Elon Musk wants to connect all corners of the planet via space, but his thousands of orbiting routers may pose a risk to satellites and to science.
When the call connects and I ask Angel Chavarin if I'm speaking to AWN-hell or AIN-gel, there's a familiar pause. I can hear the faint echo of my own words finally reach the cellphone's speaker on the other end of the line a few seconds later, and then a voice responds:
"Yep, it sure is. AIN-gel works. No one around here calls me AWN-hell except my dad."
It's a delay I recognize from using satellite phone connections while on assignment on the Alaskan tundra and other remote areas. The signal carrying my words must travel over 22,000 miles (35,000 kilometers) to a satellite in geostationary orbit and then another 22,000 miles back to Earth to reach the person on the other end of the call.
But Chavarin isn't speaking to me from a satellite phone in the Alaskan wilderness or any other end of the Earth. The 40-year-old is on a regular cellphone in the tiny Oregon community of McKenzie Bridge, about 50 miles east of Eugene, where he helps run the general store, writes fantasy novels and, until recently, looked after his father, who is at extremely high risk from COVID-19.
"It's pretty rural. There's not a whole lot out here," he says. "There's little communities about every 10 miles or so, but about half of those have been destroyed."
That destruction was wrought by the Holiday Farm wildfire, which tore through the region in September. One person died as the fire torched over 170,000 acres and a few small towns, as well as a lot of fiber-optic and copper lines that kept communities in the area online and in touch.
So now Chavarin's cellphone might as well be a satellite phone. It's connected to a temporary, mobile cell tower that certainly sounds as though it's routing our conversation through geostationary orbit.
Such temporary infrastructure is his sole point of internet access, where the latency -- those delays in the conversation -- is also obvious. That made it hard for his father, who has a compromised immune system and has had pneumonia multiple times, to continue to work from home.
"Most everyone here was working from home anyway, and now they can't do that."
While 2020 has lumbered on in an epic conflagration of storms, fires, a global pandemic, recession and civil unrest, SpaceX has been scrambling to improve life a bit by creating a new kind of satellite service it calls Starlink. It's technology that could be just the thing for folks like Chavarin and his dad.
It's already been put to use by emergency responders helping with the rebuilding effort in the wildfire-torched town of Malden, in adjacent Washington state.
Elon Musk's space company, and competitors including Amazon, are aiming to send hundreds or even tens of thousands of small satellites into orbit. These so-called mega-constellations of flying routers could drape almost the entire planet in an invisible blanket of broadband connectivity.
To Musk, it's a way to both solve a problem on Earth and test systems that might eventually prove useful to his grander ambitions to set up human colonies on Mars. On our home planet, a system like Starlink could help mitigate the array of catastrophes that seem to be on the rise.
But what's good for the internet and local communities could pose some serious problems for astronomers and add significantly to the clutter of machines and debris enveloping the Earth.
SpaceX began launching its Starlink satellites in May 2019 in batches of about 60 at a time. The metal birds are much smaller than the large telecommunications satellites in use now, and they also circle our planet in low-Earth orbit, or LEO, at an altitude of 341 miles (550 kilometers), or less than 2% the distance of geostationary orbit. This allows for much lower latency and the ability to provide a broadband connection to just about any location on Earth, once everything is in place.
Chavarin has followed the development of Starlink since before the wildfire in hopes it might offer an improvement on the DSL service he was previously using. His father had been living with him in remote McKenzie Bridge to avoid the risk of contracting the virus at the community college where he works in Eugene as library technology services coordinator. But when the fire destroyed broadband access, he had to find a newer living situation closer to work, and closer to the virus.
Earlier this year, Chavarin registered his interest in being part of the Starlink beta test. SpaceX began sending out invites to its Better Than Nothing beta program in October. For an upfront investment of $499 to purchase an antenna/router and $99 per month, the program offers data speeds from 50 to 150 megabits per second and latency of 20 to 40 milliseconds.
It certainly would be better than the temporary connection Chavarin's been using since the fire, which sometimes registers over 700 milliseconds of latency.
So far, he hasn't received an invitation to join the beta test.
I've heard from dozens of hopeful Starlink watchers online, as well as from neighbors, friends and family who are eager for a new alternative to mobile hotspots or subpar DSL.
