WASHINGTON — The number of smart engineers with lots of career options who want to work on LEO-orbit broadband constellations has increased with the news that Amazon is looking at developing its own system.
Amazon has deep pockets and a reputation for sober decision-making. It has started hiring outside Seattle, where a pool of experts formerly working for SpaceX’s Starlink project is available for duty for a company with patience and cash resources.
It is not yet possible to gauge how serious Amazon is about this, but the company’s entry into the LEO broadband world alongside OneWeb, SpaceX, Telesat and others does give pause, even to a confirmed GEO-orbit broadband believer like Viasat Inc.
Viasat Chief Executive Mark D. Dankberg and David Ryan, president of Viasat’s space and commercial networks division, outline here the company’s thinking about LEO satellites’ latency and cost metrics, Amazon’s move, Inmarsat’s GX Ka-band competition, China’s in-flight-connectivity market, the status of the terabit-per-second Viasat-3 global broadband system and the Viasat-2 satellite’s antenna anomaly.
But first, Dankerg addressed the possibility that satellite fleet operators SES and Intelsat may find themselves flush with cash in a U.S. government-approved auction of C-band spectrum.
It now looks likely that the Intelsat-SES C-band clearing proposal to the U.S. government on could result in a large, possibly transformative, liquidity event for both those operators. Any implications for Viasat?
This is hard to handicap. Intelsat has a lot of debt so there is the consideration of what their shareholders get out of it and what do they do on the debt side. SES is a different story.
We are competing in an environment where multiple players are putting multiple billions of dollars in essentially the same segment [broadband connectivity]. So this [C-band windfall] will not be an enormous change.
People ask us: How are we going to compete? We are aiming to compete on productivity. We can come up with space systems that have way more useful throughput. That’s still the case. Nobody’s undertaken satellite projects like we have. There are learning curves we are going down that others haven’t started on. We are in a very capital-intensive business and asset productivity has a large effect. That’s what we are investing in.
And if they suddenly have a few billion in cash and need to deploy it?
Anybody who has money to burn can buy lots of assets. But if the assets aren’t productive, you won’t get a good return. Markets have been pretty punishing in the long run when that’s the case.
There’s also the M&A consequence of two operators whose video business is struggling now having cash to purchase others — for example, Avanti of London, with which you have a relationship.
Yes, but all our relationships are based on contracts, and to the extent we have binding contracts with them, that’s pretty much it.
Does Avanti have a role in your future Viasat-3 global network in terms of providing an orbital slot?
No, we have not discussed that with them. We have established a relationship with them that is mutually beneficial, but in the overall scheme of things, they are not [on the critical path]. They are trying to get a return that will work out for their new investors and we’re trying to work out a way to be helpful.
You mentioned the billions being invested in satellite broadband. OneWeb now has enough financing to launch satellites; Telesat’s LEO project is moving, even if financing is TBD; SpaceX will launch a group of Starlink satellites this month; and you saw the Amazon announcement in satellite broadband.
Might there be something to LEO that you missed?
Amazon is a very pragmatic company. The fact that they think there is something there is interesting. They have plenty of ways to make money and grow. They don’t have to do satellites. The fact that they think there’s a business there is something of a validation of the space business itself.
We’ll see how that progresses. The other thing about Amazon is, if you think about transmission as a means to an end, Amazon has the ends which they could put to good use.
Facebook does too.
Yes, Facebook, Apple, Google, Microsoft and Amazon — they’ve all have got the wherewithal to do it if they want, and they have got applications.
The fact that Amazon is going to do something may motivate some of the other ones to think about it. But at the end, this issue of whether low Earth orbit is the right way to deliver bandwidth. Nothing Amazon has said changes the fundamental questions about low Earth orbit versus GEO.
You go into a restaurant and leave unsatisfied. A year later, someone with good taste eats there and loves it. Do you not think about trying it again?
It’s the consumers, the users of those systems, that matter. What we have said, and it’s still true even in rural markets, is that we don’t have enough bandwidth. We have the most productive assets and we know we don’t have enough. We have a next generation that is multiples of this, and that still won’t be enough.
Go back to the latency issue. Latency only is a factor if you have enough bandwidth. If your network is congested, latency is irrelevant because the delays from everything else are so long.
Even in emerging markets, the demand for bandwidth is enormous. You go into these rural villages in Mexico and they want to watch video as much as anybody else. So the real issue to us is whether these low-orbit systems are going to be more productive than what we can do at GEO. We don’t see that.
But Amazon’s Jeff Bezos has hired smart people who presumably don’t have skin in the game, and who told him: We can make this work.
