Cheap, Many, and Armed: YC's Defense Cohort Has Picked a Side in the Attrition Economy

The wave right now

For a decade, "defense startup" meant one of two things at YC: a dual-use company that whispered the word, or a hardware moonshot that died waiting for a contract. That era is over. The 2026 batches are funding companies that say the quiet part in their one-liners — Tenet Industries literally pitches "low-cost, mass-produced strike drones," and 9 Mothers is building an autonomous point-defense turret to shoot drones down. The organizing idea of this wave, imported directly from Ukraine, is attrition economics: cheap-and-many beats exquisite-and-few, and the winner is whoever can manufacture, coordinate, and counter at scale. The same logic is spilling into orbit, where the constraint has shifted from launch to *operations* — power, coordination, and getting things back down.

If you're a founder deciding what to build this quarter, the takeaway isn't "defense is hot." It's that two specific quadrants are already saturated, and the adjacent ones are nearly empty.

The landscape today

Attritable mass is a manufacturing thesis, not an airframe thesis

The strike-drone companies in this cohort don't compete on the vehicle — they compete on the production line and the autonomy stack. Tenet Industries (Stockholm + SF) leads with design-for-manufacturing and ready-to-scale platforms; Maquoketa Research is vertically integrating an all-USA, NDAA-compliant stack from terminal guidance down to propulsion; Seeing Systems bets on modularity plus an agentic control layer, with NATO customer relationships already in hand. Even the exotic end — Ornadyne's flapping-wing surveillance birds — is really a bet on evading the *other* side of this list. The pattern: the airframe is commodity; compliance, autonomy, and throughput are the moat.

P26

Tenet Industries

Low-cost, mass-produced strike drones

ActiveMass-Production Strike Drones4 ppl

P26

Maquoketa Research

The intelligence layer for drones

ActiveAutonomous Weapons AI5 ppl

W26

Seeing Systems

Modular AI-Commanded Drones for Defence

ActiveDefense Drones2 ppl

P26

Ornadyne

We're building robot birds for surveillance.

ActiveBiomimetic Surveillance Drones4 ppl

Every drone wave funds its own counter-wave — and this is the crowded quadrant

Counter-UAS is where the cohort is densest, and where the differentiation is thinnest. 9 Mothers is building the effector (the Edda turret for fast FPVs). Surtr Defense Systems is betting the layer above it — ParallaxOS, an open, multi-sensor fusion OS for Western counter-UAS networks. Milliray attacks detection with high-frequency radar tuned for small, low-signature targets, while Arlo Industries goes passive with a mesh sensing architecture for drones *and* missiles. Four companies, two adjacent batches, one kill chain. Notably, three of these (9 Mothers, Surtr, Milliray) are renamed companies — founders who pivoted into counter-UAS rather than starting there, which tells you how strong the pull of this market is right now. The one genuinely uncontested flank: AICE is alone in the cohort taking the same swarm logic underwater, defending subsea infrastructure.

P26

9 Mothers

AI mission systems

ActiveAutonomous Counter-Drone Weapons18 ppl

P26

Surtr Defense Systems

The open operating system for counter-UAS

ActiveCounter-UAS Defense Systems2 ppl

W26

Milliray

Technology to detect and track small drones

ActiveDrone Detection Radar3 ppl

P26

Arlo Industries

Passive aerial sensing mesh to track drones and missiles

ActiveAerial Threat Detection3 ppl

The quiet trade is the supply chain underneath all of it

While the platform companies fight for the same program offices, two companies are selling to *everyone* in the fight. Kyten Technologies (Starlink pedigree) builds custom aerospace-grade battery packs — the choke-point component every drone, satellite, and aircraft on this page needs and can't easily make in-house. GUILD is going wider, positioning as a "neoprime" rebuilding the defense supply chain itself, with vertical supplier integration as the moat. This is the picks-and-shovels position in a war economy, and it's strikingly under-populated relative to the platform layer.

