Balderton

A primer on Drones and UAVs (Part 2): Where are the investment opportunities?

This post was written by Balderton Capital Summer Associate Animish Sivaramakrishnan, and was originally published on his Medium.

In Part 1 I talked about how best to segment the drone market, the size of the market and regulation.

Now in Part 2 I’ll dive into investment trends and discuss some of the major drone companies out there right now, before going on to talk about the various market segments in more detail and tell you where I think the investment opportunities lie.

Current investment trends

Here are a couple of different cuts of where investment has been going in the drone ecosystem [1][2]:

According to CB Insights in August 2015, 42% of all investment in the space since 2012 had gone towards software or services, and 40% towards hardware. This isn’t surprising when you consider that the drone software market is expected to eventually be as large as the platform market. We’ve seen in so many other initially hardware-driven markets how software rapidly catches up and takes over as the key value driver, once the hardware is relatively commoditized.

We’re also in the midst of a significant step-up in terms of the amount of drone investment taking place. The following chart from CB Insights makes this clear:

Looking at investment on a more granular level, the following picture emerges:

This might indicate that drone investment is slowing down. Admittedly, in dollar terms things have been quieter in 2016 than in 2015, but activity measured by volume has remained robust, and steadily rising since 2012. Part of the reason for large dollar figures in 2015 was due to sizeable investments in a few companies: DJI ($75m from Accel), Yuneec ($60m from Intel) and 3DR ($50m from Qualcomm).

Indeed, these three have been some of the biggest beneficiaries of drone funding. Below is a list of the most well-funded drone companies [3]. I’ve highlighted companies with a significant software offering (even if they are formally listed as hardware businesses). And as we at Balderton are most interested in European trends, I’ve specifically highlighted European companies.

Let’s talk a little about what some of these companies actually do.

DJI: By far the biggest drone platform manufacturer, dominant when it comes to both consumer and commercial drones (more on this later). Interestingly, DJI and investor Accel have teamed up to start the SkyFund, an investment fund for drone startups.

3D Robotics: Another platform manufacturer. Anecdotally, they’re pivoting to be more commercially focused in light of DJI’s dominance in consumer drones. They’re also trying to get more involved in the software stack, and have forged partnerships with the likes of Autodesk to allow commercial users to do mapping and analysis.

Yuneec: Also a platform manufacturer — both consumer and commercial drones. The Typhoon series is quite a highly rated consumer drone and at $1,300 is said to be a commercial level drone at consumer price. I’ve not used a Yuneec drone, but they’re meant to have quite advanced sense and avoid technology with Intel’s Realsense technology

Mapbox: It’s not really fair to put this down as a “drone” startup, as Mapbox is all about custom maps, with data taken from both open data sources such as OpenStreetMap and NASA, and from proprietary data sources such as DigitalGlobe. However, you can also feed in aerial data from drones and generate maps — but this is not the primary use case.

Skycatch: Their drone platform hasn’t yet been released, so for now they’re more of a software play. They’ve got an app which you can use to control a DJI drone and use to capture the data you need. They’ve also got an analytics dashboard to do things like overlay construction plans, annotate images, track progress etc.

Aeryon: A commercial drone platform specialist. Their drones are meant to perform well in harsh environments, extreme weather, high winds etc.

Airware: They want to be the universal OS for drones. Through some hardware plugins, you can plan flights, control a drone and make use of its various sensors / payload etc. Data collected is fed back to an aerial information platform, which can be used to carry out analytics on the data. They’ve also launched a “Commercial Drone Fund”, another fund for drone startups (which was actually launched on the same day as the DJI/Accel SkyFund).

PrecisionHawk: It’s listed above as a hardware company, but in reality it is both hardware and software. They do have their own fixed wing craft but are also very involved in the software play. They’re involved in the full software stack, with an analytics package focused on precision agriculture.

Drone Deploy: They offer a suite of software products covering the entire stack from drone control to image processing to analytics. The industries they tend to focus on are construction, mining and precision agriculture.

Kespry: Again, I’ve highlighted them as a software company because in addition to their own drone they do offer the full software stack targeting the construction industry

Sensefly: The French company Parrot has a controlling stake in Sensefly, which primarily makes fixed wing drones for commercial data collection. They do also have some involvement in sensor hardware (they have a multispectral depth camera), and the company is integrated such that Parrot can offer end-to-end solutions for particular verticals (e.g., the “Sequoia” solution for precision agriculture).

Delair Tech: They can offer both hardware (in the form of long range drones), or software for image analysis, or drones-as-a-service for commercial clients in construction, agriculture and infrastructure. Interestingly they are also able to mix satellite data for analysis

AGI: This is actually more of a satellite company than a drone company: it’s all about modelling / simulation software for satellites. For example you can model sensors and payloads, run simulations on the transmitters, receivers etc.

Pix4D: One of the most widely used 3D reconstruction / modelling software packages for aerial data. Pix4D also offers some basic analysis tools like volume estimation etc.

