Former Skydio partnerships lead on where value accrues in the drone stack

This former Skydio executive helped scale the company’s go-to-market through partnerships with T-Mobile, Oracle, and NVIDIA, and led integrations across both horizontal tools like Slack and S3 and industry-specific platforms like SAP and Genetec. In this interview, we talked about how value is shifting across the drone stack, why vertical integrations tend to be stickier in enterprise sales, and how the commoditization of drone hardware is pushing vendors toward full-stack Apple-like business models.

Can you tell us what’s unique about building partnerships and integration ecosystems in the drone category?

One of the things that’s unique in the drone space is that you're working not just with software companies when you're doing partnerships in the enterprise space. You're also working with hardware companies, companies that might make sensors, attachments, motors, batteries, chips, GPUs. So the canvas for partnerships is broader because you're dealing with robotics. And so the partnerships could be really anywhere on the technical stack.

It could be on the hardware side. It could be on the firmware side. It could be on the AI side, and it could be on the SaaS and cloud side as well. And then the other piece there is with software partnerships, you're typically building integrations. You might do something similar on the hardware side, but on the hardware side, there's just more complexity when it comes to building integrations, testing integrations, and ensuring that whatever it is you're integrating with the drone does not impact safety during flight or does not impact things like flight time, range, endurance, etc.

So those are some of the considerations that are different from your classic enterprise software partnerships.

In technology-related industries, people often talk about the value chain. The question is always—and it comes up a lot in the context of AI these days—where is the value accruing? Can you speak to what you learned about where value accrues in the drone stack and whether you think it will change moving forward?

The drone space is really interesting on the hardware side. If you look at technologies like smartphones, PCs, etc.—you would have these same conversations that revolved around value accrual.

One difference between those products and the drone world, and robotics in general, is that the manufacturing for robotics is not yet at scale.

I would say it's not quite that same level of maturity [relative to PCs and smartphones] as a space, as a product category. Therefore, it's quite dynamic in terms of where the value accrues today. My guess is there's going to be a lot of change in the next ten years as supply chains mature, as manufacturing processes mature, as hardware manufacturers are able to improve economies of scale in manufacturing.

But it's not like with Dell and Windows, where there was a debate about who's really accruing the value in the relationship. Is it Intel? Is it the PC maker Dell? Or is it Microsoft, the company that makes the operating system that runs on the PC? A lot of people thought it was Wintel, but then it became just Windows.

At the moment, if you're a pure play hardware drone company, you're probably making some money, but you're not making a lot because margins on hardware are generally going to be lower than software. There's also more complexity in building hardware products than building software products.

When I look at Skydio and companies that are understanding this dynamic and looking to figure it out, a lot of companies are saying, "We don't just want to be hardware companies. We want to be like Apple." We want to own the hardware. We want to own the operating system, so iOS in that example. And then we want to have an App Store and our own services offerings, whether it's Apple News or Apple Music or Apple Health, so that we can really look at all aspects of the value chain going from the hardware up all the way to software and applications and try to maximize our ability to drive revenue and margins through that entire spectrum.

So I see something similar happening with the drone space and with robotics in general. I think there's going to be a push towards companies building not just hardware, but really full-spectrum products. But then thinking strategically about how to integrate with third parties such that you're giving customers what they need, meeting customers where they are, but without abdicating large segments of the market to third parties.

That makes a lot of sense. And in terms of software integrations specifically, I was curious about what you saw working there. Obviously, that's a great way to expand your ecosystem and leverage that for distribution, etc. What have you seen been successful in terms of software integrations to grow that software footprint and offer hooks into which other products can integrate into the software part of a drone product?

The way we looked at it at Skydio, is that there was a set of horizontal integrations, and there was a set of vertical integrations.

Horizontal means those integrations could apply to the vast majority of customers. These are common things that all customers could benefit from. An example might be integrating with Teams or Slack or AWS S3 or Azure Blob Storage or Box. These integrations could help facilitate the storage of media captured during flight and also improve collaboration between teams that are flying the drone and processing the imagery coming out of the drone and then taking some action.