"You have still 10 to 20% of the population, even in developed markets … you still have a significant chunk that have an average DSL or bad 3G connection, and therefore doing video or doing higher-requirement usage is a challenge," says Alexandre Menard, a senior partner at management consultancy McKinsey and a leader of the McKinsey Center for Advanced Connectivity.
For decades now, governments and companies have been looking to orbit for a solution to the challenges of connecting the more remote nooks and crannies of our planet, or at least to provide an option that can theoretically be accessed from anywhere.
So far, the results have been less than revolutionary. Satellite phone and internet service providers including HughesNet, ViaSat, Iridium and Inmarsat offer connectivity for remote locations, but it often comes with sluggish speeds and high latency, made all the more frustrating by high prices and poor customer service. Not to mention the dreaded data caps that have become increasingly crippling in a world that now exists largely in videoconference calls and HD streams.
The satellite internet landscape is also littered with ventures that have failed or run out of funds dating back to the 1990s. Projects like Teledesic and Celestri were among those abandoned around the turn of the century. Potential Starlink competitor Oneweb filed for bankruptcy earlier this year as the onset of the COVID-19 pandemic complicated the startup's financial situation.
Menard says that until relatively recently McKinsey had been skeptical that broadband from LEO had any prospects.
"We thought that it was way too expensive to actually come to life at scale in the foreseeable future. … You need to design, manufacture, launch and then operate [hundreds of] satellites."
That's potentially billions in upfront costs before collecting any subscription revenue.
But in the past half decade the advancement of some key technologies and the involvement of a few tech giants and other investors have changed the landscape of what's possible.
McKinsey cites advances in efficient use of the radio spectrum in the bands where the LEO constellations will operate, improved antennas and processing, and the development of artificial intelligence algorithms to help manage what could be tens of thousands of satellites in a constellation.
Menard also calls the pace of launches that's now possible "phenomenal."
Most of those launches so far have come via SpaceX and its workhorse Falcon 9 rockets, which are quickly closing in on 1,000 total Starlink satellites launched over about 18 months. OneWeb managed to launch 74 satellites out of a planned 650-bird constellation before its bankruptcy filing. SpaceX and OneWeb didn't respond to a request for comment for this story.
Amazon's Project Kuiper and Canada's Telesat are still working toward their initial launches.
Amazon didn't make anyone available for an interview, but directed us to its recent FCC filings. On July 29, the FCC approved its application for a LEO constellation made up of 3,236 satellites. Telesat has recently signed an agreement with the government of Canada to move forward with its own constellation.
There have been rumblings that Apple and Facebook also have ambitions to launch their own satellite systems. Apple didn't immediately respond to a request for comment. A Facebook spokesperson said the company launched a lone experimental satellite but doesn't plan to launch a constellation or become a provider of satellite connectivity.
On top of this, in recent weeks there have been rumors of a Chinese venture aiming to launch over 12,000 satellites of its own to serve the global broadband market.
Given that they're taking up residence in LEO, these nascent constellations have wandered into the fields of view of many astronomers -- literally.
Almost immediately after SpaceX launched its first big batch of Starlink satellites, some scientists began gasping in horror at what they were seeing from observatories around the world.
Victoria Girgis of the Lowell Observatory in Arizona posted an image to Twitter that showed 25 diagonal lines marring an observation of a distant galaxy cluster, each line coming from the orbital path of a Starlink satellite as it moved across the exposure. At the same time, people around the world reported naked-eye sightings of the bright "trains" of Starlinks moving across the evening sky.
And that was with just 60 of the satellites in the sky. SpaceX has since filed paperwork for plans to eventually expand its Starlink system to over 40,000.
Earlier this year, scientists and representatives from the satellite industry came together at a special workshop to address the coming era of new, huge satellite constellations. An ensuing report, released in August, suggested that a new phase of astronomy requiring intense collaboration with satellite operators may be the inevitable result.
"Existing and planned large constellations of bright satellites in low-Earth orbit will fundamentally change astronomical observing," the report begins.
There are a handful of options that could reduce the impact on astronomy, such as limiting the altitude of the satellites, making them less reflective, increasing and improving image processing, and coordination to avoid pointing telescopes at the satellites.
But none of these will totally eliminate the effects of adding thousands of orbiting robots to the sky. Particularly affected will be the upcoming generation of giant telescopes designed to have a very wide view of the cosmos like the Vera C. Rubin observatory now being built in Chile.
"There is no place to hide," Phil Puxley of the Association of Universities for Research in Astronomy said in August.
