A look at the burgeoning drone software and AI industry, a focus for an estimated 65 US startups, often for military uses
Patrick McGee / Financial Times : Tweets: @thuttag and @carlbfrey Tweets: Thu Thu Aung / @thuttag : Drone innovation moves from hardware to self-flying software https://giftarticle.ft.com/... “[We are at] the tip of the iceberg in terms of capabilities that are going to be unlocked by software,” Adam Bry, Skydio's chief executive, said. Carl Benedikt Frey / @carlbfrey : According to the Federal Aviation Administration, the use of registered industrial drones surged from 12,000 in 2015 to nearly 500,000 by 2020. This “appears to be at an inflection point demonstrating powerful stages of growth”. https://www.ft.com/...
Context & Ripple Effects
This piece lands after a full cycle in US drones: the 2019 shakeout, when many well-funded startups closed or pivoted to services (closed or pivoted to services as the market matured), then the DJI blacklist reshaping who could fund a US drone maker at all — Skydio's $171M round led by a16z was billed as the first significant post-blacklist fundraising. What changed by this article is where the value is said to sit: not airframes but self-flying software, with an estimated 65 US startups focused on drone AI, often for military buyers.
The demand side backs the claim — FAA-registered industrial drones grew from 12,000 in 2015 to near 500,000 by 2020 — and Skydio CEO Adam Bry frames current autonomy as 'the tip of the iceberg' of unlocked capability. The counterweight already visible in coverage is China's low-altitude economy push, leveraging its production dominance below 1,000 meters.
First-order effects
- Skydio and the ~65 US drone-AI startups gain a clear anchor customer: military demand, which funds autonomy development while consumer hardware margins compress.
- The FAA's registered-drone growth curve turns industrial operators into the addressable market, pulling software vendors toward inspection, mapping, and enterprise fleets rather than hobbyist sales.
Second-order effects
- Hardware-led rivals face a squeeze: if Bry is right about capabilities being software-unlocked, airframe makers without an autonomy stack compete on price against Chinese incumbents like DJI.
- The services pivot documented in 2019 accelerates in a new form — instead of selling flights, software firms license autonomy itself, changing revenue models from per-mission to per-seat or per-fleet.
Third-order effects
- If military procurement keeps anchoring US drone software, the industry structurally splits along national lines: US autonomy stacks versus Chinese production scale, echoing the low-altitude economy build-out already underway.
- State purchasing decisions become the de facto filter for which drone-AI companies survive — a repeatable pattern where defense demand substitutes for the consumer market that collapsed in 2019.
The trend: Drone-industry value is migrating from airframes to self-flying software, with US military procurement replacing the consumer market as the funding engine.