Why Starlink Matters for Spaceflight -- Updated

 Why Starlink Matters for Spaceflight

Originally published on weekly-spaceflight.webnode.com in June 2021. Updates in white.

There has been a lot of talk about Starlink, SpaceX's work-in-progress satellite megaconstellation. Will it provide a valuable and much needed service to millions without access, or is it just for gamers in Alaska? Is it a scourge to astronomy or is SpaceX doing a good job minimizing the impact? Is it a danger to spaceflight or are the Kessler Syndrome fears overblown, not taking into account the effectiveness of active collision avoidance and the self-cleaning nature of low orbits? Is it driven purely by profit or are they really going to try to finance Mars missions with this? Will it be profitable or is it a scam? Have Starlink launches gotten boring or is every launch exciting? And will SpaceX go bankrupt trying to do build it?

Amidst all that debate, it's easy to miss arguably the biggest impact of Starlink on spaceflight: its full frontal assault on market inelasticity.

For decades, the biggest barrier to the space age future the Apollo generation expected has been the cost of access to space. It costs thousands (and in some cases, tens of thousands) of dollars just to launch a kilogram to orbit, let alone to other planets, and that's if you buy in bulk. The cost hasn't — or until recently hadn't — changed much in decades, and not for lack of trying.

What it boils down to is that orbital launch systems are very expensive to develop, and they're somewhat less expensive to develop if they only need to launch a few times or a few dozen times a year. To recoup their development and operational costs, launch providers have to charge a lot if they launch relatively infrequently. And because launch is expensive, payloads are small and rare — nobody wants to overpay, so they size their space systems to be as small as possible while getting the job done. But because payloads are small and rare, the per kg cost of space access doesn't come down, since you still have thousands of people who're employed to provide launch services.

This is the chicken and egg problem of space launch costs. No cheap rockets, no big increase in payloads. No big increase in payloads, no cheap rockets.

The classic proposal to solve this has been along the lines of "build it and they'll come." If we built a rocket (maybe with some development cost savings thrown in) that we could price lower, people would build bigger and more numerous payloads, driving a virtuous cycle of ever cheaper rockets and ever bigger / more frequent payloads until we get to a future where space transport is almost as ubiquitous as air or sea transport.

That has been the dream, and the history of space launch systems is littered with the remains of projects that have died trying to achieve it, from cheap pressure-fed booster concepts to attempts at reusable space launch. And all the while the age-old counter-argument — one that ULA tends to bring up when explaining why they're not more seriously pursuing reusable boosters and cost reduction — has been that the launch market is inelastic, meaning demand won't rise much as cost drops. Reduce the cost of space access, they say, and you won't get substantially more launch demand. You'll just end up launching the same payloads for less.

To date, the most ambitious cost reduction projects have failed due to technical, regulatory, and/or development funding issues. So the idea hasn't really gotten a fair shot. But in the last decade or so, we've seen launch prices drop by about half, spearheaded by SpaceX's Falcon 9 rocket and followed by industry around the world — we've gotten a taste of what might be possible with cost reduction. In the meantime, launch demand has not seen a corresponding uptick.

The glass-is-half-empty folks look at this and say "see, it doesn't work", while the half-full folks say "give it some time; it'll take a while for payload providers to catch up, and the demand is on its way; just look at all the megaconstellations in the works."

Personally I lean more on the half-full side, but there's a kernel of truth in both perspectives. And there's an added wrinkle to the launch demand picture: the projected massive increase in the number of satellites is largely enabled by the miniaturization of electronics and the shift from geostationary orbit to low Earth orbit. The actual increase in launch demand is not as large as one might expect, and it's far from clear how much of it can be attributed to launch cost reductions.

SpaceX itself, as it ramped up its launch cadence while simultaneously increasing the performance of Falcon 9 to be able to launch over twice as much as originally spec'ed, experienced a temporary peak in launch rates in 2018. 2019 saw less than 2/3 the number of SpaceX launches compared to 2018. This was due in part to a periodic lull in the global launch market and in part to SpaceX exhausting their launch backlog after building up their cadence. In 2020 and so far in 2021, the picture hasn't been much different, except for one thing: Starlink.

