SDA’s Nudge Exposes Legacy Challenges

In February 2022, the Space Development Agency (SDA) awarded three contracts to three companies: Lockheed Martin, Northrop Grumman, and York Space Systems. Each company will manufacture 42 satellites and have them ready for launch by September 2024. York Space is pretty much a company purpose-built for large-scale smallsat manufacturing. But at the time, I wondered just how realistic the expectation was for companies not used to quickly mass-producing satellites–specifically Lockheed and Northrop–to meet that deadline. Especially since each one of the satellites they manufacture must have the following:

• Optical Communications Terminals (OCTs) enabling a minimum of four simultaneous optical communications links,
• Link 16 mission communications payload,
• Ka-band RF mission communications payload, and
• BMC3 module enabling on-orbit data processing, storage and fusion.

Each satellite is a military communications hub, like SpaceX’s Starlink, but integrates legacy and battle management systems. Together, the 126 satellites form a modest (when compared with the commercial companies) communications network for U.S. combatants and their systems to tap into. The optical terminals should provide high speed and decent bandwidth between satellites, while satellite-provided Link 16 communications allow combatants and military systems to network through the satellite.

Why Subcontract?

Highlighting those requirements demonstrates that the satellites aren’t simple, single-payload systems. Like nearly all military DoD satellites, they are complex, usually leading to schedule violations and increased DoD spending. Lockheed and Northrop are old pros at that game, notorious for dragging out military satellite manufacturing to the point that a program could be a person’s entire career. Still, they have many resources, including their satellite buses. Will they adapt to SDA’s schedule requirement? Can they?

The answer appears to be: they seem to be trying. Maybe.

Lockheed can tap into its range of LM series satellites insofar as having access to company resources. The sizes that might have worked for the contract’s purposes would have been the LM 50 and the LM 400. However, the LM 50 might be a scosche too small, and one of the biggest showstoppers for the LM 400 would be the time required to build a satellite using that bus. From the site:

“The LM 400 bus can be delivered as quickly as 24 months from order and requires minimal engineering for payload integration.” (bold is mine)

It takes Lockheed Martin two years (on the optimistic side) to deliver just the LM 400 spacecraft bus. That length of time alone is probably why Lockheed Martin subcontracted the work to Terran Orbital instead of building satellites in-house. Possibly pertinent: Lockheed invested in Terran Orbital over five years ago. But Terran Orbital seems to be working with cubesat-inspired design, which limits crucial capabilities, such as power generation and batteries, necessary for optical terminals. Whatever the reason, the upshot regarding Lockheed’s capabilities for building satellites–large or small–in its factories quickly and within the SDA’s budget is that it can’t.

Northrop appears to be in a similar position, even though it inherited a decent “smallish” satellite bus through its Orbital ATK acquisition–LEOStar. It even looks to be sized just right for the SDA Tranche 1 satellites; however–the company’s brochure notes similar time constraints to Lockheed’s. At best, LEOStar would be ready after 24 months, not including the time required for integrating the mission payload. Also, Northrop appears to be downplaying the existence of LEOStar, as it’s not displayed on its site between the company’s GEOStar and ESPAStar offerings.

LEOStar’s manufacturing time and seeming retirement are probably why Northrop contracted Airbus U.S. last week to manufacture 42 satellite busses for the SDA’s needs. Airbus will use a version of the Airbus OneWeb Arrow bus (the platform used for the first generation of OneWeb broadband satellites). The version, identified in job listings as D2M Arrow-450, offers a little more flexibility in that it can host more mass–in the range of 300-500 kg.

While the Arrow-450’s mass capability is essential, what may have attracted Northrop to Airbus even more, is how quickly the company manufactures OneWeb satellites, at two satellites per day. That’s two completed, ready-to-launch satellites rolling off the factory floor daily. And the company may be able to double that production, as it initially advertised the ability to manufacture four satellites in a day. Moreover, unlike Terran Orbital or York Space, Airbus OneWeb has manufactured and deployed hundreds of its satellites. That might give Northrop some justified confidence in Airbus’ offering and the company’s ability to meet the SDA’s deadline. In addition, it will undoubtedly give Northrop some design and project schedule flexibility.

