If you are a defense-tech investor or founder, the budget cycle just handed you something rarer than earnings season: five institutional confessions in one week, and a pace-mismatch that narrowed in five places at once. The Pentagon confessed in its FY27 request — $75B across drones and counter-drones, and a 24,000% raise for a SOCOM office most of the building could not find on an org chart six months ago. The Army confessed by buying 13,000 Ukrainian-tested $15K interceptors in eight days and walking them into the Gulf. Anduril and Shield AI confessed by hot-swapping their autonomy stacks on the same airframe in one sortie. Ukraine confessed by running 22,000 ground-robot combat missions in a single quarter. And the FCC confessed by asking, for the first time in a decade, whether its own rules are the problem. That last sentence is worth rereading: the pace-mismatch between autonomous-systems capability and institutional absorption just closed in five separate layers at once. It is the rarest signal this newsletter tracks, and most of the defense trade press is covering it as five unrelated stories.
The Pentagon's $75B FY27 Drone Ask Is a Quiet Verdict on the Exquisite-Platform Era — DAWG Gets a 24,000% Raise
The Pentagon filed its FY27 budget request on Tuesday, and if you know how to read it, it is the most expensive admission of doctrinal failure in modern procurement history. The top-line is ~$54B for military drones and ~$21B for counter-drone — numbers that would have been unthinkable in an FY23 request. But the line that actually tells the story is buried three layers down: the Defense Autonomous Warfare Group (DAWG), a small SOCOM office most of the building could not find on an org chart six months ago, goes from $225.9M in FY26 to a requested $54.6B — a 24,000% raise. Read that again: a 24,000% raise. Not a typo, not a total, not a misplaced decimal. And $53.6B of it is parked in the reconciliation pot rather than the base budget, which is the procurement equivalent of announcing dinner reservations before checking whether the restaurant exists. The "Drone Dominance" campaign keeps its 200,000-unit FPV target for 2027 beside it, almost as a footnote.
Program lines don't move by 24,000% unless an institution has decided something fundamental has shifted. This is the pace-mismatch closing at the doctrinal level: the Pentagon has decided that mass-produced, autonomous, low-cost systems are about to displace the exquisite-platform heuristic that has dominated U.S. procurement since the 1980s — and that DAWG, not the Air Force or Navy weapons programs, is the vehicle to buy them through. For allocators, two reads. DAWG-reconciliation is the single most fragile bet in the portfolio: $53.6B sitting outside the base budget means the real test is whether the HASC Emerging Threats Subcommittee and the SASC Emerging Threats and Capabilities Subcommittee actually report it out in markup, not whether the request holds as printed. More consequentially, the $21B counter-drone line — nearly half the offensive drone budget and the most under-covered category in the FY27 request — is a tacit admission that the U.S. faces, at home and in the Gulf, the same Shahed arithmetic it is helping Ukraine solve. That is the capital flow most likely to reward fielded-hardware incumbents and the first wave of low-cost interceptor startups.
The $15,000 Interceptor That Ended Patriot-Era Base Defense — in Eight Days
Army Secretary Dan Driscoll told lawmakers this week that the U.S. Army has deployed the Merops interceptor — a mobile, fixed-wing counter-drone system developed by California firm Perennial Autonomy (formerly Project Eagle) and combat-hardened in Ukraine since mid-2024 — to shield U.S. forces against Iranian Shahed-136 attacks in the Middle East. The Army purchased 13,000 units within roughly eight days of the Iran conflict beginning. Merops carries a 2-kg fragmentation warhead, uses onboard sensors for terminal homing, runs ~$15,000 per unit today, and is expected to drop below $10,000 at scale — against a Shahed-136 cost of $30,000 to $50,000.
This is the trend from Editions #4 and #6 compressed into a single procurement: Ukraine's battlefield knowledge pipeline is now feeding U.S. acquisition in days, not years — the pace-mismatch running in reverse, for once — and doing it for systems that field-strip air-defense economics. For the first time since the missile age, the defender's shot costs less than the attacker's warhead. That is the arithmetic that made Patriot-vs-Shahed unsustainable, and the arithmetic that makes interceptor swarms newly doable. Saudi Arabia and the UAE are already reportedly evaluating the same system. Expect a Gulf-state order inside the quarter, and expect Merops — not Patriot — to be the template for how the U.S. defends bases in the next decade.
Anduril's YFQ-44A Hot-Swapped Two AI Stacks Mid-Flight — and Turned CCA Autonomy Into an Interface Market
In newly disclosed testing, Anduril's YFQ-44A Fury — one of two Collaborative Combat Aircraft (CCA) prototypes competing for the Air Force's unmanned-wingman role — completed a single sortie running two different mission-autonomy stacks back-to-back. The aircraft flew its task set under Shield AI's Hivemind, then switched mid-flight to Anduril's own Lattice autonomy and repeated the objectives before returning. Separately, Experimental Operations Unit airmen at Edwards AFB ran a full mission cycle with Anduril's Menace-T command kit — two Pelican cases and a ruggedized laptop handling mission planning, autonomous takeoff, in-flight tasking, and post-flight data checks.