Market segmentation

In Part 1 I laid out my approach to segmenting the drone market, identifying the different product categories and customer types. Let’s now dive into each one and understand some of the trends and investment opportunities.

Hardware — Components

For a VC investor like Balderton, this is probably the least interesting part of the drone ecosystem. This is the part of the ecosystem that will get most quickly commoditized and offer the lowest margins. There are four key areas:

  1. Sensor hardware — Includes components such as camera systems, distance sensors etc. These sensors are not necessarily purpose-built for drones, and may be applicable to many other industries too. They’re often also developed by the drone platform manufacturer (especially for cameras)
  2. Power source — Most drones use commoditized Li-ion batteries (which are then white-labelled by platform manufacturer). Heavy lift drones are typically gasoline powered. Although we haven’t seen much of it yet, we might see fuel cells in the future — but the technology needs to mature first
  3. Propulsion system — Typically designed by the platform manufacturer
  4. Processor / chips — to process the data coming in from the various sensors and link drone hardware to software as efficiently as possible (e.g., low power). This area is dominated by large chip manufacturers like Intel (RealSense chipset) and Qualcomm (Snapdragon flight)

The difference between consumer and commercial drone hardware components is primarily in the level of sophistication of the components (e.g., commercial drones may require higher resolution cameras or heavy lift abilities)

Hardware — Platforms

On the consumer side, DJI is by far the largest drone manufacturer in a relatively consolidated space. DJI itself is valued at around $8 billion with around $1 billion in sales. Here’s what the consumer market looks like [4]:

DJI is credited for bringing drones to mainstream markets with its Phantom series of drone. 3DR and Yuneec are considered to be comparable to DJI, while Parrot is regarded as an easier-to-use option, targeting more casual users. GoPro is also expected to enter the fray with its upcoming Karma drone.

And while there is more diversity in commercial manufacturers, DJI is again the largest player. Commercial drones are defined as those used for profitable activities, and typically have longer endurance, higher lift capacity etc. than their consumer counterparts. The best way of tracking market shares of commercial drone manufacturers is to look at Section 333 exemptions, but there’s also a large “dark market” of unregistered commercial drones [5]:

Still, DJI is comfortably ahead of the rest of the pack. In fact, it’s suggested that if you take into account the dark market and consider commercial drones globally, DJI has closer to 70% share.

As consumer drones become more sophisticated, they’re increasingly meeting the requirements of commercial users. Parrot’s commercial drone sales accounted for 15% of their total drone revenues in 2015. Meanwhile manufacturers like DJI have “prosumer” lines such as their “Inspire” series which are targeted at both consumer and commercial users.

In summary — the consumer drone market is already heavily consolidated. Investment opportunities are in particular niches like self-following drones for action sports. On the commercial side, there’s a little more scope, but again, the investment opportunities are aligned with particular verticals — e.g., long-endurance drones for pipelines inspection. However, with the increasing sophistication of consumer drones, it seems to me that we’ll soon converge on a handful of commercial and consumer drone platforms which are used for nearly every activity.

Software — control and data capture

“Control and data capture” concerns the software you need in order to control a drone to capture aerial data. There are 4 elements, all of which may be done either by the platform manufacturer, or standalone providers (sometimes as part of a wider software offering):

1. Operating system (controlling the drone and its sensor payload etc.)

2. Flight controller (everything around controlling the flight of the drone, including things like the autopilot and sense-and-avoid technology)

3. Mission / flight planning (planning/mapping a route for the drone to execute)

4. Fleet management (managing fleets of drones — still quite a nascent field)

But what are the prospects here? Specifically, how can you differentiate and beat the market? Can the “best” control system win, or do you need to be just “good enough”?

I believe that in most use-cases, being “good enough” is sufficient. Consequently, the space will be won by either the platform manufacturer or the first independent provider to develop partnerships with the right (i.e., most popular) platforms. It’s also a crowded space; there are a lot of companies working in this area. Having said that, there might be room for differentiation in particular verticals. For instance, a high degree of control is required for drones inspecting offshore O&G platforms, or surveying underground mines. In such use-cases, GPS can’t be relied upon and sophisticated control technology is needed — so it may be possible for software providers to carve out niches in particular verticals.

Software — image processing

The next part of the software stack is all activities around image processing. This principally consists of two things: 1) 3D rendering and 2) Object recognition.

3D rendering has more application in a commercial setting (vs. consumer) and is considered to be quite commoditized. It’s basically all about stitching together aerial images into one cohesive 3D model that can be interacted with. In some cases, it’s done by the platform manufacturers; in other cases, by companies sitting across the entire software stack; or alternatively by 3D modelling specialists — Pix4D and AgiSoft are dominant when it comes to this.