W26

Kyten Technologies

Custom Aerospace-Grade Battery Packs

ActiveAerospace Battery Manufacturing2 ppl

S26

GUILD

Neoprime Reshaping Defense Supply Chain

ActiveAI-Driven Supply Chain2 ppl

Orbit is getting an ops layer

The space half of this cohort has stopped pitching launch and started pitching operations. Dispatch is building high-frequency cargo *return* — reentry vehicles and in-house heat shields — on the bet that space manufacturing needs a down elevator. Constellation Space is the control plane: an AI operating system for constellations scaling from hundreds to tens of thousands of nodes, where ground-based orchestration is already failing. Beyond Reach Labs attacks the power constraint with football-field-scale deployable solar, and AxionOrbital Space converts all-weather radar into optical-grade Earth observation with foundation models — the data product defense and insurance buyers actually want. Add General Astronautics' microgravity robotics and you have a coherent stack: power, coordination, labor, sensing, return.

P26

Dispatch

Satellites for manufacturing in space

ActiveSpace Manufacturing Logistics2 ppl

W26

Constellation Space

AI operating system for mega-scale satellite networks.

ActiveSatellite Constellation Orchestration4 ppl

W26

Beyond Reach Labs

Space solar arrays that grow to the size of a football field in orbit

ActiveDeployable Space Solar5 ppl

W26

AxionOrbital Space

Foundation models for 24/7 Earth Observation

ActiveRadar-to-Optical Image Synthesis2 ppl

The cohort signal

This is not one batch flirting with defense. Winter 2026 carries eight genuine defense/space companies; Spring 2026 carries nine more — including Tenet, Maquoketa, 9 Mothers, Surtr, Arlo, AICE, and Dispatch — and the theme keeps running through GUILD in Summer 2026 and Degla's autonomous drone orchestration in Fall 2026. Four consecutive batches, holding or growing in density, with the Spring batch the heaviest yet. That's a deliberate, accelerating program bet — YC is underwriting the attrition economy the way it underwrote dev tools in 2012.

Lessons from the last cycle

The prior cohort's record is a warning about timing, not market size. Airware (W13) built the enterprise drone platform a full decade before procurement was ready, and is gone. Volansi (W17) — itself a renamed company — got its autonomous VTOL logistics to an acquisition, not an outcome that justified the capital intensity. Momentus (S18) made it to the public markets as a space-infrastructure company, proving the category can exit while reminding everyone how punishing the path is. Geosite (W19) pivoted through spatial data to a quiet acquisition. The common thread: every one of these was early to a demand curve that only inflected after 2022. The current cohort doesn't have that problem — its risk is the opposite one: crowding into demand everyone can now see.

If you're building here

Openings worth your quarter:

• Go underwater or go orbital-return. AICE and Dispatch are each nearly alone on their flank while nine companies fight over aerial drones. Subsea infrastructure defense and reentry logistics have the same "why now" with a fraction of the cohort competition.
• Sell components and compliance, not platforms. The Kyten/GUILD layer is thin. Every NDAA-compliant subsystem — propulsion, comms, power, test and verification — is a choke point with a dozen captive buyers on this page alone.
• Build for Europe. Tenet's Stockholm footprint and Seeing Systems' NATO relationships are the tell: European rearmament wants non-US production, and almost nobody in this cohort is structured for it.

Tarpits to avoid, by name: another counter-UAS *detection* sensor — Milliray, Arlo, Surtr, and 9 Mothers already cover radar, passive mesh, fusion OS, and effector across two adjacent batches. And another FPV strike airframe without a manufacturing or compliance edge — Tenet and Maquoketa have made mass production the table stakes.

What you'd have to believe: that the procurement thaw outlasts the current conflict cycle; that attritable-systems demand survives a ceasefire; and that NDAA compliance and supplier integration are durable moats rather than taxes every competitor eventually pays. If you believe all three, the empty flanks above are the best risk-adjusted entries in this memo. If you only believe two, sell batteries.