Latest / Business Strategy Talks with Fexingo: Competitive Advantage, Market Positioning, and Growth Planning / How DJI Builds a Moat on Vertical Integration and Flight Software
Transcript
- Lucas: There are maybe a handful of companies on the planet that hold over 70 percent market share in a global, fast-moving industry. DJI is one of them. And if you ask most people outside the drone world how they did it, they'll say 'cheap Chinese manufacturing.' That answer is lazy, and it's mostly wrong. Luna: Yeah, because plenty of Chinese companies compete on price and still lose share. DJI's real advantage has to be something stickier. Lucas: Exactly. DJI's moat is built on two things that reinforce each other: vertical integration down to the chip level, and a software ecosystem that locks in professional and enterprise users. Let's start with the hardware side because it's the most visible. Luna: And maybe the most underappreciated. Everyone sees the camera and the gimbal, but they don't see what's inside. Lucas: Right. DJI designs and manufactures its own flight controllers, its own electronic speed controllers, its own gimbal motors, and its own camera sensors. They even wind their own copper coils for the motors. That's not normal in consumer electronics. Most drone startups buy an off-the-shelf flight controller from someone like Pixhawk, a standard camera module from Sony, and a gimbal from a third party. They assemble a BOM — bill of materials — and hope the margins work. Luna: And then they compete on spec sheets, which DJI can always beat because their BOM is cheaper and tighter. Lucas: Exactly. Let me give you a concrete example. In 2013, DJI released the Phantom. It used a proprietary flight controller called the naza m — which they'd originally built for hobbyist multirotors. That controller integrated a 6-axis inertial measurement unit, or IMU, with a barometer and magnetometer. Competitors were using discrete components soldered onto separate boards. DJI shrunk the whole thing onto a single PCB and wrote the fusion algorithm in-house. That gave them a size, weight, power, and cost advantage that competitors couldn't match because they didn't control the silicon. Luna: So it's not just that they make their own parts — it's that they design the parts to work together in a way that's impossible to reverse-engineer. Lucas: Right. And over time, that integration got deeper. By 2016, DJI was designing custom ASICs — application-specific integrated circuits — for their image processing pipeline. They were doing in-house sensor calibration, lens alignment, and even propeller aerodynamic modelling. The result is that a DJI drone's propulsion system is roughly 15 to 20 percent more efficient than the equivalent competitor's, measured in minutes of flight time per watt-hour. That's a huge gap in a product category where battery life is the number-one complaint. Luna: And that gap compounds. Longer flight time means better video, less user frustration, stronger brand loyalty. Lucas: The thing is — and this is what I think a lot of Western observers miss — DJI didn't inherit this capability from some state-run industrial policy. They built it because their founder, Frank Wang, was a PhD dropout who was obsessed with flight control algorithms. He started the company in a dorm room in 2006, writing code. The vertical integration came from necessity: he couldn't buy a good enough flight controller, so he built his own. Luna: That's the same origin story as a lot of Silicon Valley hardware companies — except they stayed in hardware instead of pivoting to services. Lucas: And that's the other half of the moat. Because once you have the hardware lead, you can build a software ecosystem that creates switching costs. With DJI, the obvious example is the Pilot app, the SDK, and the enterprise platform. Luna: Let's talk about the SDK in particular. DJI has a software development kit that lets third-party developers build custom apps for mapping, inspection, agriculture, and search and rescue. That's a classic platform play. Lucas: Exactly. If you're a company like DroneDeploy or Pix4D, you've invested thousands of engineering hours integrating with DJI's SDK. You know the telemetry protocol, the camera control protocol, the waypoint navigation system. If you wanted to switch to a different drone platform — say, Skydio or Autel — you'd have to rewrite all that integration. And the market is so dominant that the third-party software companies will prioritise DJI compatibility first, which makes the competitors' platforms less useful by comparison. Luna: It's a network effect, but not the typical kind. It's a developer network effect. Lucas: Right. And DJI has also built proprietary protocols that lock in accessories. For example, the DJI Lightbridge and OcuSync transmission systems — they're not just radio modules, they're complete digital video and data links with custom compression. No competitor can make an accessory that talks to a DJI drone at the same latency and range without reverse-engineering the protocol, which would violate the DMCA in the US. Luna: So a farmer who buys a DJI Agras spray drone for crop dusting can only use DJI's mapping tools, DJI's replacement parts, DJI's service centres. That's a very deep lock-in. Lucas: And that farmer probably doesn't care about geopolitics. He cares that the drone sprays his fields accurately. DJI's agricultural drones account for something like 40 percent of global agricultural drone spraying by acreage. And the data — the field maps, the flight logs, the treatment records — lives inside DJI's cloud. Switching would mean losing years of crop history. Luna: That's a real economic switching cost. Not just 'I like the brand.' Lucas: If these conversations have moved your work forward in some small way, it's worth mentioning that listener support is what keeps Fexingo ad-free. No sponsors, no mid-roll interruptions — just the topic at hand. If that matters to you, you can find us at buy me a coffee dot com slash fexingo. Luna: Yeah, we don't run ads, and that's intentional. It keeps the conversation honest. Lucas: