A flying scale model is the most persuasive thing an aerospace startup can show, and one of the easiest to over-read. This note sets out — for our own programme and in general — what a one-third-scale prototype actually settles, and what it leaves open. We would rather write it ourselves than have it inferred.
The arithmetic that governs everything
Scale a vehicle up by a factor of three in every linear dimension and three quantities move at different rates:
- Length ×3
- Area ×9 — this includes the rotor disc area that produces lift
- Volume, and therefore mass ×27 — assuming the same materials and proportions
That mismatch is the square-cube law, and it is the reason scale flight is easier than full-scale flight rather than merely smaller.
Disc loading — thrust divided by disc area — rises by the ratio of 27 to 9, so three times. Hover power scales roughly with thrust to the power of one and a half, divided by the square root of disc area, which works out at about 47 times the power for 27 times the mass. In other words the full-scale machine needs a power-to-weight ratio around 1.7 times better than the model, from a propulsion system that also has to carry its own energy.
Nobody escapes this. It is why a hobby multirotor hovers on components from a catalogue and a passenger-carrying one does not.
What the scale prototype does settle
That the geometry works. Putting the propulsion inside the track is a configuration question before it is a power question. Whether the track can serve as landing gear, propeller guard, ground drive and obstacle-clearing geometry at once — and whether the vehicle sits, rolls and lifts without fouling itself — is answered at any scale.
That the transition is controllable. Going from driving to hovering and back is the part of a multi-domain vehicle most likely to produce an unpleasant surprise: the moment when wheels or tracks unload, the machine is neither a ground vehicle nor an aircraft. Demonstrating that sequence with a real control system on real hardware is a genuine result. Simulation is not the same thing, because the failure modes here are contact, friction and attitude at low airspeed — exactly what simulation models worst.
That the control laws close. Stability and control authority in hover, in translation and during transition are demonstrated with the actual actuator layout, not an idealised one.
That redundancy behaves as designed. An eight-unit distributed propulsion system either reallocates thrust correctly when units are lost or it does not, and the answer is largely architectural rather than scale-dependent.
What it does not settle
Energy. This is the honest constraint for every electric VTOL programme, ours included. Battery specific energy does not improve when you make the aircraft bigger; the mass fraction available for batteries does not grow with the square-cube law in your favour. Endurance at full scale is decided by cells, not by geometry, and it is the number to interrogate hardest in anybody’s programme.
Structure. Under self-weight, stresses in a geometrically scaled structure grow with linear dimension. A full-scale airframe cannot simply be the model with every part three times larger; load paths get redesigned, and composite lay-ups get designed rather than scaled.
Certification. Nothing about a scale model addresses airworthiness, and no amount of flight footage substitutes for it.
One thing that gets better with size
Not everything works against the full-scale machine. Aerodynamic efficiency generally improves. Small propellers operate at low Reynolds numbers, where boundary layers are relatively thicker and airfoils underperform their design intent; a three-times-larger blade at comparable speed operates at roughly three times the Reynolds number, in a more favourable regime. A scale prototype is therefore usually flying with an aerodynamic handicap the production vehicle will not have.
This matters for reading demonstrations correctly in both directions: the model is flattered by the power and structural arithmetic, and penalised by the aerodynamic one.
Where our programme stands
A fully functional 1:3-scale prototype flies and performs full manoeuvres. It currently uses one propeller per track; the production-grade coaxial arrangement halves the count of track-mounted units while doubling redundancy. The sealed amphibious hull is part of the design and has not been demonstrated in operational use. Performance figures from an earlier project phase are under revision and are deliberately absent from this site until they are confirmed.
We publish that breakdown on the fact sheet as a table with two columns — demonstrated, and part of the design — because the distinction is the single most useful thing an investor, a journalist or an engineer can take from a programme at this stage.
How to read anyone else’s demo
Three questions, in order of how much they tell you:
- What scale, and what is the intended full-scale mass? Without both numbers, flight footage carries no information about the eventual vehicle.
- Was it tethered, and was it free-flying outdoors in wind? These are different claims.
- What is the demonstrated endurance, at what payload? Hover time is where the energy constraint becomes visible, and it is the number most often left out.
Applied to us as readily as to anyone else. That is the point of writing it down.
Related: how the vehicle works · the patent family · mission profiles.