Starlab is a joint venture, and Voyager controls 61.9% of it
7 min read
Who do you have a claim against, contractually, on a space station built by five different partners and a $217.5 million US government contribution? And how much does the strictly military side of the work weigh in the accounts of the company building it? A filing dated 14 August 2026 with the SEC answers the second question with less precision than before, and the first with more precision than usual.
The facts, from the filing
Voyager Technologies, Inc. (NYSE: VOYG, CIK 0001788060) has filed a Form 8-K (accession number 0001628280-26-057100) with Exhibit 99.1 attached, under Item 8.01 (Other Events) and Item 9.01: “to recast its consolidated financial statements and historical segment disclosure previously reported in the Company’s Annual Report on Form 10-K for the fiscal year ended December 31, 2025”. The reason: “the Company combined its Defense and National Security and Space Solutions segments into a single Defense and Space Technologies segment in order to align with how the Company’s Chief Operating Decision Maker (“CODM”) views results.” A third segment, Starlab Space Stations, stays outside the merger.
The filing is filed, not furnished: Exhibit 23, the consent of PricewaterhouseCoopers LLP, confirms it — filed to incorporate the recast figures by reference into the Form S-8 registration statements covering Voyager’s employee equity plans. It neither amends nor restates the audited financial statements: it exists to make the three fiscal years comparable under the new structure.
Two segments, one figure that disappears
The March 10-K listed three segments: Defense and National Security, Space Solutions, and Starlab Space Stations — and it was explicit about how much the first one weighed: it “represented approximately 72.1% and 50.9% of our net sales for the years ended December 31, 2025 and 2024, respectively”. A swing of 21 points in a single year, legible only because the segment existed on its own.
| segment (net sales, $ thousands) | 2025 | 2024 | 2023 |
|---|---|---|---|
| Defense and National Security | 122,954 | 77,470 | 63,154 |
| Space Solutions | 47,583 | 74,593 | 76,771 |
| Defense and Space Technologies (post-recast) | 169,658 | 151,224 | 139,925 |
| Starlab Space Stations | — | — | — |
From now on, only the third line survives, and the reconciliation is exact: the first two sum to $170,537 thousand in 2025, the single segment reports $169,658 thousand, and intersegment eliminations fall from $4,118 to $3,239 thousand — the very same $879 thousand. Total net sales are unchanged at $166,419 thousand. Those sales between the two old segments have not vanished: they became internal, and are no longer counted either as segment revenue or as an elimination. The 2025 in which defence made up 72% of revenue and the 2024 in which it made up 51% become, going forward, one composite figure.
The joint venture, in the numbers the filing gives
On the second thread, Exhibit 99.1 is unusually precise: “We and Airbus have entered into the Joint Venture Agreement forming Starlab JV […] Following the formation of Starlab JV, each of Mitsubishi, MDA Space and Palantir acquired minority interests in Starlab JV. As of December 31, 2025, we had an ownership interest of 61.9% in Starlab JV. Voyager may nominate three directors and Airbus may nominate two.” Voyager contributed “certain contracts, intellectual property and $11.5 million in cash”. The public contribution remains the one already known: the NASA grant worth $217.5 million, with $34.3 million still to be disbursed as of 31 December 2025.
One detail worth a close read: the same Exhibit 99.1 is not internally consistent on names. Early on it lists five “international equity partners”: Airbus, Mitsubishi, MDA Space, Space Applications Services and Palantir. Later, describing the JV’s “equity partners”, it names only four — leaving out Space Applications Services — without explaining the difference. If the primary document itself is not consistent about who holds a stake, the press coverage that follows it is unlikely to be either.
Palantir also shows up outside the JV’s governance: within the Defense and Space Technologies segment, Voyager describes “AI-powered edge computing units layered with Palantir’s operating system”. Two distinct relationships under the same name: a minority shareholder in a subsidiary, and a technology supplier to the parent company.
On customer concentration: “Sales to the U.S. government accounted for 86.0%, 83.9% and 69.0% of sales during the years ended December 31, 2025, 2024 and 2023, respectively”; and “approximately 86.3% of the total dollar value of our funded backlog related to our top customer, the U.S. government.” Starlab Space Stations generates no revenue yet: the government dependency sits entirely inside the merged segment.
See the service · Talk to an engineer
The point
There is nothing irregular about this filing. Merging two segments when the CODM’s view of results changes is an ordinary accounting choice, and the recast exists precisely to make the fiscal years comparable — the 8-K says so, line by line. The point is not the compliance of the move: it is that the granularity of public disclosure depends on how management looks at its own numbers, not on what the reader needs. Until yesterday, one could read separately how much strictly military work weighed against general space services. From today that line exists only for the three years already closed; going forward, it is one figure.
The second thread runs alongside it: Starlab is not a Voyager project, it is a joint venture with equity partners — Airbus, Mitsubishi, MDA Space, Palantir, and, according to some passages of the same document, Space Applications Services — built in part with a US government contribution. Whoever buys capacity on that station will be contracting with Starlab Space LLC, not Voyager Technologies, Inc. The disclosed 61.9% says who is in control today; it does not say whether that will hold, or what happens to the other partners’ rights — or a future customer’s — if the stake shifts.
What we do not know
We do not know the Starlab JV’s shareholder agreements, nor the exact stakes held by Airbus, Mitsubishi, MDA Space and Palantir: the filing gives only Voyager’s 61.9% and the board seats; the rest is “minority interests” with no percentage attached. We do not know what role Palantir plays in the JV’s governance beyond the equity-partner label, and we do not infer one from the separate technology partnership within the Defense and Space Technologies segment. We do not know why Space Applications Services appears in one list and not the other within the same Exhibit 99.1: we flag it without interpreting it. We found no press release dedicated to this reorganisation: that is not proof one does not exist, and it is consistent with the nature of the filing — an accounting adjustment, not a market announcement.
The two axes, applied
Comply. In the register of critical dependencies, for every capability bought from a third party, the line that is almost always missing is which legal entity you can actually enforce a right against, and who controls it — not the brand on the contract, the company that signs it. For each supplier: jurisdiction, the ownership structure declared in public records, and what happens on termination — the same gap measured when a critical-data supplier changed owner overnight or when the name on a fund did not match who signed it, with a dated record ready to show an inspector or a board.
Decide. The same system unifies contracts, suppliers, archives, business systems and documents into a single operational model on which AI agents execute decisions with a human operator in command, for large enterprises, defence, the public sector and healthcare — the same logic we apply reading a supplier’s government revenue concentration, relevant to anyone in the defence sector. Always two modes: on-premises, on self-contained machines that require no deep integration into the client’s network, or a dedicated cloud with a dedicated VPN and a data centre in Italy; always shared management.
From the first session, at no cost, comes the dated list of the critical capabilities you buy from third parties: for each one, which legal entity owes it to you, who controls that entity, and what public record exists on them — including the boxes that stay blank. It stays with you even if we do not go on to work together. Talk to one of our engineers about it.