The CCA program was always sold on the premise that the U.S. would not lock in a single-vendor AI stack — a mid-flight hot-swap between Hivemind and Lattice is the flight test that proves the interface layer actually holds. That is a market-structure signal: the multi-vendor autonomy thesis is intact, which is good for Shield AI and Anduril as the current category leaders and bad for any single-vendor lock-in narrative that was quietly being priced in. It also lowers the barrier for a third or fourth autonomy stack to enter the program on merit. The Menace-T demonstration matters for a different reason: it shrinks CCA's operational footprint from a command center to a laptop, which is the difference between an exquisite program that requires a base and a deployable capability that ships with the fire team — and quietly opens a new autonomous-command-kit market tier underneath the aircraft itself. Both are prerequisites for CCA going from showpiece to line unit, and both are the specific thresholds that justify the DAWG-scale budget in Story 1.
Ukraine Quietly Reclassified Ground Robots as Infantry Substitutes Last Quarter — 22,000 Combat Missions, and NATO Hasn't Noticed
Ukraine completed more than 22,000 ground-robot missions in the first three months of 2026, up from roughly 2,000 across the previous six months, with March alone logging over 9,000 sorties, according to CNN's reporting on Zelensky's latest briefing. Missions span demining, medical evacuation, logistics resupply, and direct assault — including the first capture of an enemy position in the war executed entirely by unmanned ground platforms, with no human infantry involved. Ukraine's Ministry of Defense now plans to contract 25,000 additional ground robotic systems in the first half of 2026, roughly doubling all of 2025's procurement.
Edition #5 flagged the first all-robot position capture as a historical marker. The more consequential number this week is the rate of change: an 11× jump in ground-robot missions across adjacent quarters is not a scaling curve, it's a phase transition. Ukraine has stopped treating ground robots as experimental and started treating them as infantry substitutes — the same re-categorization aerial drones went through in 2023, and the clearest example this edition of pace-mismatch closing on the doctrinal side before the hardware side is even fully priced. The 25,000-unit half-year buy would be the largest ground-robot procurement in military history, and the signal it sends is that Ukraine's UGV manufacturing base has caught up to its aerial drone industrial capacity. Watch whether NATO armies start specifying UGV units in their 2026–27 force structure updates; the doctrinal lag will not be long.
Counter-Drone Authority Just Federalized Down to the County Sheriff — and the Drone World Is Still Treating It as a Policy Story
On April 1, the FCC's Wireless Telecommunications Bureau and Office of Engineering & Technology issued a Public Notice asking whether Commission procedures are delaying the development, marketing, or sale of counter-UAS systems authorized under Sections 227 and 1048 of the FY2026 NDAA. The Safer Skies Act provisions — now in force — extend DOJ and DHS counter-drone authority through 2028 and, for the first time, create a nationwide framework under which state and local law enforcement can be trained and equipped to disable unauthorized drones at major public events. FEMA's companion C-UAS Grant Program is disbursing $500M to state, local, and tribal agencies to stand up detection and interception capability.
Two trendlines converge into a single procurement event. On the spectrum side, the FCC's Notice telegraphs a willingness to rewrite certification and marketing rules for jammers and takedown gear — a real shift from the "narrow federal use only" posture of the last decade. On the enforcement side, the pace-mismatch just closed in the least-glamorous place possible: county sheriffs' offices now have legal C-UAS authority they did not have twelve months ago, with $500M in FEMA grant money behind them. That is a distributed procurement stream for state, local, and tribal buyers who, until this quarter, had neither the legal authority to buy C-UAS gear nor the budget to buy it with — and it is where the next wave of C-UAS revenue is going to land. Watch D-Fend, Dedrone, Anduril's C-UAS line, and Epirus compete for the early awards; watch the 2026 FIFA World Cup window (March–September) for the first high-profile intercept or friendly-fire embarrassment that shapes whether the authority gets made permanent in FY27 NDAA or re-litigated from scratch.
Also worth your attention this week
Can Congress actually stomach a $54.6B DAWG? — The FY27 request parks $53.6B of the Defense Autonomous Warfare Group ask in reconciliation, not the base budget, which makes it the single most politically fragile line in the drone portfolio. House and Senate Armed Services markups in the coming weeks will tell us whether the Pentagon's 24,000% bet has real appropriation appetite behind it, or whether DAWG becomes another Replicator-style promise that shrinks quietly between request and enactment.
Part 108 BVLOS rule — still no publication date — Our standing continuity thread. The final rule was supposed to ship by mid-March under the executive order; it didn't. Industry has now baked in a late-2026 publication with 6-12 month implementation runway. Every week of slippage is another quarter of waiver-era economics for Wing, Zipline, Matternet, utility inspection, and agriculture-at-scale operators — and another quarter in which the U.S. commercial fleet reaches for whatever the cheapest foreign airframe happens to be that month.
Crucible / Swarm Forge — vendor list incoming — Carried from Edition #6. The CDAO's April 17 white-paper deadline has closed; finalists get invited to the June 22-26 live Crucible event where heterogeneous four-plus vendor swarms have to execute "find, fix, finish" end-to-end autonomously. Expect the selected vendor list to surface in May. Whether non-traditional entrants (xAI, SpaceX-adjacent teams) make the cut against Shield AI, Anduril, and Palantir will be the most honest signal we get this quarter about the Pentagon's real AI bench.
Five confessions in one week is not the noise — it is the signal. We will be looking for the follow-through on Friday.