Customer-driven Changes

The deal is good for Airbus as it appears the company is required to merely crank out satellite busses and then let Northrop deal with satellite payload integration. But neither Lockheed nor Northrop have the in-house ability to build small satellites quickly. So instead, both went to external companies with a history with small satellites. That is a change in company behavior, even if the motivation is to win a government contract. Does it also beg the question of what in the company proposals gave the SDA enough confidence to choose them since neither intended to use in-house resources?

The behavioral change seems to count on the SDA offering contracts for companies to fulfill for a long while. If so, that seems to be incentive enough for companies like Lockheed and Northrop to somewhat get out of their project comfort zones. As noted in an earlier analysis:

“U.S. satellite manufacturing companies like Ball, Boeing, or Lockheed Martin design and build satellites in years. One of their biggest and most profitable customers, the U.S. government, appears very comfortable with that pace (and ULA and NG are happy to accommodate). Their commercial customers are also satisfied with the slow pacing, but whether this is because of low schedule expectations is unclear. This situation makes it pretty clear not to expect shifts to lower-cost and faster satellite manufacturing from the legacy companies, even though their interests are served in catering to a larger commercial market.”

“Very few established companies are willing to change their culture and processes like that.”

“However, new satellite manufacturers will eventually leap in and take advantage of the slow movers, opening the satellite market to more potential customers. They already are doing so, in the form of cubesat/smallsat manufacturers and satellite operators who also build their satellites in-house. Some are even manufacturing slightly bigger smallsats, possibly in response to SpaceX’s low cost per kilogram rideshare pricing. SpaceX can manufacture more satellites in a month than some legacy companies do in two-three years. Ultimately, it will take some time for the advantages offered by SpaceX’s Falcon 9 to trickle down.”

In this case, smallsat manufacturers are leaping in, but SpaceX isn’t the one that’s pushing this change. Indeed, it’s not pushing the change as much as I thought it would (although it is working on a different SDA contract). However, a U.S. government agency, the SDA, is. While the contracts the SDA offers are minor, those still appear to incentivize the legacy space companies to, if not change, then at least become familiar with the small satellites and speedy manufacturing times that companies such as Terran Orbital and Airbus bring.

The SDA appears to understand the significant weakness that traditional DoD acquisitions schedules and requirements impose on the warfighters’ missions. The new timelines and the inclusion of a few new companies is heartening. However, it’s a little strange that the large (and theoretically more capable) companies are merely intermediaries (although it appears Northrop will be dealing with the satellite payload).

Should the U.S. military be alarmed about this situation? After all, an established European company, Airbus, working with OneWeb, is using a capability that its large U.S. military-industrial counterparts don’t have: satellite mass production. Maybe that deficiency is something to worry about, or maybe it’s how things work in today’s environment. Some people have suggested that startups are larger legacy companies' new “research and development” branches. If the startup, such as Terran Orbital or York Space, successfully builds a magnificently-functioning satellite, the larger company could swoop in and add the startup to its collection. It’s a business model the U.S. tech sector is notorious for.

If that is occurring in the U.S. space industry, then the space startups are hoping to be bought out by the space equivalent of an Oracle or IBM, not a Google or Facebook. Unfortunately, that may not be a great outcome. The legacy companies involved, Lockheed and Northrop, aren’t known for the nimbleness and risk acceptance that Google or Facebook exhibit. Moreover, Northrop seems to be underusing its Orbital ATK acquisition, which perhaps bodes ill for any new space startup the company buys.

In the meantime, however, is the fact that a U.S. government customer is pushing a necessary change, encouraging legacy companies to hop to. Depending on the outcome of the subcontracts, it may be the legacy companies will buy into the startups (although Northrop will have to find a company other than Airbus)--or the SDA just skips the middleman altogether.

How much quicker could Airbus field satellites for the SDA, complete with its version of a communications hub if Northrop weren’t the middleman?