“Based on the company’s study of the broadband market’s rise and how to better support it, I guess it’s something we had to do. For us, you have no choice but to land at LEO.” —Telesat President and CEO Dan Goldberg (Source)

With 1,675 satellites launched and 500,000 requests for service in the United States alone, SpaceX Starlink clearly leads all would-be providers of low Earth orbit (LEO) internet service to consumers. However, it is arguable that Telesat Lightspeed is in the lead, or contending for the lead with SpaceX, in non-consumer enterprise, government, mobile backhaul, mobility and rural community markets. SpaceX will eventually offer service to non-consumers—they have already applied for mobile connectivity—but have focused on consumers from day one.

Telesat is an established Geostationary orbit (GEO) Internet service provider, but it became clear that their customers and others were interested in LEO service, so they began researching markets, evolving technology, and system design. (Goldberg estimates that about 25% of Telesat’s current GEO business could be served better from satellites in LEO).

Both companies started their LEO projects around the same time. In November 2014 Elon Musk tweeted that SpaceX was “in the early stages of developing advanced micro-satellites operating in large formations” and they launched two test satellites in February 2018. In April 2016, Telesat ordered two prototype satellites, to begin tests for the design of a LEO satellite constellation. LEO Vantage-1 was launched in January 2018 and LEO Vantage-2 was launched in November 2017.

In the interim, SpaceX has made spectacular progress in rocket and launch capability while Telesat focused on LEO system design, conducting tests and demonstrations with their GEO customers and others, working on the DARPA Blackjack constellation, and evaluating potential vendors. This work led them to the decision to ignore the consumer market, and that is reflected in the design of their constellation. Let’s consider some of the characteristics of the SpaceX and Telesat systems.

Telesat plans to launch 78 polar-orbit satellites at an altitude of 1,015 km and 220 inclined-orbit satellites at 1,325 km, while SpaceX’s first phase is 4,408 satellites between 540 and 570 km altitude and inclinations from 53 to 97.6 degrees. SpaceX has permission to launch 12,000 satellites and has applied for 30,000 more, and Telesat has applied to launch 1,600 more satellites if things go well. The mass of Telesat’s satellite will be around 700 kg and SpaceX satellites are around 269 kg. As we see below, Telesat’s satellites are larger because they do more—process inter-satellite and satellite to ground data at high rates, adapt to dynamically changing customer requirements, and have multiple thrusters and control systems for de-orbiting a failed satellite.

SpaceX is able to launch 60 of their satellites at a time using their current Falcon 9 rocket, and that will increase to 400 with the Starship. Telesat plans to launch early satellites on Blue Origin and Relativity Space rockets but is talking with others, including SpaceX, for subsequent launches. Goldberg estimates that a Falcon 9 could launch 15 or 16 satellites to an inclined orbit and “something like 13” to a polar orbit. Those numbers would increase roughly six times with a Starship.

Since Telesat satellites will be more expensive to build and launch, they are designing them for a ten-year life compared to five years for SpaceX. SpaceX per-satellite manufacturing and launch costs will be much lower than those of Telesat, but it will cost them more to make and launch 12,000 satellites than it will cost Telesat to make and launch 298. That being said, SpaceX will have the opportunity to update its technology more frequently and will have the cost advantage of in-house, high-volume mass production.

SpaceX beta testers in relatively affluent nations are paying $500 for a terminal and a monthly fee of $99, and customers will receive untiered, best-effort service with no data caps. (SpaceX may charge consumers less in poorer nations if they have unused capacity). Most beta testers are reporting download speeds of 50 to 150 Mbs and latency from 20ms to 40ms with occasional dropouts. Capacity will improve as more satellites are launched but the available capacity will be shared by more users.

Telesat will have far fewer customers but will offer guaranteed service-level contracts with speeds up to 7.5 Gbps to a single terminal or 20 Gbps to a hotspot. They will be able to dynamically configure their service—for example, requiring different performance during the day or night or increasing airport capacity during the holiday travel seasons—which will require a software-defined network operating system. (At one time, they had agreed to use the operating system Google had developed for Project Loon, but Loon was terminated and the network operating system is being designed by Thales Alenia Space, the prime contractor).

SpaceX and Telesat both plan to have polar-orbit and inclined-orbit planes, which will enable global coverage. Telesat has applied for a patent on that architecture. There may be something unique about the Telesat design and, if the patent is granted, maybe SpaceX will have to pay royalties.

