Autonomous Vehicles: Investor Guide
Updated: Oct 11, 2020
Despite markets being highly inflated (relative to macro risks), the next decade will undoubtedly see incredible investment opportunities across rapidly growing emerging technology themes such as clean energy, artificial intelligence, cloud and edge computing, nanotechnology, space tech, satellite tech……and autonomous vehicles (to name but a few).
Read up on the article links above or, for a recap, you can refer to the AV market maps below:
The AV Market is Complex!
The first thing that stands out is that the market for AVs is incredibly complex with a huge number of dependencies and relationships. What’s highlighted above is also an extremely light version, mapping just some of the key relationships.
Note — Amazon have beefed up VERY recently — adding Zoox to their already impressive stable of Aurora and Embark subsidiaries/partners. The landscape is literally changing every day. But, as an investor, how do you make sense of all of this?
If you wanted to get exposure to AV in the public market, is it best to invest in those seemingly at the forefront such as Tesla, Google and Nvidia?
This may seem an obvious place to start, however, behind the ‘blue-chip’ techs are a swarm of often overlooked opportunities.
The aim of this piece is not to provide stock recommendations but provide a reasonably thorough overview of which public companies are likely to be at the forefront of autonomous vehicle and self-driving technology (and who appear to be in reasonably healthy shape in terms of debt and profitability).
This means some big companies have been left off the list.
For example, despite Google and Nvidia (both incredible businesses) being at the forefront of AV technology, it’s still a relatively immaterial part of their overall business.
You’ll notice as well that Tesla and Ambarella have also missed the cut. Again; both are incredible businesses, however, respectively they’re trading at 88x and 122x forward earnings (EPS). This means the market is pricing in exceptional long term growth prospects which may very well be reasonable, however, any company trading around 100x earnings makes me a tad anxious — especially in this environment!
This produces a list of 18 stocks of reasonable health that have the potential to benefit substantially from an ongoing shift towards vehicle automation.
Profitability vs Net Debt of AV Companies (Source: Granite Bay Capital, Yahoo Finance)
You’ll note from the above that there are a number of companies trading at a negative earnings multiple. This is a fairly redundant data point, however, for illustrative purposes highlights that these companies are, predominantly, early-stage and have R&D (and other expenses) outweighing cash inflow from sales. Most (if not all) companies in the R&D phase are quite advanced in their development, are generating revenue and have a path to profitability in the next 2–3 years. It’s obvious in the chart above that the smaller companies (smaller bubbles) are predominantly cash-rich and earnings poor, whilst the larger companies (large bubbles) are profitable, with reasonable debt exposure (the exception being Lyft).
The final point before diving into some of these companies is that the automobile market in the current environment is very risky from an investment perspective. We’re likely 15+ years away from seeing truly scalable Level 4 or Level 5 solutions, particularly solutions with V2X (vehicle to everything) capability. What we will see however is a steady increase in the incremental adoption of autonomous solutions. We’re already seeing this in LiDAR and driver monitoring.
What’s Under the Hood
The most advanced cars rolling off the production line today, such as the Audi A8, are filled with an array of sensors — including short and long-range Radar, ultrasound, cameras and the one best associated with autonomous vehicles, LiDAR.
If you’ve seen any picture of an AV (like the image of Waymo below) the most unique technology is the dome-like camera mounted to the roof of the vehicle. This is LiDAR. In the instance below, mechanical LiDAR which creates a 360-degree view of the environment.
Source: Waymo (Google)
The other is solid-state LiDAR, a fixed camera attached the front, rear and side of the vehicle. It is this compact, relatively inexpensive form of LiDAR we are currently seeing rolled out in the new Audi A8 (below).
As you can see — in this single unit — there are many various components (laser diodes, receivers, mirrors) supplied by various organisations.
However, LiDAR is but one component of the overarching sensor (and AV) ecosystem. For true level 4 and level 5 automation, all of the sensors (radar, cameras, ultrasound and LiDAR) need to effectively communicate with each other to provide the AV system with a seamless view of the entire environment. There are pros and cons to all sensors (individually) but combined — they can navigate almost any environment.
