PARIS — Startup narrowband data-communications satellite constellation operator Kepler Communications of Canada has reached the crucial “bringing into use” milestone for international regulators following the successful Jan. 19 launch of its first satellite aboard a Chinese Long March 11 rocket.
It was Canada’s first launch with China and makes Toronto-based Kepler the first commercial operator of a Ku-band satellite in low Earth orbit. (Eat your heart out, OneWeb).
Kepler is now setting its sights on serving its first early-adopter customers for the satellite starting this spring. Later in the year it will launch a second satellite, also built by Scotland’s Clyde Space, but this time the launch provider likely will be India’s PSLV rocket.
Kepler has raised some $5.1 million in two funding rounds, including a $100,000 convertible note with Right Side Capital Management and a $5 million seed-funding round led by IA Ventures of New York and Liquid2 Ventures of California.
Kepler Chief Executive Mina Mitry said the existing funds will carry Kepler through the launch of the second satellite and into early 2019. The next capital raise will be used to expand the constellation.
In addition to Clyde Space, Enclustra of Switzerland and Comtech’s AHA Data Group contributed the payload. Kepler is working with launch integrator Innovative Solutions in Space (ISIL), Kongsberg Satellite Services and Comtech EF Data for the ground network; and Nextologies of Canada, a content and distribution network. Further down the road, it is looking at flat-panel antennas for its user terminals and has been working with Phasor.
Mitry discussed the company’s next steps now that its satellite is operational.
What is your satellite’s status in orbit?
Everything is operating as expected. We have a ground station in Canada and it is reporting healthy status, all our deployable performed as expected, with the satellite transmitting.
And the orbital injection was good?
It was actually surprisingly accurate. It was a 540-kilometer, sun-synchronous orbit.
This constitutes “bringing into use” for the regulators at the International Telecommunication Union (ITU), correct?
This can constitute BIU. Our regulatory filing would allow this to be BIU. If you follow ITU Working Party 4A, there is a high probability that they are going to try and retroactively change the rules. Some of the proposals going out within the ITU now are looking for the application of WRC-19 [the next big ITU conference] to be retroactive. But this has yet to be determined.
What is next for Kepler now that you have a satellite in orbit?
Right now we are prioritizing customer deployment. We have a few early-stage customers that are planning to take advantage of our first operating assets. That looks like it will happen around April.
After that we have a second satellite that is scheduled for launch in the second half of 2018. Then we are looking to put together another funding round that will enable us to put together our larger constellation.
Is the second satellite’s launch firm?
It looks like we’ll be going up on an Indian PSLV. It’s not finalized, but Innovative Solutions in Space has been good about giving us flexibility to be able to maneuver between vehicles. So our timeline is set, but given ISIL’s unique capabilities in purchasing capacity on a number of different vehicles, and the standardized form factor of the spacecraft, they are able to pretty easily maneuver between launch opportunities.
On the financial side are you set through the launch of the second satellite and its early operations?
Yes, we are pretty well financed through the start of 2019. The capital we are looking to bring on board this year is just to put more fuel on the fire. The intention is to launch more spacecraft to serve the customers that we are seeing.
You see your constellation size as growing with demand. But is there an optimal number if demand develops as you see it?
The unit economics are such that a single spacecraft can pay for itself given the nature of the service we are providing. Increasing the number of spacecraft, though, has a disproportionate outcome. The number of customers you can serve with just one or two satellites grows exponentially when you have a significantly higher-quality service.
So although the economics works with one or two satellites, it’s a meaningfully more exciting business when we scale the infrastructure.
And where to those lines cross?
I would say around the 50-satellite mark.
Is Clyde Space still set to be your prime contractor if you went to that volume?
We are taking a look at all the options that are available Clyde would be a great contender but we’re evaluating the various options when it comes to a larger-scale constellation.
Who is doing your Earth stations?
Our Earth stations, which act as teleports, we are deploying ourselves, purchasing equipment from various suppliers like Comtech EF Data, which is a vendor to us. And we are deploying to KSAT [Kongsberg Satellite Services] in areas that they have available. As an always-on communications service it becomes important for us to have our own teleports.
And for the user terminals?
For the user segment it’s been pretty open. We have a partnership with Phasor in the U.K., helping to develop user equipment for flat-panel antennas. We’re also looking at repurposing existing VSAT terminals. Anything that is deployed as a tracking VSAT could be repurposed to use the Kepler network.
The near-term service deployments are looking to repurpose existing tracking VSATs to connect to our network. Then the longer-term play is with flat-panel antennas.
Have you run into any issues with the Ku-band geostationary-orbit operators?
We really don’t have to do anything with GEO operators. Per the ITU regulations there are requirements on equivalent power-flux densities. The ITU has a great piece of software that calculates statistically your likelihood of exceeding that. Since we abide by the EPFD limits that are provided. We don’t foresee any coordination or anything with the GEO operators.
Part of that is being driven by our software-defined radio, wherein we vary our communications bandwidth and power and elevation angle by which we communicate.
So you have run your network through the Transfinite Systems software recommended by the ITU?
You look at your market as only secondarily Internet of Things?
Let me put it another way. We are a data-focused communications constellation. IoT is one of the sectors we serve by being data-focused.
We are more targeted toward industrial IoT and data-heavy sensors as opposed to some businesses that require a very limited amount of data, which I think a lot of the other players are targeting.
That’s true but it’s still a crowded space right now for LEO constellations, isn’t it?
No, we look to distinguish ourselves in looking at different market opportunities and what we look to offer. On the other side of IoT, to draw together a real IoT business you really need scale. It’s millions of devices connected. This is very hard to do. Look at Globalstar, Iridium, Thuraya: The time it took them to reach 100,000 subscribers was non-trivial. Trying to make the leap to millions of subscribers to make the business cases a little more interesting, it’s very difficult.
And the second part of this is that some of the other spectrum offerings are pretty limited: L-band, S-band, and even into UHF and VHF, the total spectrum available doesn’t make it easy to reach this level.
You have said that Ku-band can offer a service equivalent to terrestrial wireless but that it needs “substantial technology advancements.” What were you referring to there?
Ku-band has 2 GHz of downlink spectrum and 500 MHz of uplink spectrum. For true large-scale IoT you are talking about deploying hundreds of millions of devices. And even though any individual device would occupy very small amounts of data, you still need sizable spectrum bandwidth to be able to serve that massive need.
You have a high amount of bandwidth in Ku-band, which allows you to operate large numbers of devices, whereas in L-band, S-band and UHF/VHF the available spectrum is pretty limited. So they would have to prioritize customers that have higher ARPU or they can’t serve the mass market required.
What do you estimate as the capex required for a 50-satellite constellation? Less than $200 million?
I can definitely say it will be far less that $200 million, far less than the cost of a single GEO satellite.