors, AST SpaceMobile’s service had two other big attractions for telcos around the world.
One was that, although no prices have been published, there had been consistent industry feedback that AST is offering telcos a cheaper deal (one of the reasons its signed up 60 partner telcos to Starlink’s 10).
Shahzad also noted that Starlink’s pricing varied hugely between the likes of the United States, Australia and New Zealand and countries like India and Brazil, where it was far cheaper.
READ MORE
The other was that AST had demonstrated satellite-to-mobile speeds of close to 100 megabits per second with its “Block 1” BlueBirds – that is, close to the download speed of ultra-fast broadband (UFB) fibre when it was first launched in New Zealand – and has promised twice that for its “Block 2” satellites.
Starlink’s satellite-to-cell service offers a modest seven to 17Mbps (Megabits per second) – which is why One NZ, Spark and other telcos have had to restrict their customers to special stripped-back or “lite” versions of popular apps.
AST could potentially offer a full-blooded service from launch. (2degrees has yet to offer local service details, but says its satellite-to-mobile offering will include voice, text and data.)
Starlink promises much faster satellite-to-mobile speeds once it gets its third-generation satellites into orbit next year, which will support up to 150Mbps per user.
Shahzad sees Starlink as “at least two years ahead” of rivals, particularly in terms of being able to offer continuous service anywhere on the planet, but its lead is not as unassailable as it looks.
While it currently has 10,800 satellites in orbit, with permission for another 4000, problems with its Starship rocket mean only around 650 are the second-generation model that supports voice and data for mobile, which is why telcos have had to crimp services so far.
Rivals multiply
Amazon-owned Amazon Leo says it will have just under 400 satellites in orbit by year’s end, or enough to launch its commercial service in Australia and other countries (there’s no NZ launch date yet, but its aforementioned ground station is underway).
Shahzad says Rocket Lab is also poised to compete with Starlink via its US$8 billion ($14b) acquisition of US satellite network operator Iridium, which has 80 birds in orbit and specialises in government, military and corporate contracts – markets where its profitable business has some 2.55 million total users.
A United Launch Alliance Atlas 5 rocket launched 29 Amazon LEO satellites, taking the putative Starlink rival’s constellation to 375. Amazon says it’ll have 400 in orbit – or enough for its commercial launch – by year’s end. It has FCC approval for 7727.
And although little appreciated in the West yet, Shahzad said China has three private LEO (low-Earth orbit) satellite network operators with Starlink-sized ambitions – which will likely offer developing and conflict-involved countries an alternative to Musk’s service.
And neither is it all traditional competition. Shahad noted that Chinese car maker Geely (owner of Polestar, Volvo, Lotus, Zeekr and other brands) is part-way through launching its own 52-satellite network for services for its own vehicles, plus IoT and other broader users, with long-term plans for more than 5000.
Performance is increasing, and the emerging competition will drive down prices, Shahzad said.
The size of the market today
People and companies – or “end users” in telco-speak – are already spending more and more on satellite to mobile, either indirectly by upgrading to more expensive plans that include the service or by paying extra to access the service (Spark charges those not on eligible plans $10 per month for satellite to mobile data, One NZ $20).
A new Gartner report says direct-to-device satellite service connections across Australia and New Zealand have increased from 251,000 to 673,000 this year and are projected to rise to 1.52 million next year (underpinned in part by Amazon Leo’s contract to provide satellite broadband to 300,000 remote Australians from later this year under a contract with the Government-owned National Broadband Network).
Gartner says Australia and NZ end users spent US$5 million ($8.58m) on satellite-to-mobile in 2025 (when Starlink introduced its Direct to Cell service); this year it will hit US$23m, next year US$66m.
“Current demand for LEO D2D [direct-to-device] services in Australia and New Zealand is primarily being driven by consumers requiring mobile connectivity in rural and remote areas, particularly for safety, recreation and off-grid travel,” said Shahzad.
“However, enterprise demand is steadily increasing for field operations in industries such as mining, agriculture, forestry and transportation, where continuous connectivity is critical for operations and worker safety.”