The comprehensive report did suggest one drastic approach that would keep Starlink and other coming constellations from tainting our view of deep space:
"Launch fewer or no LEO sats. However impractical or unlikely, this is the only option identified that can achieve zero astronomical impact."
Unlikely indeed. New SpaceX Starlink batches are lifting off from Earth every few weeks. The company's permit from the FCC to operate a broadband constellation actually requires that it have its first 2,212 satellites orbiting and operational by 2024.
But SpaceX, Oneweb, Amazon and others have been working closely with the scientific community to address the problem. SpaceX has experimented with ways to make its satellites less reflective.
"We set out two goals," SpaceX's vice president of satellite government affairs, Patricia Cooper, said in October during a webinar organized by the Satellite Industry Association and the American Astronomical Society. "One of them was to reduce brightness. … The second goal was to make the satellites invisible to the naked eye."
Cooper says that over 350, or close to half, of the Starlink satellites deployed are equipped with VisorSats, a sort of shield to reduce the reflectivity of a satellite. SpaceX can also change the orientation of satellites to reduce brightness.
"SpaceX is doing more than promises, they have taken some real actions, which is nice," says astronomer Jonathan McDowell of the Harvard-Smithsonian Center for Astrophysics. "I think they have done enough to ensure the naked-eye sky will be preserved, but I'm still worried about the impact on professional observations."
The United Nations' Online Index of Objects Launched into Outer Space lists over 10,000 objects that have been lifted beyond Earth's gravity well since the start of the Space Age. Of those, maybe half remain, and closer to a quarter are operational.
So it's possible that if all the planned broadband constellations come to full fruition, the total number of objects launched to space by humanity will quintuple over the next decade or so.
That rising volume means an unprecedented new risk of collision. There's probably no reason to worry about a dead Starlink falling on your head. (The small satellites that occupy LEO are designed to easily reenter Earth's atmosphere and burn up completely.) But it does present a threat to other satellites.
In September 2019, the European Space Agency performed an emergency maneuver to move one of its weather data satellites out of the way of a Starlink satellite to avoid a potential collision. SpaceX later blamed the incident on "a bug in our on-call paging system."
"There will be accidents and collisions if the really big (30,000-100,000) version of these constellations happen, and it will be bad," McDowell said via email.
SpaceX has long touted Starlink's autonomous collision avoidance system. To its credit, hundreds more Starlinks have been launched, without incident, since the ESA near-miss. But the real risk may come, as McDowell notes, when thousands of competing satellites are also sharing nearby space. Imagine an operator going bankrupt and leaving hundreds of abandoned robots whipping around Earth at high speed like a driver asleep at the wheel.
Starlink currently leads the way with its ongoing beta test, while Oneweb is now reorganizing itself under new ownership, with the British government and Indian conglomerate Bharti holding the largest stakes.
Project Kuiper and Telesat have yet to begin launching their respective constellations, though Telesat launched a single prototype demonstration satellite in 2018. But both have the means to do so, so there's reason to take all four leading players seriously.
"We think that at least one or two of these are going to come to life in the next two years and start offering concrete services to customers," Menard says.
He says the demand is there for the services the companies hope to offer. If you spend time in rural parts of the world, you don't have to ask around much before you'll meet potential customers.
"No hope of ever getting fiber here and the best we get is 4 MB ADSL," says Matthew Vermeulen from the small town of Ugie in South Africa via an online chat. "I personally am a huge gamer and do all my work through the internet so being able to have the speeds and ping that people in the cities and overseas get to have would be great."
In Ector County in West Texas, many residents find themselves in a similar situation. The county comprises the city of Odessa and the stark, dry, often treeless flatlands to the west of town. Nearly 40 percent of households responding to a survey said they had unreliable broadband service or none at all, according to Mike Adkins, director of communications for the Ector County Independent School District.
The dire state of connectivity came to a head when the COVID-19 pandemic hit and schooling in the county went online. As of October, nearly a third of students were still attending online.
Over the summer, the district connected with SpaceX, which offered to run a pilot test of Starlink in Ector County next year, starting with 45 families and later expanding to 135 households.
"It's just a moral imperative that we find solutions," says Adkins, "because we have so many kids who can't connect with school once they leave the school building."
And for Mars-obsessed Elon Musk, it may be a small step toward solving the problem of how to create an Earth-like environment on the red planet someday.