A rendering of a Telesat Lightspeed satellite based on MDA's new software-defined platform. Credit: Telesat

Telesat finally has funds for a low Earth orbit (LEO) broadband network after pivoting to smaller but equally capable satellites from MDA following production issues at Thales Alenia Space.

MDA announced a contract Aug. 11 to deliver 198 satellites for Telesat’s Lightspeed constellation for launches starting in mid-2026 — six years later than the Canadian operator had originally planned to expand out of geostationary orbit (GEO). 

Telesat had held off picking a manufacturer until 2021 to get a better deal on parts, only for Thales Alenia Space to run into pandemic-related supply chain issues that derailed the project.

MDA’s satellites would be about 75% smaller at 750 kilograms each. Telesat has said the $3.5 billion it has already lined up in equity and Canadian federal and provincial financing is enough for 156 of them, enabling initial polar and global services.

Key to the cheaper satellite design is digital beam-forming array antennas that Telesat had initially deemed too immature for the constellation, which would also at least be partially deployed by Blue Origin and Relativity Space rockets that are still under development.

As OneWeb and SpaceX’s Starlink seek to densify already operational LEO networks, and while Amazon prepares to begin deploying Project Kuiper around the same time as Lightspeed, Telesat CEO Dan Goldberg talks to SpaceNews about the 50-year-old company’s attitude to risk as it races to get its LEO plans back on track.

Telesat CEO Dan Goldberg. Credit: Telesat

What are the next steps now you have the funding for a LEO constellation after all these years? 

We’re working intensively with MDA to launch the program and finalizing contracts with other vendors. We’ll make additional announcements around launch providers, ground segment providers, and things like that soon.

We’re hiring a ton across the company and now have something much more concrete to speak with the customer community about. So it’s finally all about execution across all those different fronts.

Does the plan call for more prototypes, or will the next launch for Lightspeed be full-scale satellites?

The first launch will be the actual satellites, although maybe we’ll launch just a couple of them and do a bunch of in-orbit testing and whatnot before getting on a very active launch cadence to populate the constellation.

When does MDA expect to kickstart mass production? They plan to be capable of producing two satellites a day, although it’s just one daily in your agreement.

My expectation is mass production will probably be in the middle of the second half of 2025.

You have previously talked about launch agreements with Blue Origin and Relativity Space. Is that still the plan?

We have those arrangements in place, and we’ve got another announcement that’ll be coming as well [Telesat announced a contract with SpaceX Sep. 11 for 14 Falcon 9 launches, enough for deploying the entire constellation].

Rockets from the two you’ve announced [before SpaceX] are still in development. What are you hearing from them and others about whether the launch capacity will be there for Lightspeed?

I’d say Blue Origin is the more meaningful of the two we’ve announced so far, and based on everything we’re seeing from them, I think New Glenn will be totally available for us when we start launching mid-2026.

They have a different approach than some others out there, potentially, just in terms of how they move forward with their development, but they’re good. They’re doing all the right things, have very capable people, are well-financed, and their facilities are very advanced.

To get up to speed, can MDA take advantage of any of the early work Thales might have done while you were waiting to close the financing — where were they with the project?

We had done a lot of work with Thales, and they were trying to build prototypes and things. The reality, though, is not really.

First off, MDA was always building the antenna, even when we were on the Thales path. The antenna they’re going to be building now is a digital beam-forming antenna. The one they were going to build before was analog. In addition to building the antenna, they’re also building the bus and the onboard processor, which will be their own design. So it’s not leveraging anything that Thales did.

There was other work that Thales was doing on the original path, mostly around software development that’s customer-facing and software that relates to the operation of the system. I think we’ll certainly be able to leverage some of those things.

Construction has commenced for the first Telesat Lightspeed landing station at the company’s Allan Park facility in Ontario. Credit: Telesat

COVID-19 didn’t help supply chains and production, but why was it so difficult to close the financing for the constellation on the Thales path?

You’re right; COVID was a real kick in the teeth. It did two things: It adversely impacted schedule because of all the supply chain issues, and it adversely impacted the cost that we thought we were going to incur on the Thales path, just because of all the inflation that also came in the wake of COVID.

It was less than two years ago that Thales informed us that the schedule and the [$5 billion] program cost that we had been talking about was something they were no longer going to be able to support.

It then took six, seven months just to get an updated proposal. They had to go all the way back out to their supply chain, and it was tricky to get suppliers to hold up their hand and get firm, fixed prices because we were in this inflationary environment. We looked at different funding sources to fill that gap, including talking to some equity investors.