Because it’s such a commoditized aspect of the software stack, I don’t think this is an interesting investment opportunity — outside of one particular use case which is real-time reconstruction. No-one has yet cracked this, but if a drone could build a 3D model in real-time, that would be a game-changer.

Object recognition on the other hand is a much more interesting (and also harder) problem. For instance, in a consumer context it would enable a drone to to autonomously follow a user. There are a bunch of consumer platform manufacturers trying to develop this capability from the likes of DJI to pre-launch startups like Lily.

In a commercial context, object recognition would facilitate advanced analytics — like being able to identify different crops and measure their growth (in precision agriculture). Again, this may be done by the existing software analytics companies or by startups focusing on image processing.

In any case, as a machine vision problem which hasn’t yet been fully cracked, and with a host of relevant applications, it’s a very interesting area and one to watch closely.

Software — analytics

Although there isn’t much of a consumer use case, in my opinion drone analytics software is the key commercial market opportunity.

More so than other elements of the software stack, data analysis is sector-dependent, thus there’s opportunity for differentiation. There are also so many different sectors that drones can be used in — just take a look at the Section 333 exemptions by end-market [6]:

The key point is that each vertical has very different analytics requirements. For example, a farmer using drones for precision agriculture might be interested in measuring plant densities or being told when and where to apply fertilizer to optimize growth. On the other hand, a construction site using a drone would be more interested in surveying the land, measuring soil movements and comparing the progress of the site to architectural plans. Similarly, the analysis of the aerial data collected by drones in industrial inspection, pipeline inspection, O&G, mining, security, mapping etc. is quite different in each case.

Currently the analytics is done by either:

· Platform manufacturers: well-known platform manufacturers like 3DR and Parrot offer analytics through partnerships with Autodesk and Airnov (respectively), while others like Kespry and Delair manufacture their own drone and build out the entire software offering as well.

· Standalone analytics providers: aerial data is captured through some other means and then the software provider carries out the analytics

· End-to-end software providers: i.e., software providers who are involved in everything from control and data capture through to analytics.

The end-to-end software provider approach is most common. And already we’re starting to see some large players emerge:

Supporting Systems

Drone supporting systems are a disparate set of products, but there are some interesting prospects. I see 3 key areas for venture investment:

1. Drone marketplaces: There are several types of drone marketplace — those for drone bookings (whether pilot / drone / or “Drones as a service”); those for drone equipment; and those for aerial data (which may also encompass satellite data). Of these, the drone booking market may be riskier in the long run. While on the one hand drones are expected to become more ubiquitous and there will be an increasing need for customers to get access to drones, cheaper and easier-to-use drones may eventually result in businesses owning and operating their own drones rather than outsourcing to a specialist

2. Airspace regulation: This is all about enacting preventative measures to avoid drones flying into restricted / sensitive airspace e.g., near airports. It consists of working with government authorities to regulate drone airspace by setting no-fly zones and detecting drones, as well as with drone manufacturers to control drone movement. Another part of this is working with users to help them understand when and where they can fly their drones (increasingly relevant as commercial drone use becomes more widespread) and helping them get permission from the various authorities when planning missions

3. Anti-drone technology: The logical next step to airspace regulation — countermeasures to disarm drones after they fly into restricted airspace. This might range from remotely taking control over drone mid-flight, to aggressive anti-drone systems such as nets or lasers

Drones as a Service

Finally, we come to the “Drones as a service” market — which really sits across many of the segments I’ve outlined. In this model, a drone might be booked through the service provider, which would then work with the client to collect and analyse data — feeding back the findings to the client. The platform itself may be bespoke and tailored to the client, or may be an off-the-shelf platform like a DJI.

However, I believe this model has numerous disadvantages and is not a long-term investment opportunity. The model requires the company to be involved across the entire value chain, and meaning that it is difficult to develop a competitive advantage in any one area. Working closely with the client means that it is a business model that is difficult to scale (rather consulting-esque). It is also a very crowded market at the moment and there is a plethora of 2–4 man bands offering drones-as-a-service. But most significantly, as the price of drones goes down, commercial users are increasingly likely to own and operate their own drones (outside of some highly specialized applications).

Admittedly there is some way to go. From a cost-benefit perspective, only a tiny proportion of farms would presently benefit from owning their own drone currently and forecasts suggest that industrial inspection, construction and precision agriculture will be the largest beneficiaries of drones-as-a-service in the near term. These are all large end-markets. But for an investor looking at long term returns, the recommendation would be to look elsewhere

Well, that concludes this outline of the drone market. I hope it’s been useful in helping you understand the landscape, and investment trends and opportunities. Thanks again to the whole team at Balderton and James Wise for their support through this!

References

[1] http://drones-startups.silk.co/

[2] CB Insights

[3] Drone Industry Insights

[4] Goldman Sachs Investment Research

[5] Drone Analyst

[6] FAA Aerospace Forecast

Stay in touch with Balderton

Sign up for our newsletter to stay up to date on news from Balderton, and our portfolio.