When you look at vertical integration, it tends to be more industry-based. For example, if you have customers in the energy utility space, they're using drones to do asset management and asset inspection—assets in this context being substations, distribution towers, transmission towers, or power lines. In that scenario, you often have very specialized tools, products that are used by those organizations to do asset management or enterprise asset management (EAM). It's a category defined by Gartner, and companies in that category include IBM Maximo, Oracle, and others.

If I'm working with a utility company and they're using SAP for their asset management solution, then I'd want to find a way to integrate with SAP in order to push the data that's being collected on those assets by the drone into systems that are used more broadly to do asset management.

Another example of vertical integration might be with public-safety organizations, like law enforcement agencies. They have their own systems for CAD (computer aided dispatch, in this context). When you call 911, your call ultimately ends up in a CAD system that's used by dispatchers to dispatch police officers to your location. In the case of using drones for public safety, you might be integrating with those systems.

A third example, again public-safety related, might be integrating with video-management systems like Genetec that consolidate video feeds from hundreds or potentially thousands of CCTV cameras in their region. When the drone's flying, it's essentially a flying camera, so integrating with those video management systems provides a common operating picture to the operator or control room so they can see what the drone is seeing and then make decisions accordingly.

And would you say that, in this context, vertical integrations prove stickier, more differentiated, and get you deeper into folks' workflows vs. horizontal integrations, or would you say that's not the case?

I would say that's generally the case. There are some horizontal integrations that are very useful and can drive a lot of value to operators and people adjacent to them. When I say operators, I mean people operating drones, flying the drones.

An example of a horizontal integration that's incredibly valuable would be an integration with Esri, which is mapping software used by pretty much everyone. That integration allows customers to pull in their own maps and annotations into the Skydio cloud system in order to have their own reference points to see where the drone is and where it needs to go.

But broadly speaking, you're right. The vertical integrations tend to be much stickier and tend to be more of the gatekeepers when it comes to going from a pilot to a full-scale implementation. A law-enforcement agency might say, "For the first five drones that I'm buying from you to run these pilots in these small parts of the city, I'm fine not having you integrate into my CAD system. But if you want to sell me 100 drones, then there's no way I can proceed without it."

In terms of large enterprises, could you speak a little to what the triggers are for a large Fortune 500-type org to seek a formal engagement with a company like Skydio or someone else in the drone ecosystem? What are the push and pull factors?

Absolutely. There's partnerships, which are typically people that you are integrating with or co-selling with or both. And then there's customers, people that are going to be buying the drone and driving revenue.

On the customer side, it is vertically oriented. You can be selling drones to consumers or you can be selling drones to enterprises in B2B scenarios. Skydio started out as a company selling drones to consumers in 2014. Back then, there was really no market for enterprise drone usage outside of very expensive and sophisticated drones used by the military.

If you look at the civilian market, it didn't exist. It's really the last five to six years that it started to evolve and develop, and it's growing very quickly. The way it initially happened was that, for example, utilities were doing inspections of their assets manually, climbing up using a ladder to inspect a distribution tower or transmission tower, or inspecting power lines using a helicopter. So very manual, very expensive, and potentially very risky and dangerous processes.

I think a few drone enthusiasts started bringing their DJI drones and saying, "Hey, can I try and use this instead?" There was skepticism, like "this may or may not work," but enough people got the opportunity to try it that organizations started to see this was an interesting technology that could potentially help them. That's what led to them starting to look for solutions in this area.

Once that started happening at enough places, there was enough critical mass that people at the top started saying, "Can we use drones? And if we use drones, does that mean we have a better safety record for employees? Does that mean we can do more inspections at a much lower cost? Does that mean we can do more frequent inspections? Does that mean we have better data because maybe the drone can capture not just color pictures, but also infrared pictures, which give you new information on the asset?" Those were the conversations that started happening, and over the course of the next five years, it became more of a real thing.