We have interactions with Amazon. They are disciplined and smart. My own sense of what happened is that there is a perception that reducing the cost of launch is a big enabler for LEOs. SpaceX has low-cost launch, and [Amazon-owned] Blue Origin does. My understanding is that’s a big part of where they started.
The other thing is that SpaceX let go of the leaders of their Starlink project, and the perception was that Elon [Musk] was in a big rush, and that these people wanted to be more methodical, and actually they were aiming higher.
So if you look at the Starlink capability now, compared to what they were aiming for a year ago, it’s far reduced in order to meet that schedule.
You refer to the modification that reduce the orbit for part of the Starlink constellation?
If you look at the filings, there are quite a few changes that will substantially reduce the capability of the first generation to meet that schedule.
David Ryan: Which reinforces what Mark was saying before: The physics starts weighing against you in having a network that doesn’t saturate when it’s over cities and is totally dormant when it’s over oceans or rural areas.
Dankberg: So the Starlink altitude is lower, I think the EIRP is lower, and in order to get the geographic coverage they have increased the look angles they are going to support. It’s not clear they are going to have cross-links.
Surely they’ll include inter-satellite links.
I’m just saying it’s not clear. What happened is, a number of people at Starlink said: We have this visionary system that got cut because of time and expediency. So they went to Amazon and said: There is this great system we think is possible, SpaceX is not pursuing it, you guys have a longer time horizon and you can make a bigger investment. Amazon says: We’re in this for the long run, we want to do that. Come over here. They have a system that they’ve filed and they’ve done a little bit to try to increase the fraction of time that the satellites are over populations.
So they are looking at it. I don’t know that they are committed to it.
Is it clear what frequency Amazon plans to use?
We think it’s Ka-band.
I know this sounds audacious but one Viasat-3 is comparable to the whole OneWeb first-generation constellation. Starlink was going to be more ambitious and there are estimates of what that was going to be. What we’re aiming for in our next generation would be almost the same — a satellite that has the same throughput as a whole constellation that costs $5 billion to $15 billion.
If we can do that, we’re going to be fine. Our capital costs are going to be so much lower. Our satellites’ lifetimes are going to be a lot longer.
Understood. But how long does the line of smart people have to be before you think: Maybe there is something to this LEO movement.
I’m not saying that what Amazon is doing is dumb. In their context it may make complete sense. In our case, we have a number of applications that are very bandwidth-intensive. Video is a big part of it. Amazon has video but they have e-commerce and other things that are not so bandwidth-intensive.
I don’t know what they’re thinking. All I know is that if you look at what the assessments were of the full-blown Starlink that they are not yet pursuing, that was a pretty capable system, very sophisticated. Still we don’t think it was economically better than what we’re doing.
For the markets we’re going after, we’re going to be fine.
Suppose you’re right: These systems aren’t business successes but do get launched. With the magic of U.S. Chapter 11, they’re purchased for pennies on the dollar. They’re still a factor, like Iridium and Globalstar from 15 years ago.
Iridium fared way better than Globalstar, partly because the satellites lasted far, far longer than they were expected to. They were way over-built — $5-$6 billion for 70-plus satellites.
But if you look at what’s going on now, these satellites are not being overbuilt. They’re built for a five- to seven-year life. That turned out to be a big problem for Globalstar even though their satellites also lasted longer than people expected. The deterioration of that constellation was a big issue.
OneWeb, which is bent-pipe, is going to have issues even if they go through bankruptcy. I don’t think these satellites are going to last 10-12 years.
SpaceX is trying to get something up that’s a first generation and then they’ll reinforce it. You’ll have a pretty short lifetime for those satellites.
But the general issue of what happens if these systems get restructured is a fair question.
Inmarsat is looking at standard satellite buses to which payloads would be added. The idea is a lower cost per delivered megabit for a constellation in GEO orbit to fill in their Global Xpress Ka-band service. What do you make of it?
Let me go back to productivity. The thing you need to sell a lot of bandwidth at a low price is assets that are extremely productive.
They might get that with this new generation of satellites.
I don’t think so, and I’ll tell you why. Throughput is dependent on the illuminated bandwidth and the power that it has. Software-defined satellites use a lot of power that do not deliver capacity.
We were the first who did this. We did free bandwidth on JetBlue and Qantas, and we did Netflix. We’ve been the main catalyst for people’s increasing expectations for in-fight connectivity.
More people want to use it, and those that use it want more bandwidth. We have the most productive satellites in the world, and it’s still a struggle. Smaller satellites, with less power, aren’t going to have the structural ability to do some of the things that we’ve done. And you have to do all those things in order to get the productivity.
