Two runs through the gates

1. Dec 2023Gate 1

   DASA Rapid Impact Open Call contract

   A public-sector R&D contract to develop a 12ft variant of its K4MANTA foiling USSV. No equity taken.

   Read it through the frameworkThrough the framework

   A contracting-clock innovation contract used exactly as the methodology prescribes: a cheap, fast assumption test that doubled as requirement validation. The call responded to maritime capability areas of interest the funder itself had published; the buyer literally printed the gap.

2. Jun 2025Gate 2

   Strategic round: NATO Innovation Fund + NSSIF + Superangel

   Strategic, sovereign and intergovernmental capital, invested while the company was already delivering scale production orders to NATO countries.

   Read it through the frameworkThrough the framework

   The sovereigns did not fund a promise; they formalised a procurement fact. NIF and NSSIF arrived only after production orders existed, which is precisely the recorded behaviour of the sovereign layer across the dataset.

3. Jul 2026Gates 3+4 fusedfusion moment

   $175M Series B at a $1B valuation, led by DTCP

   Rheinmetall and Inocea (Davie Shipbuilding) participating as strategics; British Business Bank and NIF returning; early backers NIF, NSSIF, SmartCap, Notion Capital and Speedinvest converting prior stakes.

   Read it through the frameworkThrough the framework

   The fourth lane at full power: Rheinmetall is simultaneously investor, manufacturing partner and channel, with Anduril and Davie diversifying the lane. The dual-lane doctrine executed precisely, sovereign backers and primes on the same cap table. One press release carried contracts, primes and the round together.

Between gates 2 and 4, the recorded proof: contracts from the UK MoD, NATO European partners and USSOCOM, platforms deployed in support of multiple ongoing conflicts, and manufacturing partnerships with Rheinmetall, Anduril and Davie Shipbuilding. The headline lesson is not that any single step was fast; it is that every proof-point was in phase, contract, then sovereign validation, then prime design-in, then consensus. No gate was pitched with the prior gate's proof.

1. Aug 2025Gates 2+4 together

   $62M at a $500M valuation, led by Sequoia

   With 8VC, Thiel Capital, NATO Innovation Fund, In-Q-Tel, Project A and Döpfner Capital. Product: the OWE-V Virtus strike UAV and the Minerva C2 system.

   Read it through the frameworkThrough the framework

   A tier-1 lead flanked by two sovereign-linked funds in a single round, barely a year after founding. Three compression enablers are visible: a repeat founder whose prior credibility substitutes for early gates, a white-hot category where the published capability gap needed no shaping, and Germany's budget shock compressing the budget clock. Compression is real but rare; it is what the top of the barbell looks like.

2. Feb 2026The anchorfusion moment

   Founders Fund joins; the valuation crosses EUR 1B

   Recorded alongside a potential EUR 536M German government procurement of its strike drones, in the same news cycle.

   Read it through the frameworkThrough the framework

   The methodology's central claim in one data point: the valuation is the procurement path, priced. Tier-1 investors did not underwrite a drone; they underwrote a German budget line converging on a company. Proof-point and capital event landed inside one window, the 90-day fusion rule executed.

3. Jun 2026Full consensus

   EUR 500M at a EUR 3.5B valuation

   Sequoia, Founders Fund, NIF, Project A, Air Street, 201 Ventures, Advent and Döpfner; more than 80% of proceeds committed to R&D and manufacturing.

   Read it through the frameworkThrough the framework

   Capacity capital raised after the procurement anchor existed, not before. Raising EUR 500M for factories before the EUR 536M signal would have been uninvestable; ninety days after it, it was oversubscribed logic. Sequence matters.

A dataset honesty note: Stark's history before August 2025 (any seed round, early Project A entry) is not captured in the dataset; the reconstruction begins where the records do. Case-study discipline is stating coverage limits rather than smoothing over them.