Skip to content

Drone Industry Cycles: How the 2016 Bubble Burst, and What's Different This Time

Aug 6, 2026 1 min
Home › tech
TL;DR The 2016 consumer drone bubble left specific wreckage: 3D Robotics stopped making hardware, GoPro recalled all 2,500 Karma units six weeks after launch and cut 15% of staff, Parrot cut 35% of its drone workforce, and Lily Robotics collapsed after taking $34M in pre-orders. In 2023 even Skydio — $570M raised — exited consumer, and three years later it is valued at $4.4 billion. This wave runs on a completely different engine, but three things are exactly the same.
#drone#uav#industry-history#supply-chain#hardware
Table of Contents
  1. Wave one: 2014–2016, betting on consumers
  2. The collapse: inside a single year
  3. Why it broke: two reasons, neither technical
  4. The decade after: consumer is a dead end
  5. Wave two: post-2022, a different engine
  6. Three structural differences from last time
  7. But three things are identical
  8. The bottom line
  9. References
Series: Taiwan's Drone Industry, Taken Apart (12 / 38)
BVLOS in Three Jurisdictions: Taiwan Has No Framework At All Taiwan's Drone Supply Chain: Where the 267 Companies Are, and Which Layer They're Stuck On

🌏 中文版

Every industry boom runs into the same question: is this real, or another bubble?

Drones are worth asking about specifically, because this one has already burst once — and burst thoroughly. The wreckage from 2016 is concrete enough to use as a measuring stick.

This is the judgment entry in the drone series. Earlier pieces covered industry structure and Taiwan's position; this one covers the timeline.

Wave one: 2014–2016, betting on consumers

The narrative then was that camera drones would enter every household the way digital cameras had. Money followed:

  • 3D Robotics, founded by former Wired editor Chris Anderson, raised roughly $120 million and shipped Solo in May 2015, built around GoPro camera integration
  • GoPro announced its in-house Karma in May 2015, pouring its profits into development
  • Lily Robotics collected $34 million in pre-orders on the strength of a throw-it-and-it-follows-you video
  • Parrot crossed over from consumer electronics and built a drone division in the hundreds

The collapse: inside a single year

March 2016: 3D Robotics announced it would stop manufacturing drones and pivot to enterprise software. An anonymous former employee's line to Forbes still gets quoted: "3DR was a $100 million blunder based on ineptitude."

19 September 2016: GoPro Karma launched. Six weeks later, on 8 November, GoPro recalled all ~2,500 Karma units sold since 23 October — the drones were losing power in flight, later traced to the battery decoupling from its connector. The recall led to roughly 15% of GoPro's workforce being laid off; Karma limped to discontinuation in 2018.

Early 2017: per Forbes's account, Parrot cut 290 people — 35% of its drone workforce — and Lily Robotics folded amid a consumer protection lawsuit alleging false advertising.

Within a year, the US and European consumer drone field was essentially wiped out.

Why it broke: two reasons, neither technical

First, demand was badly overestimated. Market research from 2016 found that fewer than 5% of US broadband households owned a drone, and fewer than 4% said they were interested in buying one. "As common as digital cameras" never happened — aerial photography is a hobby, not a need.

Second, DJI's cost structure made survival impossible. Drone Analyst's analysis notes that while some called it dumping, the evidence doesn't support that — DJI priced consistently worldwide, and its 2015 $75M Accel investment was on par with what Yuneec and 3DR raised. The real difference was vertical integration and manufacturing automation: DJI could profit at price points where competitors simply could not.

Together that's the worst possible combination: a price war against a lower-cost opponent, in a market smaller than you thought.

The decade after: consumer is a dead end

The strongest evidence isn't 2016. It's 2023.

Skydio announced in August 2023 that it was ending its consumer business to focus entirely on enterprise and public sector. By then the company had raised roughly $570 million at about a $2 billion valuation, and was the only American maker to genuinely beat DJI on autonomous obstacle avoidance. CEO Adam Bry's explanation:

The impact we're having with our enterprise and public sector customers has become so compelling that it demands nothing less than our full focus and attention.