In terms of the pull factors, we talked about economics and safety. Another pull factor you see a lot in the utility space and potentially in other spaces where inspection is relevant is that many of the people doing inspections (called linemen typically) have been with the utility for 10, 20, 30 years, and many are retiring. The utilities are worried about a loss of institutional knowledge that resides in the heads of these linemen who've been doing the job for decades.

They're trying to figure out how to bring in a new generation of workforce and equip them with better, more modern tools. This is all part of a bigger digital transformation story that many energy companies and utilities are going through—everything from using drones instead of cameras and ladders, to using iPads instead of notepads for taking notes.

In terms of data capture, you start to see drones as being very useful for digital transformation.

When you look at other countries, like Japan, it's not just about people retiring. They also have a demographic issue where they just don't have enough young people to replace the older people who were doing those jobs because of changes in the population demographic. They're proactively looking for robotic solutions for a large number of industries because they have a unique demographic challenge.

There are different reasons why organizations or governments are choosing drones—from being unable to hire people to demographic changes to digital transformation, and usually a mix of all three.

And when you worked in the past with organizations like NVIDIA, Oracle, and T-Mobile, I assume those were some form of distribution or technical partnerships? What does that kind of engagement look like? What possibilities for partnerships or distribution channels do they represent for drone startups?

It's a great question. With T-Mobile, when you're flying drones, typically you're flying it with a remote control just like you might have a remote control car as a toy.

The drone can only fly as far as the remote control, and the drone can communicate using radio waves coming out of the remote control or coming out of the drone and going back to the remote control. That creates issues because if you need to have a human being with a remote control flying a drone, that person can only fly the drone as far as they can see visually with their eyes, and they can only fly the drone as far as the signal might reach. If you get any kind of occlusion because of a building or something else, you can't communicate with the drone, and that could potentially lead to a dangerous scenario.

The T-Mobile partnership revolves around leveraging 5G as a command-and-control mechanism to communicate with the drone, between the drone and either the remote control, the ground station that the pilot is using, or some kind of digital command center where people are flying the drone remotely. They could be sitting in California while the drone is flown in Texas or Alaska, and they don't need to worry about either line-of-sight or radio waves because they're using 5G to fly the drone.

Oftentimes with these large companies, there's a very strong integration or partnership value proposition. They might also drive distribution, but it's less obvious how a Skydio might drive distribution through a T-Mobile. T-Mobile is used to selling data and wireless plans and devices related to providing connectivity. They aren't as familiar with selling drones, and selling drones is a very different discipline—you have to know a lot of things before you can answer questions that customers would have.

It’s not always the case that telecoms are not able to do hardware distribution in this market. For example, British Telecom has spun up a drone-focused organization within their company. This organization is trying to build software, but also stand up a drone services provider that is part of British Telecom. The initial use case was using this drone service provider to inspect their own telecom assets—their own cell towers, etc. But then they started thinking, "Why don't we also provide this to other companies that need inspections?"

So there are some telecom companies coming up with ways to monetize drones in more interesting ways. I would say that's probably a more unique scenario, however. I haven't seen too many other telcos doing that. But I have seen lots of telcos in Europe that are very interested in drones. They see drones as a new vector for revenue because you need to buy a data plan to support flying the drone on 5G. These data plans are not cheap, they're much more expensive than your typical consumer data plan. So it becomes an interesting revenue vector for them five, ten years down the road where they anticipate there'll be many more drones flying in the sky, doing everything from inspections to delivery.

I wanted to go back to image and data capture and then the handoff to other systems. From the outside looking in, it seems a company like DroneDeploy has really focused on that middle stage, the data-processing layer. They style themselves as a reality-capture company. They're really going after this idea that the value is in the data, in the software to process the data, and whatever vertical or generalist use cases that result. Do you think that's inevitably where value will go if drones get commoditized with more scale on the production side? Obviously, the business value doesn't come from raw data or raw images. It comes from something more refined.

