INFLCT ← All theses

DEFENCE / AUTONOMY

The incremental defence dollar is shifting from primes to drones and counter-UAS

60/100 conviction
▲ +2 up

Last reviewed 15 July 2026 · our own structural assessment, not a price target

The thesis

Defence spending is not just rising — its composition is shifting at the margin from legacy primes (manned platforms, exquisite missiles) toward attritable drones, autonomy and counter-UAS, a lesson written in Ukraine. The direction is clear (a first-ever US dedicated autonomy budget-request line, record defence-tech VC, large counter-drone awards), but the shift is still small in absolute budget share and the sharpest pure-plays are private, so listed exposure is indirect.

The case for

Defence-spending composition is shifting toward attritable drones, autonomy and counter-UAS. The FY2026 US budget REQUEST (DoD rollout, 26 Jun 2025) broke out autonomy as a dedicated line for the first time — $13.4B, of which $9.4B for aerial drones — plus $3.1B for counter-UAS (DefenseScoop). The FY2027 request goes further, seeking more than $70B for drones and counter-drone systems (about $53.6B for autonomy/platforms and $21B for munitions/C-UAS), as reported by DefenseScoop and USNI News on the FY2027 budget documents. Ukraine was on track to receive about 3 million FPV drones in 2025 (Ukraine MoD / Defence Minister Denys Shmyhal, 24 Dec 2025); Ukrainian government figures put total drone output of all types near 4 million in 2025 with a stated 2026 target above 7 million. Private neo-primes are funded at scale: Anduril raised $2.5B at $30.5B (Jun 2025) then $5B at $61B (May 2026), with reported revenue roughly doubling to about $2.2B in 2025 (TechCrunch/Forbes; Anduril is private). Defence-tech VC set a record at $19.8B across 262 deals in Q1 2026 (PitchBook). Counter-UAS is becoming a hard procurement line — AeroVironment won a three-year $500M IDIQ ceiling supporting JIATF-401's "Domestic Shield" homeland C-UAS program, with a first $80.5M task order for its Titan multi-sensor system (AeroVironment, 6 Jul 2026).

The case against

The shift is real at the margin but still small in absolute dollars, and the cleanest exposure is not in listed equities. Legacy primes still capture the great majority of DoD contract obligations, and the autonomy/C-UAS budget lines remain a low-single-digit share of the total defence budget — a derivative bet, not yet a level bet — while prime order backlogs are still growing. The sharpest pure-plays (Anduril, Shield AI, Saronic) are private, so listed exposure is indirect and bundled. Valuations look stretched: Anduril CEO Brian Schimpf said he thinks there is "a bit of a bubble" in defence tech (Fortune Brainstorm Tech, Jun 2026), with Anduril itself at roughly 28x reported 2025 revenue (Forbes), and PitchBook's median venture-growth pre-money valuation nearly tripled to about $1.4B in Q1 2026 from about $500M a year earlier. The flagship Replicator program underdelivered and was absorbed into the Defense Autonomous Warfare Group (DAWG) in late 2025 (Washington Times / Breaking Defense). Two commonly overstated figures, corrected: Germany signed a framework contract for 19 Rheinmetall SkyRanger units (18 serial vehicles plus one demonstrator) in February 2024 (Bundeswehr / Rheinmetall), with a further 500-600 an anticipated Berlin planning requirement, not yet a signed contract; and NATO's Hague commitment is at least 3.5% of GDP for core defence plus up to 1.5% for broader security by 2035 (NATO Hague Summit Declaration, 25 Jun 2025), not a flat 5%.

What would change our mind

Re-scored 58 -> 60 on 2026-07-14 after a fresh source and counter-case review. The directional evidence held up and strengthened: the FY2027 request seeks more than $70B for drones and C-UAS (reported by DefenseScoop/USNI on the FY2027 budget documents) and the Replicator initiative was consolidated into the much larger DAWG autonomy program — both point to composition shift outrunning the "still tiny share" framing. Corrected material errors before publication: the "bubble" quote is Anduril CEO Brian Schimpf, not founder Palmer Luckey; FY2026/FY2027 figures are budget requests, not enacted; Ukraine's ~4M is all-types (the ~3M official figure is FPV delivered); Germany's SkyRanger count and NATO's headline 5% were overstated. Removed unsourced prime-obligation and backlog figures that could not be reproduced from a primary source. Held at 60 because the cleanest exposure remains private and VC valuations show froth.

Where the exposure sits

Where the exposure sits: the sharpest pure-plays are private — Anduril (~$61B reported valuation, May 2026), Shield AI (~$12.7B reported) and Saronic (~$9.3B post-money, Q1 2026 per PitchBook). Listed exposure is indirect and bundled: AeroVironment (NASDAQ:AVAV — layered C-UAS and loitering munitions), Palantir (NASDAQ:PLTR — autonomy software/C2), and legacy primes with drone/counter-UAS lines such as RTX (NYSE:RTX) and L3Harris (NYSE:LHX). European counter-drone exposure runs mainly through Rheinmetall (ETR:RHM, SkyRanger).

Show our work — the evidence

The key evidence behind the score, dated and sourced — supporting and disconfirming. This is how the conviction is built, and what we re-check as it moves.

Tags mark how each fact is sourced: Company disclosure (the company said it) · Official data (a government or agency figure) · Analyst estimate (a third party's number) · Reported (via the press).

Get each thesis as it moves.