2020 saw SpaceX do more Starlink launches than launches for paying customers, while breaking its 2018 launch record. And in the first 5 months of 2021, we saw 3 (plus a fraction, counting rideshares) paid SpaceX launches but a whopping 13 Starlink launches. SpaceX is well on its way to shattering its annual launch record again, by launching mostly Starlink satellites. In the first 11 months of 2021, there were a total of 26 SpaceX launches, and 15 were Starlink. The rate of non-Starlink launches increased, while there was a big hiatus in Starlink launches as SpaceX transitioned to v1.5 hardware. SpaceX is still poised to break its annual launch record (26 in 2020) but no longer by a wide margin.

So why is this a good thing? They're not getting paid for those launches, so it means the market isn't expanding, right?

Actually, it is. Pretty much every commercial project starts with upfront investment, with the hope of making a profit in the future. For both aerospace and communications projects, the investment tends to be very large (typically in the billions of dollars), and the expected profit tends to be years down the road.

This is exactly what we're seeing with Starlink. The only thing unusual is that the launch provider happens to also be the launch customer. While OneWeb is raising billions to develop its satellite constellation and paying for Soyuz launches, SpaceX is raising billions to develop its satellite constellation and launching its satellites on its own rockets.

The onrush of competitors, investor interest, and economic analysis of projected demand at achievable service prices all suggest the communications satellite megaconstellation market is far more likely to take off than to fizzle out. Starlink, with its technical, first-mover, and low-launch-cost advantages is poised to be highly successful. So future Starlink profits, with high probability, will end up paying for the launch of the initial constellation and for frequent launches for upkeep and expansion.

A recent leaked email and public comments by Elon Musk suggest the financial situation is more precarious than previously thought, but in context the disparity is not that great. What it seems to boil down to is that v2 hardware that's already in development is needed to be highly profitable in the market conditions of the future, but building the v2 constellation requires the upmass capabilities of Starship. Delays in Raptor engine production have slowed down Starship development, which in turn risks delaying Starlink v2. If those Raptor issues aren't solved in time and Starship's pace doesn't pick up in the next year or so *and* there's a recession and liquidity crunch, Starlink v2 (and by extension SpaceX) could be in trouble. The upshot is that employees are urged to put in more hours to resolve Raptor production issues and get Starship moving fast again. This is not yet an existential threat to SpaceX, and Starlink is still projected to be highly profitable.

And why is that a big deal for spaceflight? SpaceX put a lot of money into making the Falcon 9 first stage and payload fairing reusable. That reuse can cut launch prices down by another factor of two over current levels, but not with annual launch cadence in the teens — they still have all those employees to pay. With cadence in the several dozen per year, however, Falcon 9 launch prices can drop to near marginal cost and still be profitable.

And then there's Starship, promising far more drastic price cuts. But again, launch demand has to go up to realize those savings.

Starlink has gotten the ball rolling, and still has plenty of room for growth. Starlink satellites have a planned service life of around 5 years, and SpaceX already has plans for a constellation size of around 40,000 satellites. Maintaining such a constellation would take the equivalent of over 130 Falcon 9 launches a year. Substantially more with the larger v2 satellites. And it seems SpaceX isn't intending to increase Falcon 9 cadence anywhere near such levels. They are planning on Starship taking over soon.

And there's no theoretical reason for them to stop at 40,000. Perhaps even more promising is the potential for substantial increases in the size of Starlink satellites. Larger phased array antennas would enable both higher total bandwidth and tighter beams, opening up higher population density markets and the ability to serve more data to more customers, and with smaller user terminals. The v2 satellites have already started going down this path, it seems.

Starlink could provide Starship with the bulk of the launch cadence it needs to get safe and cheap enough for the next level of high mass applications like in-space manufacturing, large-scale space tourism, space mining, space-based solar power, and interplanetary colonization. Not all of these potential ventures may be viable, but if any of them are, Starlink is the bridge.