Even the best-funded, most technically capable American startup walked away from consumer. That's not failure — it's admitting the unit economics don't close.

And the decision has since been vindicated. In April 2026 Skydio announced a $110 million Series F at a $4.4 billion valuation, with CEO Adam Bry emphasizing how little was raised — the company now has "a strong core business generating hundreds of millions in annual revenue, with strong unit economics and hypergrowth," and its capital needs are falling. Per DroneXL's account, Skydio has shipped over 60,000 drones to more than 3,800 enterprise customers, and its March 2026 US Army order — $52 million for nearly 3,000 aircraft — was the largest single-vendor small UAS purchase in US military history.

Three years after exiting consumer, the valuation went from $2 billion to $4.4 billion. That turns "consumer is a dead end" from a defensive claim into an offensive one.

As of 2026, Drone Industry Insights still dispatches it in one line: "The recreational market is flat." Ten years on, that judgment hasn't moved.

Wave two: post-2022, a different engine

Nothing driving the current wave overlaps with the last one.

Demand shifted from consumers to militaries and public agencies. The war in Ukraine turned small drones into consumables; Ukraine produced roughly 4 million systems in 2025. This demand needs no convincing — spending hundreds of dollars to destroy millions of dollars of equipment makes its own argument.

The shape of the money changed too. Per Drone Industry Insights' 2026–2035 report: drone investment fell 52% in 2024, then hit a record $3.86 billion in 2025, with 77% flowing to dual-use companies; the first two months of 2026 alone drew $1.7 billion.

Taiwan's numbers move in step. Output grew more than 2.5x and airframe exports 21x from 2024 to 2025, with the Czech Republic, Poland, and the US as the top three markets.

Three structural differences from last time

① The demand source changed, and predictability improved. Last time it was consumer discretionary spending — it evaporates in a downturn. This time it's defense and public budgets, visible in yearly increments once appropriated. That's a genuine improvement, but it swaps one risk for another: political risk replaces cyclical risk.

② The unit economics need no selling. Last time you had to convince a consumer their life needed an $800 camera drone. This time it's using a $500 device to destroy a $2 million armored vehicle. The first is a marketing problem; the second is arithmetic.

③ The moat shifted from cost to certification. Last time the contest was who could drive costs lowest — DJI won, decisively. This time it's who can pass NDAA, Blue UAS, and Green UAS supply chain and cybersecurity review — and DJI is on the excluded side of that contest. The rules changed, and that is precisely the window for Taiwanese, Japanese, and European suppliers.

But three things are identical

This is the part I'd watch most closely.

① Capacity is being built ahead of orders landing. In 2016 the lines were built and demand never came. In 2026 countries are expanding capacity while most procurement is still inside budget and legislative processes. Taiwan's defense special statute still sitting in the legislature is a concrete instance of that gap.

② Narrative growth rates far exceed the base. "Exports grew 21x" and "output of NT$12.9 billion" are two faces of the same number. The last cycle had the same sentence structure — the difference is whether you ask for the denominator when you hear the growth rate.

③ Consumer never came back. All of this wave's growth is military and commercial; the recreational market has been flat for a decade. Any thesis premised on drones entering every household was wrong in 2016 and is still wrong.

The bottom line

The core error in 2016 was betting on the wrong demand side. The technology worked and the products shipped; not enough people wanted to buy them.

This time the demand is real — battlefield consumption, public-sector inspection, non-Chinese supply chain reshuffling, all backed by budget and policy. But real demand doesn't mean every company survives, and the failure mode will differ: last time companies died because they couldn't sell. This time they're more likely to die because they can't deliver — specs unmet, certification not passed, acceptance testing failed.

One sentence to separate the two cycles: last time the risk was on the market side; this time it's on the execution side.

References

The first bubble

Shakeout and pivot

The second wave

On this site

Related

The Drone Industry Map: Components, Regulatory Ceilings, and the Non-Chinese Supply Chain Rebuild 1 min Taiwan's Drone Supply Chain: Where the 267 Companies Are, and Which Layer They're Stuck On 1 min Taking Apart Two TTSB Crash Reports: Neither Was the Operator's Fault 1 min