Start with the cost line, not the diagram. On June 23, 2020, Lockheed Martin was granted US10689131B2, “Sectioned self-mating modular satellite buses” — a structural design classified in B64G 1/10, the spacecraft-construction art. The inventor of record, Raymond Edward Fraze, also appears on the earlier Vector Launch filings in the same family, which is itself a small story about where modular-bus IP migrated as the new-space shakeout played out.
A satellite bus is the chassis — the structure, power, and plumbing that everything else bolts onto. The claimed innovation is a bus built in sections that self-align and mate without elaborate fixturing. For a one-off exquisite satellite, that saves little. For a program building dozens of near-identical spacecraft, it attacks the single most stubborn cost in the model: touch labor during assembly, integration, and test.
Here is why a markets desk cares. In a constellation program, recurring cost per unit is the number that compounds. Backlog is priced against an assumed build cost; if assembly hours per satellite fall, gross margin on the program rises without a single new contract. A modular self-mating architecture is, in financial terms, a lever on recurring unit cost — the same lever automotive learned decades ago.
The disciplined caveat: a granted patent is not a shipping production line, and a manufacturing claim is only as valuable as the volume it is applied to. Lockheed builds many spacecraft as bespoke national-security assets where this kind of standardization is hardest to apply. The patent describes a capability; it does not prove the cost curve has actually bent.
But the direction is the read. When a prime patents around assembly-labor reduction rather than performance, it is telling you where it thinks the next margin dollar lives — in the factory, not the payload. For an analyst pricing a constellation backlog, that is the more durable story.