As well as sensor technology, there is an array of hardware and software required to develop the final solution including:
Processors and semiconductors
Driver and passenger monitoring
Interiors (touch screens, seating)
V2X Solutions (home connectivity, vehicle to infrastructure technology)
Mapping / GPS
The companies outlined below cover various aspects of these solutions.
Listed AV / ADAS Companies
LiDAR (3D Flash LiDAR, Camera With LiDAR, Short Range LiDAR), Cameras (mono, stereo, multi-function, surround view, rear, mirror view), Radar (long-range/short-range radar), control units, vehicle networking, brake systems, speed sensors, V2X communication, low-speed manoeuvring, cruise control, blind-spot detection, cruising chauffer, lane/speed/traffic jam and traffic sign assist). They are global giant in automotive components and will be central to advancement in ADAS/AV.
Source: Continental (L-R, 3D Flash LiDAR, Camera with LiDAR, Short Range LiDAR)
Magna is a well-established auto component supplier that does complete vehicle manufacturing and engineering, seating systems, chassis and fuel systems, powertrains and lighting. The reason they are included (and other diversified auto suppliers have not been) is that Magna have developed highly advanced AV platforms (MAX4) at Level 4 automation. Currently, Cheverlot Volt, the Cadillac ATS and Mazda 6 are fitted out with Magna’s AV platform (although Level 4 software is yet to be installed!). As well as the AV platform, Magna also produces camera modules, ultrasonic sensors, radar and a Solid State LiDAR.
Lumentum produce high-performance commercial lasers for use in optical communications, macro/micro material processing and consumer and automotive 3D sensing. Within automotive specifically, they offer a broad range of products for various use cases, all of which are expected to be mainstay solutions for all future vehicle production.
Osram is one of the world’s largest lighting manufactures and will be a major supplier to the AV market. They are in a JV with Continental (above) on the development of specific autonomous solutions and have further investments in the sector via a 25% stake in LiDAR tech company LeddarTech (along with Delphi, Marelli and Japan’s Renesas).
LeddarTech LiDAR platform. Module partner (Faurecia), LeddarEngine partners (Valeo, Aptiv, Osram, Marelli). Source: LeddarTech
Hardware & Software
Veoneer (VNE) Veoneer is a Swedish HQ’d company with an extensive product suite encompassing ADAS, vision systems, Radar, LiDAR, driver monitoring systems and ECUs. They call themselves a ‘pure-play’ company on autonomous vehicles after a recent spin-off from Autoliv. The company has active partnerships with leading AV organisations such as NVIDIA, Velodyne and Seeing Machines (see below) whilst Zenuity (their 50/50 JV with Volvo) is being brought in-house, to save Veoneer an estimated US$30-$40m p/a (~200 Zenuity staff will transfer to Veoneer). Financially, they’re sitting on ~75% Net Cash / Equity and are overall quite well positioned to take advantage of incremental shifts towards vehicle automation.
Lyft (LYFT) Most likely, the first time you will experience self-driving vehicle is through a ride-sharing platform and Lyft, in collaboration with Aptiv, are leaders so far with the largest commercial autonomous fleet in the US. As of February 2020, the two organisations had completed 100,000 paid rides in Las Vegas. However, of all the companies on the investment landscape, I would rate Lyft (and Uber) as the highest risk. R&D expenses are high, and the ability for ride-sharing companies to utilise the tech for margin improvement (i.e. driver redundancy) is a long-term play (at present, all autonomous trips have a safety driver behind the wheel). Further, Lyft (and Uber) will likely find themselves competing more and more with companies they simply cannot compete with financially (Amazon, Apple, Google) as they continue to gradually increase their exposure in the space.
Uber (UBER) Uber’s Advanced Technologies Group (ATG) is the AV unit of Uber, with investment from companies such as Toyota, Denso and Softbank and a core partnership with Volvo. The third-generation of its AV (developed with Volvo) will start testing on public roads in 2020. However, as with Lyft, it’s risky business. Uber are only EBITDA positive on their rides, however, the takings from that business is down 18.5% Q-Q. Uber Eats is reporting -US$313 EBITDA (adjusted) whilst ATG reported -US$108m EBITDA.