It's a really great question, and I think you're onto something there. When you look at DroneDeploy specifically, they do have the word "drone" in their name, but they are not a hardware company—they only do software.

They've built software to interact with various kinds of drones, including Skydio drones, and they've also built software to interact with ground robots such as [Boston Dynamics’s] Spot that can be used in a construction site for ground-based LIDAR inspection or other things.

I think they're really smart because there are a lot of companies doing what DroneDeploy started with, which is photogrammetry. Photogrammetry is where you take multiple images, it could be hundreds or thousands, and stitch them together to create a composite image that represents the scale of whatever you're mapping, or creates 3D models. There are many companies that do photogrammetry, like Pix4D and BentleySystems.

What DroneDeploy has done is say, "We don't want to do photogrammetry just for any industry because we don't think there's a ton of value there, and there's a lot of commoditization and competition at the horizontal photogrammetry level. So we're going to build a very bespoke photogrammetry solution for the construction industry."

To do that, they have to layer in things like building information model (BIM) data, maps, project-management data for construction sites. For example, track when a contractor is supposed to move a mound of earth from point A to point B. They've incorporated a lot of that data into their system with different layers. You can almost think of DroneDeploy as a system that enables a layer for your drone data, a layer for BIMs, a layer for structural models, and all of these layers and data can interact in a way that leverages AI to help construction companies make better decisions and go faster.

I do see that as being a really useful pattern for what companies in the space are doing. If you look at what Skydio is doing, it's not that different because Skydio is saying, "We want to build the best drones we can for the industries and customers we serve, but we also want to provide all the relevant software capabilities because that's a really important and margin-accretive part of the value chain."

For example, in the law-enforcement space, they've built very bespoke software products that you license along with the drone, which act as an operational center for not just people flying the drone, but people that might need to make decisions around what to do in certain situations when responding with a drone to a 911 call. How do you pipe that data? How do you zoom in on a suspect? How do you follow a suspect if they start running? How do you take a picture of a license plate?

There are lots of very bespoke use cases, scenarios, and capabilities. Those capabilities are all being built by Skydio because they recognize that you need to have something that really suits a customer for that specific vertical and use case.

In terms of first responders, I know that Skydio also sells a lot to the government. Did you learn anything about navigating government and military contracts while you were there that you think would be useful market knowledge?

Skydio is expressly a dual-use technology. They have made it clear to the market that they're not just providing drones for civilian use cases, but also for military use cases. But they draw the line where they say, "We will not weaponize the drones."

So the drones, even ones used by the military, are used in an ISR capacity—more for things like surveillance and reconnaissance, and less for things like dropping a bomb on a tank or piece of artillery.

The defense market at Skydio falls under the global government organization. It's not just the US buying these drones: it's also allied countries like Canada, Taiwan, Mexico, or countries in Europe as part of NATO that are looking to stop buying DJI drones. They want to buy drones made by European vendors or American vendors because that gives them more confidence that the data will not be used in ways that could harm them.

In terms of how you sell, each country is different. In the US, you end up partnering with companies that are doing procurement for the government, and that already have contracts with the government. They might have contracts with the DoD, NSA, CIA, or State Department. Essentially, you leverage those existing contracts in order to sell into the government.

You still have to build relationships with the end users and sell to them. But the contract vehicle used to actually do the sale process is often one that's already established by a government sales partner, which then Skydio and other companies use to sell to the government.

In that regard, it's not that different from what happens in enterprise software with companies like Salesforce or SAP. It's quite similar in that you are selling directly to the users and to the people that own the budget, but the actual contract process goes through a partner that the government has authorized to sell to the government.

That's a helpful analogy. Military and governments obviously operate a lot of infrastructure and facilities and have a lot of inspection and maintenance use-cases as well, I assume. I wonder if in these conversations around Pentagon or Homeland Security spending, Coast Guard, etc., if it doesn't get lost in the mix that even as a dual-use technology there are more prosaic use cases for drones like maintaining security on a base or making sure everything's where it's supposed to be, or helping them develop a new facility?