In any event, I think we would have gotten there. We were making progress, but we were also in parallel exploring whether there was a more cost-effective way to move forward with the constellation, which was when we hit upon a really compelling alternative with MDA. That took sometime. We had to get our heads around whether the digital beam-forming antenna was ready for primetime. Telesat is pretty methodical and maybe not the most risk-embracing organization on the face of the Earth.

We started late last year to really start to explore this MDA path, and it was eight months later that we announced something. So, maybe not light speed, but there was a lot of careful engineering work and review that we needed to do.

Pandemics are tricky to predict but would you have done anything differently in hindsight?

I don’t know. It’s always easy to look back and say, maybe we should have pivoted earlier. I’m not somebody who spends any time in my professional or personal life looking back with regrets. I’m just not wired that way.

To be honest with you, I give us reasonably high marks for being resilient, innovative, and open to exploring alternative paths.

We’re big believers in this enterprise, global broadband market that we seek to address with Lightspeed. In the industry, we were reasonably early in believing that LEO is the right solution. We attempted doing it with Thales — we have a high regard for them, and I think they were the right company at the time to partner up with to do it.

For sure, we did not foresee a pandemic. But I also believe that life almost always works out the way it should, and things work out for the best, and I think that’s what happened here. By getting delayed, it allowed this other technology to develop to a point where it was, from our perspective, viable and the right path to get on.

And it’s a much more cost-efficient path. I mean, it’s always the case if you wait, there’s going to be more advanced, more capable, maybe more affordable technology that’s out there to leverage. But we’re saving so much money here, and the business case as a result, is so much more compelling.

Smaller satellites from MDA have shaved $2 billion off what you thought the constellation would cost, but the delay has affected your early LEO start. Is it now going to be harder to compete?

I never thought that we were going to be a first mover. SpaceX and OneWeb were ahead of us. There’s that old line about how you can tell who the pioneers were. They’re the ones with the arrows in their backs.

There are some risks to going first. If we had pulled the trigger and stayed on our original path, we would have come to market two years early or something like that, but in retrospect, with a solution that would have been less cost-efficient.

Some other things have improved. User terminals, in particular flat panel, electronically steered user terminals, which are a key enabler to the [non-geo- stationary satellite] market, are more advanced.

Hats off to SpaceX and how fast they’re moving. They’ve done a great job of demonstrating how transformative LEO satellite constellations can be. We haven’t seen exactly what Amazon’s constellation is capable of, but they’re pulling the curtain back more on it, and it looks like it will also be a really powerful solution.

Both have fundamentally designed their constellations in the first order to serve the consumer market. They’re certainly also focused on the enterprise market, but they are different markets. The government and enterprise markets demand CIRs [committed information rate agreements], service level agreements (SLAs)… I think the customer community appreciates our heritage of being a business-to-business, enterprise-focused service provider.

So we believe there’s going to be a big addressable market for us when we come to market in 2027. We only need a small share of it to be successful in terms of meeting our business plan objectives.

Do I wish I could bring this capability to market next year? Yeah, we’d love that, but I still think we’re going to be very effective and the customer communities will embrace the solutions we’re bringing.

Having an existing GEO fleet was going to be a differentiator. However, OneWeb looks set to merge with Eutelsat around the end of September also to offer multi-orbit services — how has that changed the competitive landscape?

First off, the OneWeb-Eutelsat combination was not a surprise. We’ve both been big proponents of moving to LEO.

I don’t think it changes the competitive landscape so much. Again, I think this is going to be a very big market. We can all be successful, just like we’ve all been successfully serving the GEO market over the last four or five decades.

It’s been a highly competitive market, and the customers have benefited from that. Still, I think all the operators have been able to achieve attractive returns on their investment, create equity value for their shareholders, and do a good job serving the customers.

You also need to get deployment deadline extensions from regulators to keep hold of your priority spectrum licenses. Where are you in that process?

It’s at two levels. There’s the International Telecommunication Union level where we have a number of regulatory filings, and we’ve taken steps to try to preserve rights here. We just launched our LEO-3 satellite, which will support customer demonstrations as a primary mission but will also preserve some of our spectrum rights at the ITU level.

And then there’s the market access that you need from each country that you’re trying to serve. We have authorizations from Canada, the U.S., and some other countries. Certainly, they have milestones associated with them, and we’ve been engaging with regulators to make sure that we can leverage our constellation when it’s available.

I don’t expect anything on the regulatory side, either on the ITU or the more domestic market access level, that will prejudice our ability to meet our business plan.

There will be some countries like there have always been where full, unfettered market access will be difficult. We’ve taken that into account when we put our financial projections together. I think the regulators are delighted to see Lightspeed moving forward. The regulators tend to want more entrants and competition to benefit their domestic users. So I think we will be in really good shape on that.

