AINEWS
Twelve companies are building competing satellite interceptors in parallel. None of them can judge the contest.
On July 31, the U.S. Space Force handed a coordination problem to Johns Hopkins Applied Physics Laboratory (APL), a university research center that doesn't sell weapons. APL will be the independent technical direction agent for the Space-Based Interceptor (SBI) program, the most expensive single piece of the Pentagon's planned Golden Dome missile defense architecture. Its job is not to build interceptors. It is to write the technical interface documents that decide whether the twelve companies designing them can make their systems work together.
Twelve companies are running parallel prototype efforts under Other Transaction Authority (OTA) contracts, the procurement tool the Defense Department uses to fund fast, non-traditional vendor work. Twenty of those agreements, worth roughly $3.2 billion in aggregate, were awarded in late 2025 and early 2026. The choice of APL is not ceremonial. Each team is free to design its own orbital kill vehicle, its own sensor package, and its own ground software. None of them are required to make those designs compatible with anyone else's. Without a referee that owns the interface, the program is twelve prototypes that cannot interoperate. With one, the program is a network built by a dozen different vendors.
The competing field is heavy with primes that are themselves gunning for the production contract: Lockheed Martin, Raytheon, and Northrop Grumman each have an SBI prototype in the field. APL is the only institution that fits the seat. Asking any of them to police the interface documents they would also have to obey is a conflict by design, not by accusation. APL is a university-affiliated research center, not a defense contractor, and it doesn't build interceptors. The Space Force picked it because it can own the rules without having a stake in the winning design.
The Space Systems Command (SSC) runs cost, schedule, and contract decisions on SBI. The technical direction agent role is narrower: APL writes and owns the interface specifications, including the data formats, the physical and logical link definitions, the test and verification protocols, and the change-control rules that govern how an interface can be edited once prototypes start talking to each other. In a multi-vendor architecture, the document that defines how systems connect is the program. The companies are now designing to a specification APL controls, and they can't edit it without going through APL.
Commercial primes cannot be neutral because they are competitors, and parallel competition creates a coordination problem the acquisition structure cannot solve on its own, the source analysis notes. SBI's OTA field was set up to maximize parallel prototyping speed, which is exactly the property that produces incompatibility by default. The technical direction agent is how SBI tries to fix that. The Space Force is running the program the way an internet standards body runs a protocol: separate the referee from the implementers, give the referee ownership of the interface, and let vendors compete above the line.
The gating constraint is no longer who builds the best single interceptor; it is whose design can be made to interoperate cheaply with the other eleven. A team that can't interoperate doesn't get bought, no matter how good its kill chain is. The vendors who win will be the ones who can document their interfaces cleanly and accept APL's revisions quickly.
APL's mandate is narrower than the coordination problem. The lab owns the interface documents. It doesn't own launch, ground software procurement, or the classified data formats that come from the intelligence community. If those layers aren't aligned to APL's interface, the network still won't come up. The first multi-vendor integration test in 2027 will be the first public proof: if it succeeds, the program is a network. If it doesn't, the Space Force owns twelve prototypes with a shared stamp.