With 85,000 customers for its dish-based service, Starlink will also be drawing around $185m from its fixed-broadband service in NZ alone.
Elon’s million-satellite dream – can it happen?
The Financial Times reported in June – again, in the build-up to SpaceX’s IPO (initial public offering) – that Musk was mulling whether Starlink should launch its own phone service, and even make its own handset, cutting telcos out of the picture altogether.
SpaceX has spent billions to buy up spectrum licences via a deal with US satellite operator Echostar, which could help make it happen.
There are a couple of issues, however. One is whether the average punter is ready for a phone that uses Grok as its interface.
Another was a practical issue in urban areas.
Shahzad said the problem with Starlink, Amazon, and AST’s “celltowers in the sky” is the same as celltowers on the ground: the more mobile phones that connect at once, the greater the congestion (some commentators have seen Starlink connection “demand surcharges”, and some areas missing out on specials as indications of congestion today with its first and second generation satellites).
Even with Starlink’s third-generation satellites, there will only be enough for 1000 to 10,000 people to use a satellite’s bandwidth at once (the number will vary depending on how many are, say, texting versus mainlining Netflix).
Elon Musk’s SpaceX has applied to the FCC to boost Starlink’s constellation to 100,000 satellites. It currently has permission for 14,000 (10,800 of which have been launched). Ahead of SpaceX’s IPO, Musk talked-up the possibility of a million-satellite network.
So even if the US Federal Communications Commission agrees to SpaceX’s request to license 100,000 of Starlink’s third-generation satellites, it still won’t be enough capacity to service more than perhaps a 10th of city dwellers around the globe.
In June, as SpaceX’s IPO approached, Musk ramped up the stakes by suggesting a new target of “up to a million” satellites (with some larger models serving as AI data centres in space, assuming some gnarly engineering challenges around radiation shielding and cooling in a vacuum can be resolved).
But Shahzad said that regulatory, political and, most of all, practical space junk collision challenges would limit the total viable number of satellites in low-Earth orbit to around 100,000 – “with perhaps 40,000 to 42,000 of them SpaceX”.
We’ll get a taste of where things are heading when the FCC rules on SpaceX’s request for 100,000 Gen3 satellites, which would be a whopping 300 to 400 square metres each with their arrays extended – a supersizing that would make Starlink’s satellites bigger than AST SpaceMobile’s (223sqm).
Uphill from here
Shahzad noted that “all of the easiest orbits are already taken”. If satellites circled the Earth higher to avoid space junk, then their internet bandwidth would degrade and more satellites would be required to provide the same level of service.
If they wanted to fly lower, the increased tug of gravity would mean more fuel used on course adjustments, so they’d fall out of the sky earlier (LEO satellites have a five- to seven-year lifespan until their fuel source runs out and they can no longer course-correct and fall out of the sky, burning up in orbit).
So things could get more expensive for satellite network operators from here, amid more competition putting a lid on what they can charge their customers.
The 5% – it ain’t enough
The kicker: Even assuming Musk could use his (undoubted) engineering genius and (doubtful) political wiles to supply 40,000, 100,000 or a million birds in orbit, there just wasn’t the demand, Shahzad said.
There was no problem to solve for most urban customers, who already have fixed-wireless (cellular) or fibre broadband that performs better and is cheaper than a satellite-to-mobile connection.
“In New Zealand, Australia and the US, local telcos already cover 95% of the population, so the remaining 5% is the only target. And in Africa, India and Latin America, people just don’t have the buying power.”
So satellite-to-mobile won’t eat the telcos’ lunch. In fact, it’ll help pay for it.
Shahzad said telcos were “highly motivated” to offer satellite-to-mobile “to reduce churn, differentiate their offerings, and capture an estimated 10% to 30% attractive average revenue per user per month uplift without the capital expenditure of building terrestrial towers in challenging geographies”.
Chris Keall is an Auckland-based member of the Herald’s business team. He joined the Herald in 2018 and is the technology editor and a senior business writer.