Air · Reusable VTOL interceptor
Banshee
If it does not fire, it comes back, refuels, and goes again. Nobody touches it.
Mission
Air and missile defense is an arithmetic problem. Raids arrive faster than magazines reload.
The cheapest way to add magazine depth is not to buy more rounds. It is to make the rounds you already own generate a second sortie inside the same raid. BANSHEE launches vertically, intercepts, and — if the track is serviced by something else, reclassified, or leaves the geofence — flies home, lands on the launcher, and puts itself back in the revolver as a ready round in nine minutes with no human in the loop.
-M is the munition: a 158 lb twin-turbojet round for subsonic cruise missiles and Group 4–5 UAS. -R is the rapid round: a 14 lb electric interceptor for Group 1–3, twelve to a launcher. Both carry their own seeker. Neither needs a ground illuminator, a cloud, or a reachback.
Key figures
Specification
Two rounds, one launcher philosophy
The cheapest effector that can service the target is the one that fires. Both rounds come home if they do not.
| Parameter | BANSHEE-M — munition | BANSHEE-R — rapid |
|---|---|---|
| Role | Counter-cruise-missile, Group 4–5 UAS | Counter-Group 1–3 UAS |
| Length | 6.1 ft | 26 in |
| Span / diameter | 3.4 ft deployed | 21 in across the ducts |
| Launch weight | 158 lb | 14.2 lb |
| Propulsion | 2 × 120 lbf turbojets on vectoring pylons | 4 × electric ducted fans, 2.1 kWh pack |
| Launch thrust-to-weight | 1.52 | 2.4 |
| Dash speed | 0.85 M | 210 kt |
| Loiter speed | 140 kt | 42 kt |
| Time to 20,000 ft | 41 s | — |
| Intercept ceiling | 30,000 ft | 12,000 ft |
| Endurance, loiter | 22 min at 12,000 ft | 14 min |
| Endurance, full dash | 4.2 min | 2.5 min |
| Engagement slant range | 15 nm | 4.5 nm |
| Max target closure | 1,200 kt | 400 kt |
| Seeker | Ka-band AESA (380 g), EO/MWIR and passive RF | EO/MWIR and passive RF |
| Effect | 22 lb multi-mode blast-frag / continuous rod, proximity and contact fuzing, selectable inert ram mode | 1.4 lb frag sleeve or inert ram head |
| Ground illuminator required | No | No |
| Recovery | VTOL to the launcher service deck, 0.4 m CEP | VTOL to the launcher service deck, 0.4 m CEP |
| Unattended regeneration | 9 min | 4 min |
| Consumables per sortie | $410 | $18 |
| Life between depot | 40 sorties, or 6 engagement-profile recoveries | 120 sorties |
| Launcher | 6-round revolver | 12-round magazine |
| Unit cost at rate | $118k | $16.4k |
| Hosts | CAIRN ground node, MARROW bed, STYX mission deck, or a 20 ft flatrack |
|---|---|
| Reload of expended tubes | 4 min, 2 people |
| Cueing | REVENANT MESH from CAIRN, VIGIL, STYX or any off-board sensor. No cloud, no reachback. |
| Recovery pad | Co-located, or decoupled up to 800 m from the launcher |
| Terminal recovery guidance | Visual-inertial against a fiducial. GPS not required. |
| Authorization | Engagement requires a valid signed ROE token. Every intercept is written to REVENANT LEDGER. |
Concept of operations
Nine minutes, no hands
Cue, commit, prosecute or abort, regenerate. The fourth step is the one the whole system is built around.
Four phases
-
Cue
A CAIRN node, a VIGIL tower, a STYX deck or any off-board sensor detects inbound and hands the track over REVENANT MESH. No cloud. No reachback. If the network is degraded, the launcher fights on its own sensors.
-
Commit
The launcher fires the cheapest effector that can service the target. BANSHEE-R for a Group 2 quadcopter. BANSHEE-M for a subsonic cruise missile. The gun on the CAIRN node for anything inside 1,200 m. It does not spend a $118,000 round on a $2,000 target.
-
Prosecute or abort
If the track is serviced by another effector, reclassified as non-hostile, or leaves the geofence, the round is recalled. It does not self-destruct. It flies home.
-
Regenerate
Vertical landing on the service deck to 0.4 m. The arm rotates the round into the service bay in 45 s. Purge, 11 lb of JP-8 in 3.5 min, avionics battery hot-swap, fin actuator re-torque if built-in test flags it, full BIT and signature check, back in the revolver. A human touches it only if BIT fails.
Engineering rationale
Why it is built this way
Three decisions, and the price we agreed to pay for each one.
Vectored turbojets instead of lift fans
Two 120 lbf turbojets on vectoring pylons do both jobs — vertical launch and a 0.85 M dash — with no separate lift system to carry. That is what buys 41 seconds to 20,000 ft on a 158 lb round. The tradeoff: vectored turbojet VTOL is fuel-hungry in the hover phase, and the recovery landing costs about 1.9 lb of fuel. On a round that only lands when it did not fire, that is a good trade.
The launcher is a service station
Most of the engineering in this program is not in the round. It is in a robotic service bay that purges, refuels, swaps a battery and runs a full built-in test without a technician. That is what converts a recovery into a sortie. The tradeoff: the launcher is heavier, dearer and more mechanically complex than one that only launches. It is also the only part of the system that turns six rounds into ten engagements.
Every round carries its own seeker
A Ka-band AESA in a 380 g package means no ground illuminator, no uplink during terminal, and no single emitter whose destruction disarms the battery. The tradeoff: a seeker on every round is the largest single line in unit cost, roughly 31% of the -M. A semi-active round would be cheaper per unit — and would die with its illuminator.
Program candor
Open engineering risks
Nine minutes is a bench number. Here is the rest of it.
Four of them
- Canister conditioning is a launcher dependency. Nine minutes holds from −40 °C to +60 °C because the sealed round never sees ambient — but that assumes launcher power for the heater and desiccant circuit. On generator-out the canister holds conditioning for 14 hours, and after that regeneration time is no longer specified.
- Turbojet hot restart after an aborted dash is the least mature part of the cycle. A round that spooled to full dash, aborted and returned hot has a longer service cycle than one that loitered.
- A returning round localizes the launcher. It lands where the launcher is. Decoupling the recovery pad up to 800 m and randomizing the approach corridor reduces the problem. It does not remove it.
- Inert-mode kill probability is model-derived. 0.61 is not yet defended by terminal test data against representative cruise-missile targets.
What cues it and what layers over it
All air systemsAir & missile defense
Bring us your raid
Send us the threat profile and the magazine you have. We will model what regeneration does to it, including the fourteen-minute case.
Every figure on this page is a design target derived from first-principles analysis, not a measured result. No BANSHEE has been built or flown. Where a number here is later measured and comes out worse, we will publish the measured number. No export-controlled information is published on this site.