You’ll see the low end of these systems come up. But our objective, working with the airlines, is to raise their expectations and have them use it more. We are more encouraged that that is the right track when we look at some of the things our airline customers are doing.
Dave Ryan: One of our other strengths is that we’re very vertically integrated. If you’re building the terminal, the network and the satellite and you optimize that as a system, you can be much more efficient in how you deliver higher speeds to more people and more places.
It would be nice to be able to quantify that advantage. Not everyone thinks it’s significant. You saw the Eutelsat announcement with Konnect over Europe and Africa — Thales Alenia Space satellite, seven General Dynamics 9-meter gateways, Hughes ground network.
For Viasat-2 we use gateways that are half that size, and we have way more gateways. You can’t have more user link bandwidth than you have feeder link bandwidth. Having 45 gateways [for the current Viasat-2 network] gets you a lot more bandwidth than having seven gateways.
And [Konnect] is a smaller satellite with much less bandwidth. We’re doing tons of testing on our Viasat-3, and it’s got around 100x more throughput than where we were 10 years ago. To get there, we’ve have to change almost everything, with hundreds of gateways. We also have things like in-flight-connectivity coverage.
One of the things we put in there, which I still don’t see on any of these other satellites, is this notion of dynamic beam hopping so we don’t have to put a fixed amount of bandwidth in every beam. We can flex wherever the demand is.
We have looked at the filings for [Hughes Network Systems‘s] Jupiter 3 and they say flexibility is over rated, because they know where the demand is.
More regional satellite operators are announcing Ka-band broadband payloads. The most recent is Measat, with an Airbus satellite. Is that an issue, especially in Asia, as you plan the deployment of the Viasat-3 global network?
One of the things we’ve done in going into these countries is play to our strength in being integrated. We can do networking and operations. You see from our work in Brazil and Australia that they don’t need to use our satellite. We’re going to do that in China.
In China, on the aero side, there is the domestic air market but also the international market. On the international side, the Chinese airlines want to be competitive, and the international carriers want to be competitive outside their borders and inside of China. So there is some room there for a deal.
For Measat, we could be a very big consumer of these third-party satellites. one of the really good things about Viasat 3 is the flexibility. Countries can say to us: We have our own national satellite but we like the way you bring things to market. We like the aero, we like the defense things, we want to cooperate.
We have plenty of other places to use that bandwidth, we don’t have to use our satellite capacity in Brazil.
What milestone should we look for in the China Satcom agreement you recently announced on in-flight connectivity?
The most important thing is agreements with airlines. Thats what Chinasat wants. They want to be in the in-flight business and that means airlines. Chinasat has a very good satellite, it covers a lot of air routes.
What we have is a good reputation in the industry, especially at Ka-band. So there’s a potential for a good match.
Who goes to the airlines with this, you or Chinasat?
We both do. What the airlines want to know is: Can we get the JetBlue or American Airlines service? They want to know whether the satellites will support that. I think it’s a good combination. But really, the test is signing up airlines.
What’s the status of your Brazilian business? Are the legal challenges behind you?
The GESAC [government e-learning initiative] is progressing. It is very important to Telebras, and we are supporting them. We are very close to having the last approval from the TCU [Brazil’s Federal Accounting Court] on the contract. They acknowledged that the contract is legal, but they wanted some amendments to it. They need to rule on those final amendments. The Supreme Court has dismissed the other lawsuits about the contract.
The main thing we have been able to do is the GESAC contract. Bringing the rest to market hinges on the TCU.
Now that the insurance claim has been received, can we talk about what caused the Viasat-2 antenna anomaly? Boeing satellite, Harris Ka-band antennas and what you described as a problem in the antennas’ deployment in orbit.
David Ryan: We’re pretty confident that we know what happened, that the problem is stable and we’re not going to lose more capacity on the satellite. And we don’t think it’s going to affect our operations in the future now that we’ve adjusted for it. Here you can go back to our vertical integration. If you understand how it’s all working as a system, then you can adjust the network along with understanding how the satellite’s operating.
Mark Dankberg: Our point is that this happened, we’re still buying satellites from Boeing, and we’re still interested in those types of reflectors. Let’s learn from it. It was a very specific failure mode related to the combination of the spacecraft and the antennas.
It couldn’t repeat on Viasat-3?
No. First of all, it was a specific mechanical failure mode. And even if that antenna effect were to occur due to some other specific combination of events, the architecture of the satellite is robust. Once of the things with Viasat-2, and it’s not super surprising, is that a very minor mechanical dislocations can have a pretty substantial effect.
One of the things that we have wanted to do all along is make it so that we are robust to those types of mechanical issues. We feel confident in it.