The telecom press in Canada has been gushing over the news that Elon Musk and his company SpaceX has visions of coming to Canada to provide super-fast Internet service to rural and remote Canadians via their StarLink LEO (low Earth orbit) satellite constellation.

Please take a step back. Pause. Take a deep breath. Just stop it.

This is not imminent, it may never happen. In fact it probably will never happen.

First off, SpaceX only applied for a BITS licence. BITS is Basic International Telecom Service. AurorA has a BITS licence (in fact I believe AurorA had the first BITS licence provided to a non-facilities based carrier). This does not provide StarLink the spectrum required to operate in Canada. OneWeb had licensed Canadian spectrum (through 1021823 B.C. Ltd.) and so does Telesat for their proposed LEO constellation. SpaceX hasn’t yet applied for a Canadian spectrum licence (as far as I know).

The model of “connecting the unconnected”, providing broadband Internet service to the 3.5 billion people in the world that don’t have it has been around for decades. I remember the old Teledesic and Skybridge proposals, ICO, Iridium and GlobalStar who all went bankrupt. Iridium emerged from Chapter 11 eventually after a huge amount of drama (see my review of a great book on Iridium). More recently OneWeb who had already launched 74 satellites went bankrupt in March (read here) and before that LEOSat last November. As an aside, even Intelsat went into Chapter 11 bankruptcy protection in May.

Elon’s SpaceX has 362 satellites in orbit, but to actually provide service their plans call for 4,000 satellites just in the first phase deployment. To do this takes money; SpaceX need billions, and soon, to satisfy its high cash burn rate. It’s not making near enough revenue launching satellites (and astronauts) from others to cover the costs of it unpaid Starlink launches. This is also why they are desperately trying to be included in the FCC’s Rural Digital Opportunity Fund auction to get a piece of up to $16B in funding, even though they cant demonstrate that they can actually provide service (and may be fudging the latency numbers, hard to prove since you cant actually test it) !

There are very few investors that can write the cheques big enough to fund an LEO sat constellation. Even large investors like Softbank balked at committing additional risk capital to a single investment like OneWeb. The history of bankrupt constellations would bear that out. COVID-19 has made raising money for risky ventures even harder. Building the business case is difficult; it costs too much money to make money, even for a SpaceX that has re-invented rockets and dropped the price of launching satellites.

LEO satellite constellations, especially using small nano satellites can provide great services for many use cases; Earth observation, low data rate iOT or M2M, remote asset tracking, weather data and even imaging. GEO sats can provide great high throughput required for TV and telecommunications, but they provide poor connectivity at Northern latitudes like Canada’s Arctic. The use case of providing broadband Internet for the unconnected is just not suited for LEO.

To connect those 3.5 billion people to the Internet requires not only the huge investment in satellites and ground infrastructure (plus legal/regulatory, and non-trivial insurance costs) but also an inexpensive antenna/receiver for the consumer to be able to plug in themselves to access the service. Given that the terminals will be in remote areas they will need to be low power usage. By definition those places are off the grid. Since the LEO sats move quickly across the sky, the tracking antenna must be sophisticated, like a electronic phased array. No one has yet to produce such a low-cost, low power terminal device.

When the constellation is finally setup, the satellites would cover the entirety of the Earth’s surface. Ninety percent of the time this would be over open water, frozen tundra, wilderness and deserts devoid of people. So a business case dependent on a use case like consumer broadband would not generate revenue 90% of the time ! In areas like Africa or the Amazon, the surface is covered by jungle. The signal won’t penetrate through thick vegetation, limiting its usefulness. Finally, how much would people actually pay to use the service ? Has Elon gone out and talked to the people in emerging markets and asked them what they need and what they can afford ? What about when that area finally gets served via a cable or fixed wireless that can provide service for an order of magnitude less ? Eventually even Nunavut will get service by one or more sub-sea fiber optic cables.

So where exactly are the customers going to come from to justify the immense capital cost of building the constellation ?

The other problem that LEO sat constellations have to solve before we actually see this service launched is the one that they create themselves by launching thousands of satellites to do so. Space debris. Low Earth orbit is already crowded with dysfunctional satellites and space debris. What happens if, no when, a StarLink satellite malfunctions or becomes derelict or collides with other satellites or space debris. The attraction of low cost sats is that they are made with commercial off-the shelf components for a fraction of what traditional GEO sats cost. But given their disposability we are going to have a huge increase in space junk.

Without strict regulatory oversight (and who would be responsible for this in space ?) and the non-zero chance of failures and malfunctions, we could see a runaway feedback loop creating tons of space debris, called the Kessler syndrome. These concerns and liabilities have still not been fully addressed for the large volume LEO constellations.

Until StarLink (or any other LEO Sat constellation) actually builds and operates a constellation without going bankrupt, solves the physics of providing low cost-low power consumer terminal equipment and can guarantee they won’t destroy low Earth orbit with space debris for succeeding generations don’t hold your breath for cheap sat based broadband service in Canada’s rural and remote areas.

Instead of waiting for Elon, Canadians should take control of their own future and build more Arctic fiber optic cables.