Telesat’s orbits are at higher altitudes than those of SpaceX, so their satellites will have larger footprints, hops from one satellite to another will be less frequent, and their customers will be farther apart, reducing the likelihood of interference. There will also be less congestion and the likelihood of a collision at the higher altitude but, if a satellite fails and becomes debris, it will take longer to de-orbit so Telesat satellites have GPS and Gallileo receivers, multiple thrusters and control systems, and grappling hooks for tracking and powered de-orbit if necessary.

SpaceX will be in commercial operation before Telesat, but Telesat has gained cutting-edge experience with their work on the Blackjack constellation and will be able to use later technology. For example, SpaceX initially planned to have five inter-satellite laser links (ISLLs) on each satellite, but the technology was not affordable when they began launching satellites, so they had none. (They now have ten satellites with four ISLLs each in orbit). In the meantime, Telesat worked with Mynaric on the Blackjack constellation, and when they finally begin launching their satellites, they will each have four ISLLs, improving performance and lowering ground segment cost. (Telesat ISLLs will be full-duplex, capable of ingesting 4 x 10 = 40 Gbps and transmitting 4 x 10 = 40 Gbps simultaneously).

Terminals are another key technology. SpaceX will require millions of consumer terminals. They are estimated to cost $1,500 today, but the cost is expected to fall to a few hundred dollars within a year or two. Still, the total cost of consumer terminals will be substantial, and they will have many customers to service. SpaceX has applied for permission to deploy mobile terminals on ships, planes, and trucks. (How long will it be before a Starlink terminal is offered on a Tesla car)? Telesat will have a variety of fixed and mobile terminals when they launch their service, and I expect that SpaceX will also by that time.

It looks like neither SpaceX nor Telesat will have trouble financing their first production constellations. Telesat says it will need $5 billion, and it has existing GEO revenue, government support for rural connectivity, revenue from the sale of C-band spectrum and loans, and is planning a public stock offering to complete Lightspeed funding. SpaceX says it will need $10 billion, but it has received five grants and completed 15 funding rounds. The latest funding round, completed in April, valued SpaceX at $74 billion, up from $46 billion in August 2020. SpaceX will also receive a subsidy for rural connectivity and has a profitable launch business and government contracts.

SpaceX is vertically integrated, manufacturing satellites and components like phased-array antennas, ISLLs, and user terminals, as well as launching their own satellites. SpaceX will develop most of their system in house but will partner with Microsoft and Google on the ground segment and also uses LeoLab’s satellite tracking service. Telesat is handling marketing and customer relations but will outsource launch service, system integration, satellite design and manufacturing, laser terminals, antennas, the operating system, etc. to others. Perhaps they are considering contracting with Google and/or Microsoft for ground infrastructure and service.

I’ve singled out SpaceX and Telesat for comparison because they have made significant progress, but they are not the only LEO broadband contenders. OneWeb is also focusing on non-consumers and they will be offering service before Telesat. I imagine Amazon Project Kuiper will also focus on non-consumers since Amazon has been an infrastructure company since it was founded. Finally, with plans for 12,992 satellites, my guess is that China SatNet will serve consumers and non-consumers in China and countries with Belt and Road infrastructure projects. Starlink, Lightspeed and the others can all succeed if we are able to avoid collisions in LEO.

Update May 31, 2021:

Telesat had hoped to earn $300 million from C-band specturm sales, but looks like the Canadian Government, not Telesat will run the auction for its spectrum and it may or may not compensate Telesat when the sale is held in 2023.

Update Mar 24, 2022:

Supply chain problems and inflationary pressure during delays may cause Telesat to cut the size of their constellation by as much as one hundred satellites or raise more money. Schedules have slipped and they now plan to launch the satellites in 2025 and begin service in 2026—three years after close competitor OneWeb hopes to begin service.

Update May 9, 2022:

Telesat reported that it reduced the size of its planned LEO satellite constellation, Lightspeed. The company still plans to spend a total of $5 billion on Lightspeed, but now plans to operate a total of 188 satellites instead of 298. They will still offer global connectivity but will have less capacity, which will limit revenue.

That’s the bad news. The good news is that they received a $30.65 million contract to demonstrate Ka-band transmission between two Earth observation satellites, between LEO satellites and Telesat-operated satellites flying at a higher altitude (in geostationary orbit), and optical transmission between LEO satellites.