That's a really great insight, and that is actually what's happening. Base security, base inspections, ship inspections, aircraft inspections—these are all things that drones can help speed up and make more effective. So those are all potential vectors of growth within the government space for drone companies.

In terms of pricing, I imagine it's kind of a nightmare trying to figure out economics and pricing for Skydio and other drone companies that are combinations of hardware and software. I imagine there are some service components too in terms of implementation and getting these integrations up and running. Can you walk me through how that works at a high level? What are the risks for the vendor, in terms of not getting this right?

High level, how it works, and this is probably generally true of drone companies, is that you buy the hardware outright, and then you have a SaaS model for the software. So let's say you're buying five drones, you pay for the five drones upfront, and then you might have a three—to five-year contract where per drone, you pay for the software you're licensing or subscribing to for that vehicle. So you might buy five units of Skydio Cloud and five units of some other software that is relevant for that use case.

It is complicated and challenging because it's a fast-moving environment, and companies like Skydio are innovating quite quickly. There's also customers pushing back on pricing and saying "this doesn't feel right," which leads to reevaluation and changes. So there's a lot of change when it comes to pricing. Some of that is because it's a relatively new market and a relatively new product and a new sales cycle. People haven't been doing this for twenty years, so there's a lot of new stuff here in general, and that causes things to change more often.

The risks are that somebody might say your pricing is too high and try to build their own software or buy some off-the-shelf drones and use them and build their own software. So the build-versus-buy debate is something we often see with customers.

The interesting thing is that most companies or departments within companies that are buying drones—unlike when you're buying SAP or Salesforce—don't have very strong IT staff or dedicated IT staff that can work just with the internal drone team or the inspection team or the law enforcement team. So it is hard for them to build these things on their own and then maintain them over time.

It is something that comes up, but more often than not, customers understand that while it's not going to be cheap, if they were to try and do it themselves, it would likely be more expensive and potentially even not possible to keep up with the kind of innovation that Skydio can deliver.

Makes sense. Last question for you. If you had to start a drone-adjacent company today, what sort of market need or underserved niche would you go for? What comes to mind?

I can give you two examples: one on the hardware side, one on the software side. This is just my viewpoint, so other people might have totally different perspectives.

From my perspective, one of the gaps is around better radios. When you're flying a drone, let's say you're flying it on 5G with T-Mobile, ideally, you could have multiple virtual SIMs in the drone such that you could automatically switch if the signal from one carrier got weak but the signal from another carrier remained strong.

That is something we are starting to see on the cell phone side, but it's still really hard to do on the drone side, and it requires a mix of both hardware, software, and firmware. So that's one area where companies could be building better radios that are more resilient, more flexible, have more intelligence built in, and can tap into multiple networks and carriers. I think that could be really interesting.

On the software side, one of the challenges I ran into at Skydio is that drones capture a lot of data. If you send a drone out for inspection, you could easily be taking a thousand photos and lots of high-res video in the span of one flight. If you're doing lots of flights with lots of drones, you have potentially a massive number of image and video files, and there aren't very good systems to help customers manage that data today.

They're often trying to build their own image repositories or use something like Box or AWS S3 for storage. But there are a lot of things you could do in terms of analytics, discovery, search, and exploration—for example, you might want to compare the same photo taken two months or three months apart of the same asset. Those kinds of things are quite hard to do today.

I think there are opportunities to create better software that, given the exponential rise in media capture occurring because of more use of drones, can help manage that need. We don't yet have the best tools for this.

This transcript is for information purposes only and does not constitute advice of any type or trade recommendation and should not form the basis of any investment decision. Sacra accepts no liability for the transcript or for any errors, omissions or inaccuracies in respect of it. The views of the experts expressed in the transcript are those of the experts and they are not endorsed by, nor do they represent the opinion of Sacra. Sacra reserves all copyright, intellectual property rights in the transcript. Any modification, copying, displaying, distributing, transmitting, publishing, licensing, creating derivative works from, or selling any transcript is strictly prohibited.