Alright — back to the moat. The big challenge facing DJI right now isn't technological. It's regulatory. The US government has been progressively restricting DJI's ability to sell to federal agencies, and in 2024, the Countering CCP Drones Act effectively banned new DJI drone imports into the US Department of Defense and other federal agencies. Several states are following suit. Luna: So the question becomes: can a company that controls 70-plus percent of the global market survive losing its biggest single-country market? The US is maybe 25 to 30 percent of global drone spend. Lucas: It's survivable, but it's painful. The interesting thing is that DJI's moat actually makes them harder to replace than a typical hardware vendor. You can't just order 10,000 units from Skydio and switch overnight. The software ecosystem, the pilot training, the accessory ecosystem — it all takes years to replicate. And in the meantime, DJI continues to dominate in Europe, Asia, Africa, and Latin America. Luna: But the US market is also where a lot of the high-value enterprise innovation happens. If DJI is locked out of contracts with US oil companies, utilities, and telecoms, they lose the feedback loop that makes their products better. Lucas: That's a real risk. The counterargument is that DJI has such a massive installed base globally — over 10 million drones sold by some estimates — that the data and feedback from China, India, and the Middle East is enough to maintain their lead. And they've started to localise. They opened a factory in California in 2020, though it's not clear how much production happens there. Luna: Let's also look at a competitor that's trying to attack the moat. Skydio is the most credible us based challenger. They raised over $500 million, and their drones are known for best-in-class autonomous obstacle avoidance — what they call 'the drone that flies itself.' Lucas: Right. Skydio's approach is pure software differentiation. They use a set of nine cameras and neural networks to build a 3D map of the environment in real time. It's genuinely impressive. But here's the problem: Skydio doesn't manufacture its own cameras, flight controllers, or motors. They buy Sony camera sensors, Nvidia Jetson processors for the AI, and standard motors and propellers. That means their BOM is significantly higher than DJI's for a similar spec, and their margins are thinner. Luna: So Skydio can't win a price war, and DJI has a huge cost advantage even before you factor in scale. Lucas: Exactly. Skydio's only path is to be so much better at autonomy that customers pay a premium. And for some use cases — like bridge inspection where you need to fly close to structures without GPS — they are better. But for the vast majority of drone use — aerial photography, mapping, surveillance — DJI's obstacle avoidance is good enough, and the price gap is wide. Luna: And the gap is widening. DJI's latest consumer drone, the Mini 4 Pro, costs about $1,000 and has obstacle avoidance in all directions. Three years ago, that level of capability cost $2,000. Lucas: That's the vertical integration flywheel. Because DJI controls the IC design, the optics, the motor windings, and the software, they can improve performance and reduce cost simultaneously. A competitor that buys components from different vendors can't do that — each vendor improves on their own roadmap, and system-level optimisation is limited. Luna: So the moat is really two overlapping moats: a cost moat from vertical integration, and a switching-cost moat from the software platform. And each reinforces the other. Lucas: Right. And there's a third layer that's underappreciated: brand and community. DJI has essentially defined what a consumer drone looks like. The Phantom shape, the gray and white colour scheme, the four-armed quadcopter design — that's the visual archetype. If you show someone a drone that doesn't look like that, they don't recognise it as a serious product. Luna: That's a form of category ownership. Like how everyone calls a facial tissue a Kleenex. Lucas: Exactly. And DJI has also built a massive ecosystem of third-party content. YouTube is full of DJI footage. The hashtag #DJI on Instagram has over 40 million posts. That user-generated content is a marketing moat that competitors can't buy — it's earned over a decade of consistent quality. Luna: So looking forward, what could break the moat? Regulation is the most obvious. If the US extends the federal ban to state and local governments, and if Europe follows suit with similar security concerns, DJI could lose 40 percent of its addressable market. Lucas: That's the biggest risk. The other one is technological disruption from a fundamentally different architecture. For example, if someone figures out a way to deliver equivalent performance using a distributed swarm of tiny drones, or if a new form of propulsion — like hydrogen fuel cells — doubles flight time, DJI might not be able to adapt quickly because their vertical integration is optimised for current technology. Luna: But that's a long shot. In the near term, the most likely scenario is that DJI continues to dominate the global market while gradually losing share in certain government-sensitive segments in the West, but those segments are small compared to the consumer and commercial boom in Asia. Lucas: I think that's right. And the lesson for anyone building a hardware business is clear: if you can vertically integrate the core components that create the biggest performance gaps — in DJI's case, the flight controller and the gimbal — you build a cost advantage that compounds. And if you layer on a software platform that creates real switching costs, you become very hard to dislodge, even with geopolitical tailwinds against you. Luna: It's a reminder that moats aren't just about patents or brand. Sometimes the deepest moat is simply the ability to make a better product at a lower cost than anyone else, and then lock users in with software. Lucas: And DJI has done exactly that. It'll be fascinating to see whether the regulatory pressure becomes a real wedge or just a temporary headwind.