Private Defense Tech Is Marked at 24–32x Sales; Its First Public Print Cleared at 4x

Anduril doubled its private valuation to $61 billion in a $5 billion Series H on May 13, 2026, barely twelve months after a mark near $30.5 billion. By July 24, investors were discussing a round at roughly $100 billion — a figure that would make the company worth more than Northrop Grumman's entire $78.3 billion market capitalization, and would price it near 45 times its 2025 revenue. Helsing went from $12 billion to $18 billion in thirteen months, and Shield AI from $5.3 billion to $12.7 billion in a year. Every one of those marks was set by the price of a new round, not by a public clearing price.

Capital is following those marks. PitchBook counted a record $24.4 billion of defense-tech venture funding in the first quarter of 2026 and $16.4 billion in the second, still among the largest quarters on record, and Dealroom expects European defense startups to raise a record $10.5 billion in 2026, with more than 85% of that coming from a handful of mega-rounds. Mega-rounds are what set the marks, and mega-rounds are what leave investors holding the most expensive paper.

Lyntris is the cleanest available read on what a sponsor-owned defense-tech roll-up is worth in public. The Falls Church, Virginia company, bought and reshaped by Trive Capital, reported $241 million of revenue and a $13 million net loss for the first half of 2026, and about $451 million of revenue with a $12 million net loss for the twelve months to June 30. It priced its IPO on August 18 at $17.50 a share, below the marketed $19 to $22 range, and closed its first session at $15.01. By October 26 the stock was $10.28, a market capitalization of $1.19 billion against the $2.53 billion the company had targeted — 41% below the IPO price within two months and 53% below the valuation it had marketed.

The ladder is not arbitrary; it tracks what a buyer can verify. Lyntris sells engineering, integration and sustainment, and clears at roughly four times sales. Rocket Lab sells hardware nobody else can build quickly, carried record backlog of $2.36 billion at the end of the second quarter of 2026, up 137% year over year, and clears at 38.3 times sales. The market pays for verifiable scarcity, not for the defense-tech label.

That distinction is why the private marks look expensive rather than heroic. Anduril's pitch — software-defined defense sold like a platform with recurring revenue — is priced as though contract revenue behaves like subscription revenue. On the same forward-revenue arithmetic, Shield AI's $12.7 billion is about 24 times the more than $540 million it projects for 2026, and Helsing's $18 billion is roughly 32 times the $502 million forecast for 2026, while the listed verdict on comparable contract revenue is two to four times.

If the private marks were right, the listed companies exposed to the same end-demand should have re-rated upward alongside them. They did the opposite — the same end-demand pulled the private marks up and the listed proxies down, which is the strongest evidence that the marks and the market disagree about the same revenue stream.

The pattern is the point: the de-rating landed on the listed names carrying the private defense-tech story — drones, space hardware, propulsion — while the primes with contracted backlog held their multiples and raised guidance. Note the asymmetry in cause: AeroVironment's problem was contract structure, not demand, and Kratos Defense's problem was the multiple, not orders. Public investors have already voted on this story, and they voted no.

The money will be spent; the question is who converts it and at what multiple. The appropriations calendar is the hinge. Congress is operating on a continuing resolution funding the government through December 11, 2026, signed September 2 after a 370-48 House vote and a 90-6 Senate vote, with full-year defense appropriations still unresolved in the Senate. Continuing resolutions slow new contract starts, and that is the mechanism that separates the two pricing systems: primes convert signed backlog while growth-multiple companies need new starts.

The transmission from private marks to listed stocks runs through two channels that are easy to miss: suppliers who built capacity against private order books, and listed proxies that carry the private multiple. Primes sit on neither channel.

Two of those links are the honest warning lights. Voyager Technologies has put more than $500 million into propulsion capacity on the strength of private defense-tech demand; if the funding cycle turns, that capacity meets a smaller order book, and capacity built against private order books is the first thing written down. Rakuten's brokering of Helsing into Japan matters more than it looks, because the first test of an $18 billion mark runs through a foreign customer's order book, not a domestic budget line.

This call is wrong if the private prints start working. An Anduril listing at $100 billion, a Japanese order for Helsing's HX-2, or a Shield AI round at $20 billion would each show that the private market has found a clearing price the public market has not seen yet. Until one of those prints, the most recent verified public price for a sponsor-owned defense-tech roll-up is $10.28 a share, against private marks of 24 to 32 times revenue.