The Beginning of the End (of Expensive Satellites)?

A few days ago, the Space Development Agency (SDA) awarded a contract (one of two) to SpaceX. The company is now on the hook for getting four Tracking Layer satellites (using SpaceX’s Starlink bus) ready for the SDA by September 2022. The SDA is paying SpaceX $149 million to do this. The fact that overhead persistent infrared (OPIR) incumbents didn’t get selected is interesting. That fact is why legacy defense satellite manufacturers should pay attention to this government decision.

I’ve already conducted a cursory history and capability comparison overview yesterday. This analysis is all about the possible impacts of SpaceX’s presence in satellite manufacturing.

I genuinely think SpaceX’s entry into the military satellite realm should make other defense satellite manufacturers take notice. Consider who did not receive a contract for building a satellite with an infrared payload: Lockheed Martin (who manufactures and operates the Space Force’s Space Based Infrared System), and Northrop Grumman (manufacturer and operator of the Missile Defense Agency’s Space Tracking and Surveillance System). I am not privy to whether either company put their hat in the Tracking Layer contracts ring.

Why Should They Worry about SpaceX?

SpaceX has manufactured and operated more satellites than any U.S. legacy defense space contractor in an astonishingly (for the space industry) short time. It has fielded a communications constellation of over 750 satellites at an estimated cost ($233 million) that is much less than a single DoD satellite (such as $605 million for a WGS satellite). Some details:

• Monthly Starlink satellite manufacturing rate: 120
• Starlink satellites deployed: 776
• Average monthly Starlink deployment since May 2019: -46 satellites/month
• Estimated per satellite cost: ~$300,000 (probably less)
• Mass: 260 kg

Based on those statistics, we can see that SpaceX can manufacture satellites quickly. It’s a capability none of the legacy defense contractors have (if they did, we would have heard about it).

More distressingly for these contractors, SpaceX manufactures satellites inexpensively.  I don’t believe I’ve ever seen a news story touting the low costs of a missile warning satellite manufactured by the old guard. Typically, they are in the news because the Government Accountability Office is worried about how much more these satellites cost than projected.

SpaceX also has operational control of over 750 satellites in orbit. Not one of the other contractors has this experience.

If SpaceX successfully manufactures SDA Tracking Layer satellites at slightly over $37 million per satellite, and the SDA is happy with the result, it should worry the legacy defense contractors (and maybe U.S. taxpayers, too. More about that later).

It's only natural for SpaceX to branch out from manufacturing its Starlink satellites, and offer that capacity to interested customers.

And, yes, these are smaller satellites than what's typically used for communications birds in geosynchronous orbit. But these aren't cubesats, either. The Starlink bus may be good enough for most customers. The company will target other DoD customers. It will be only a matter of time for those customers to embrace the benefits of selecting SpaceX. Legacy defense providers will be left holding their...exquisitely expensive satellites.

And there’s reason to believe, based on SpaceX’s launch industry impacts, that the Starlink satellites themselves will become cheaper. While all this won’t happen tomorrow, it will happen faster than competitors anticipate.

Higher Pricing (with the Possibility of Going Lower)

So, why did SpaceX ask for $149 million for four satellites instead of $1.2 million? One reason would be development costs. Not the development of the satellite bus or infrared sensor payload, but for optical inter-satellite links (OISLs--or laser communications). OISLs are an essential capability for SpaceX’s plans (or any truly viable LEO broadband company’s plans) with the Starlink constellation itself. Since the SDA is interested in this capability and willing to pay for it, why not get involved?

Or, instead of developing, maybe use the money to buy ready-made OISLs?

OISL systems, such as those manufactured by Mynaric, are estimated to cost around $2 million. The cost alone may be a reason for SpaceX not to buy from existing manufacturers: their terminals cost more than Starlink satellites (six Starlinks, in Mynaric’s case). Additionally, each laser terminal (a satellite needs more than one) has a mass around 20 kg and requires about 60 watts to operate. Does a Starlink satellite generate enough power for multiple OISLs and an infrared payload? Or will SpaceX redefine requirements for laser communications between satellites?

Another reason for extracting a higher cost from the SDA is the complexity an infrared sensor brings to the satellite and satellite operations. The payload itself, since it’s outsourced (probably to a defense company), is likely the highest-cost component on SpaceX’s Tracking Layer satellite. Some mass trades vs. capability will need to be done. SpaceX’s schedule may be a risk if the external supplier cannot supply a quality product for four satellites on time.

The sensor is not like an infrared sounder on a weather satellite. It’s more--capable? The sensor’s mass might be small, but it probably needs to be cooled. It may require more energy to operate. It may require a sunshade. Does it have a computer for onboard processing of infrared information? Or will it only pass along raw infrared data to the ground? For a more responsive warning and tracking constellation, I suspect the former is desired.

The infrared payload is also very different from what SpaceX is using for its Starlink constellation. The differences between a passive communications system and an active infrared sensor implemented as part of an automated kill chain aren’t insurmountable. But those differences might be costly. The former probably requires communications and satellite monitoring and updates, whereas the latter will need people in the loop (and for a loop to be there).

In the end, the question surrounding the integration of OISLs and infrared sensor to a Starlink bus is: is there enough overhead in the mass a Starlink bus supports to change it to a warning and tracking satellite?

SpaceX believes so.

Ultimately, it’s **only** $149 million for four satellites--a bargain in the DoD space acquisitions world. But, even that bargain implies a "gotcha" for taxpayers.

If the SDA went only with SpaceX's Starlink for the next Tranche--and I'm not saying it will--the contract would be in the billions.  This implies that no matter who the SDA chooses to manufacture its satellites, the value of the current contracts awarded for developing Tranche 0 satellites doesn't seem sustainable.

Using my limited math skills--multiplying 200 satellites at $37 million each comes to...$7.4 billion?? Just for the Tracking Layer, which is one of the smaller parts of the SDA's architecture? Can a nation dealing with a pandemic's economic impacts even consider paying that? Even if the satellites are cheaper than what the DoD usually purchases?

That question aside, the current award means that SpaceX will get an OISL while fulfilling SDA requirements. That laser will be useful for its Starlink LEO broadband constellation and give SpaceX an edge against competitors who are already behind. Since that OISL system will be compatible (per SDA requirements) with the OISL’s on the Tracking Layer satellites, Starlink might serve as a useful backup to the SDA’s Data Transport layer satellites.

At the same time, SpaceX gets a toe in the door of satellite defense contracting, beyond which is a room filled with companies who bill high, aren’t competitive, don’t like change, and move very slowly. They’ve seen SpaceX’s impact on launch companies that exhibited the same traits: expensive, non-competitive, and slow. Legacy defense satellite manufacturers should be concerned.