The lineage we work in.

We make robots in a moment when most of them are made dishonestly. Carbon-fibre frames laid up in factories that consume the same fibre forever. Lithium cells mined in places we will not visit. Software designed to expire on a schedule. The machines that result are extraordinary, and they are also fragile in a way no eighteenth-century mill ever was: they cannot be repaired by their owners, and they cannot be returned to the earth that made them.

We do not think this is what the discipline should look like. Robotics is one of the few engineering fields where the materials, the manufacturing, and the maintenance are all still up for negotiation. We are negotiating.

A machine that cannot be repaired by the person who flies it is not a finished machine. It is a contract. — Workshop note, MMXXVI

Leonardo's notebooks were not illustrations of finished thoughts. They were how he had thoughts. The act of drawing a mechanism — committing it to a page where every line implies a tolerance, every curve implies a stress — forces a kind of clarity that no parametric CAD environment imposes.

We start every machine the same way. A fresh page, a sharp pencil, and the rule that if it cannot be drawn on a single sheet, it is not yet understood. Only after the drawing closes — after every part is named, every joint accounted for — do we open CAD.

Why this is not romanticism

A drawing forces you to count parts. It forces you to acknowledge what you do not yet know. CAD models hide complexity inside trees of dependencies; a drawing does not. The thirty extra minutes spent at the page have, in our experience, paid for themselves several times over by the time the prototype flies.

Bamboo is, by most engineering measures, an exceptional material. Its specific stiffness rivals aluminium. Its specific tensile strength approaches steel. It grows three feet in a day and captures carbon as it does so. When laminated and resin-stabilised, it becomes a structural material that is forgiving in a crash and quiet under load.

More important than its mechanical properties: it grows back. We can size our drone production in groves rather than mines. The supply curve is bounded by sunlight rather than geopolitics. This is not a small thing.

We supplement bamboo where physics requires: bio-resin at joints, walnut turnings for high-stress hubs, hemp linen wraps where vibration damping matters. We use aluminium and carbon only at the smallest possible margin, and we say so.

Read the materials study →

Every part on a Davinci Robots machine is a part you can hold, name, and replace at a workbench. There are no proprietary modules. There are no firmware locks. There are no vendor-specific connectors. If a customer wants to keep their drone flying for twenty years, they can.

This is a constraint we accept. It rules out certain miniaturisation techniques, certain integration tricks. It also rules out the slow extractive relationship most modern hardware has with its owner. We think the trade is worth it.

What this looks like in practice

• parts Every part has a name, a drawing, and a price.
• tools Every drone ships with the keys to its own service.
• software Firmware is open; no part of the machine is held to ransom.
• support We answer service letters for the life of the airframe.

Drones are tools, and tools have politics. We will not build weapons. We will not build platforms whose purpose is surveillance of civilian populations. We will not build closed ecosystems that lock customers into proprietary services. We will not chase autonomy for its own sake.

What we will build: machines for conservation, for cinematography, for inspection, for research, for the handful of remote-area logistics problems where a quiet electric drone is a better answer than a diesel truck. We will build slowly, repair as much as we sell, and stay small enough to know what our products are being used for.

Some of this will sound idealistic. Some of it is. We are not pretending the economics are easy or that we have all the answers — we have a workshop, a set of drawings, and a willingness to build slowly enough that we can be honest about what we are doing.

If you are a pilot, conservationist, researcher, or patient investor for whom this lineage makes sense, we would like to hear from you. We answer every letter.

Write to the workshop →