Telesat is also an operator that’s well known to regulators all over the world. We’ve always done a good job of being a responsible operator of providing mission-critical services in all these different jurisdictions.

MDA’s plans for the constellation hinge on digital beam-forming technology that Telesat previously considered but deemed too immature. What has changed?

The biggest thing has been improvements in semiconductors that will underpin not only this digital beam-former but the onboard processor that’s integrated in.

When we first considered the digital beam-former years ago, the chips were more power-consumptive and generated more heat, leading to thermal issues. The engineers felt it was stressing the spacecraft’s power and thermal budgets.

So we dealt with that by moving to an analog beam-forming antenna, and by having two pairs of those antennas on each satellite, we could meet our coverage and capacity allocation objectives.

What’s changed is now there’s a new generation of chips that generate less heat and consume less power. It means instead of needing two pairs of antennas — two to transmit and another two to receive, we only need one pair and will still have three times as many beams as we did on the older design.

Lightspeed is using a new software-defined, configurable satellite design from MDA, and it’s going to at least be partly deployed using as-yet unproven rockets. It’s a comparatively risky venture for a company that has historically been pretty conservative. Is this just a symptom of all the transformation in the industry?

I don’t think so. I’ll have to challenge your premise a little bit. Maybe it’s why things have taken us longer than they would take a company that is more willing to embrace unproven technologies. We’ve been all over every system, every subsystem, and on the payload on the bus of these new satellites.

Not that it’s risk-free. We’ve never launched a satellite that’s been risk-free, but we think the risk that we’re taking on here is not disproportionate to the risks we’ve taken on some of our other programs.

You really need to get deep in the weeds on the processor, the digital beam-forming antenna, the optical inter-satellite links, and on the different components of the bus. When you break it down like that, and you can break it down much more, we think this is not an extraordinary risk that we’re taking here.

I’d say the same with the launch vehicles. By the time we launch on the New Glenn rocket, it will have flown.

But I would say that if we’re going to be able to effectively meet the requirements of our customers and deliver an advanced, capable broadband experience, you need to go LEO.

It’s very comfortable to stay in GEO, but we don’t believe that we can deliver a modern, high-quality enterprise broadband experience from there.

We think you need to get close to the Earth and have a latency that’s sub 50 milliseconds. That pushes you to LEO and to a more distributed architecture.

It also pushes you toward embracing beam-forming antennas, regenerative process payloads, and optical inter-satellite links so you have a highly capable, flexible, resilient, and efficient space-based network.

We’ve been innovating throughout our 50 years. We started in C-band, went to Ku-band, and then went to Ka-band. I think we were the first to launch a multi-spot beam, Ka-band geostationary satellite.

If we didn’t do this, we think that would be the risky thing for Telesat’s future. We wouldn’t be relevant to the kind of communications infrastructure that our customers are demanding nowadays. You have funding commitments for a 156-satellite constellation, enough for initial polar and global services to generate sufficient revenues to fund another 42 proposed satellites. 

How much do you need to generate to fund these extra satellites, and when do you expect to hit that goal?

It’s about $800 million to fund the incremental satellites, launch vehicles, [and] we’d probably roll out some additional ground infrastructure at the same time, too. We’re going into service in 2027, and generating the cash flows we need won’t take long.

It might be the case we secure some incremental funding between now and the time we even launch our first satellite, and we just pull the trigger and go directly to the 198.

How would 198 satellites improve services versus 156?

The initial 156 will give us a hugely capable, multi-terabit constellation, but getting up to 198 will give us some incremental capacity and densify the network.

How has the split between the company’s broadcast and connectivity revenues changed over the years, and how do you see that evolving as Lightspeed comes online?

Right now, about 50% of the revenues are from our principally DTH video business — that’s really mostly [from customers in] Canada and then Dish Network in the U.S.

The other 50% are what we call enterprise services, but that’s really broadband connectivity for rural broadband, aero and maritime services, for the energy market, government customers, and those kinds of things.

When I first came to Telesat [in 2006] we were probably two-thirds video and one-third broadband.

Our DTH revenues were growing really quickly back about 12-15 years ago. Not only has that leveled off, but we’ve seen declines there as we’ve gotten renewals for some of our DTH satellite contracts at less capacity or lower rates.

The video business is facing some decline because of cord-cutting, whereas the broadband business has been growing because people just want more broadband, and they want it everywhere.

Certainly, with Lightspeed we think that trend will accelerate. Lightspeed is all about broadband connectivity. Some will be for video streaming, but it’ll be just a big broadband pipe.

This interview has been edited for length and clarity

Jason Rainbow writes about satellite telecom, space finance and commercial markets for SpaceNews. He has spent more than a decade covering the global space industry as a business journalist. Previously, he was Group Editor-in-Chief for Finance Information...