Sir Airffic GoWork backwards from who pays. That is the basics of business.
Products flow right, money flows left
Component and chip makers sell to aircraft makers. Aircraft and autonomy-software companies sell to operators and customers.
The final payers are companies that deliver goods, firms that inspect bridges and power lines, police and fire services, and defense organisations.
- Parts & sensors
- Chips & AI models
- Aircraft makers
- Autonomy software
- Operations & data
- Customers (delivery, inspection, police, defense)
Who pays: four big wallets
Defense & security
Largest and fastest
Since 2022 budgets for autonomous drones and counter-UAS have surged in many countries. Export controls and ethical lines are strict.
Public safety (police, fire)
Fast growth in the US
Drone as First Responder (DFR) sends a drone to an emergency before officers arrive. Autonomy and remote operation are the value.
Infrastructure inspection & survey
Japan's core opportunity
Bridges, power lines, plants and roofs. Image recognition finds cracks and hot spots. Labour shortages are a strong tailwind.
Logistics & delivery
The long-term prize
Zipline, Wing and Amazon fly real deliveries. Lowering cost per delivery needs autonomy and airspace management so one person can oversee many aircraft.
Where profit collects — and where it doesn't
| Layer | Profit retention | Why |
|---|---|---|
| Sensors & parts | Low–medium | Prices fall with smartphone and auto volumes; specialised radar is an exception |
| Chips | High (for giants) | Tool lock-in makes switching hard; entry needs huge investment |
| Aircraft | Medium | Scale matters; US–China decoupling made country of origin a value |
| Autonomy software | Medium–high | Hard to replace once installed, but competes with open source such as PX4 |
| Operations, airspace & data | High | Embedded in customer workflows and approvals; stronger as data accumulates |
This is a qualitative reading of structure from public information, not a measurement of any company's margins.
Size the market yourself
Forecasts for drone and AI markets often differ several-fold between research firms because each defines the market differently.
So this site does not republish other people's figures. It gives a formula; with it you can see which assumptions move the answer most.
- Pick one job (e.g. power-line inspection)
- Estimate how many aircraft fly that job
- Decide the share with AI features (adoption rate)
- Decide the annual software and service spend per aircraft
- Aircraft × adoption × spend per aircraft = annual market for that job
Sir Airffic GoNever swallow a number whole. Think in formulas and you can explain the market consistently to a board or an investor.
Frequently asked questions
How big is the Aerial Physical AI market?
Research firms define the market differently, so forecasts vary widely. We recommend estimating it for your purpose with: aircraft × AI adoption × annual spend per aircraft.
Where is the most profit in the drone market?
Operations and software layers embedded in customer workflows, where data and regulatory approvals accumulate, are generally hard to replace and retain profit. Components tend to face price pressure.