Aptiv (APTV.US) Aptiv is laser-focused on software for the autonomous vehicle market. This revolves around their Smart Vehicle Architecture (SVA) that pulls together data from LiDAR, Radar, cameras and other ADAS/AV systems. In partnership with Lyft they have provided over 100,000 paid rides self-driving vehicles in what they believe is the largest open-to-public commercial pilot. Further to SVA they’re also focussed on electrification, cybersecurity and driver monitoring. See more about their solutions in the clip below:
Blackberry (BB.US) Blackberry, originally known as Research in Motion (RIM), and developer of the early Blackberry smartphones have transitioned to being an enterprise software company with a core focus on cyber-security and IoT products and services. Within this is their QNX unit — a real-time operating system with a heavy focus on autonomous vehicles via Neutrino and the Blackberry QNX CAR platform. According to the company, QNX is used by 9 of the top 10 auto manufacturers and all 8 Tier 1 integrators (Aisin, Aptiv, Bosch, Continental, Denso, Harman, Magna, Panasonic, Visteon and ZF) work with QNX.
TomTom (TOM2.AS) Mapping platform TomTom, are powering the localisation, perception and path planning on AV software stacks; in particular via a collaboration with Sweden’s Zenuity (backed by Volvo and Veoneer) and Toyota. However, due to Covid19, 2020 revenues are expected to take a bit of a hit (i.e. auto manufacturer closures and significant traffic slowdowns). Certainly worth keeping an eye on though.
Telenav is working on several technologies within navigation and connected-car experiences. On the former, it provides navigation software to the likes of GM, including augmented reality navigation and ongoing development in chauffer enabled applications (i.e. calling your car from the parking lot to the mall entrance). Financially, they’re sitting on ~US$124m of cash and equivalents which is close to 50% of its market cap (so plenty of cash buffer for R&D and to ride out ongoing headwinds from Covid19). Furthermore, the company reported their first positive EBITDA quarter recently and are focussed on rolling out products such as In-Car Commerce to boost revenues.
Sensors / Monitoring
Visteon (VC) Visteon produces advanced heads up display and instrument clusters for almost all of the major automotive companies (BMW, Daimler, Ford, GM, Honda, Hyundai, Jaguar-Land Rover, Kia, Porsche, Renault Nissan Mitsubishi, Toyota, Volvo, VW). They are also very advanced in AV solutions with the development of DriveCore — a dedicated hardware and software stack with an onboard computer and external studio platform to analyse AV data. Their reasonably prudent management and balance sheet should see them ride out the current economic glut better than most.
Smart Eye (SEYE.ST) Smart Eye’s eye-tracking technology (and speech recognition) recognises who is driving and when they’re distracted, drowsy or intoxicated. The tech is built to integrate seamlessly with ADAS technologies and they are already well advanced in their R&D through collaborations with the likes of Volvo, NVIDIA, NXP, Neonode, Omnivision and Osram. An outline of their solution is provided below.
MicroVision (MVIS) MicroVision has a deep portfolio of IP in laser beam scanning technology which is applicable to AR micro-display (i.e. eyewear), interactive displays (embedded in smart-speakers), consumer LiDAR (i.e. home security) and automotive LiDAR. Specifically, on auto LiDAR, they are developing an integrated hardware and software module with real-time object classification. However, they are very deep in R&D mode, the cost of which ($3.6m for the last quarter) considerably outstrips revenues (at $1.5m for the quarter). As a result of this cash burn, and a challenging covid19 environment, the company are actively pursuing suitors for all or part of the company.
Seeing Machines (SEE.L)
As with Smart Eye, Australian HQ’d (and UK-listed) Seeing Machines have developed driver monitoring systems to detect driver distraction, drowsiness, “microsleep” events and driver engagement levels to then trigger events whereby AV operations partially (or in full) take control of driver activity. Their technology debuted in the 2017 Cadillac CT6 (below). As well as the GM collaboration, they have also recently partnered with semiconductor manufacturer Qualcomm and Swedish auto-company Veoneer. Despite headwinds in 2020, the company have implemented $12m in cost reduction measures, added $5m in recent pre-production revenue and have ~A$40m in net cash to maintain a level of balance sheet strength during the year. Revenues and R&D for 2H19 were both around the A$16m mark.
Seeing Machines driver monitoring technology debuted in the 2017 Cadillac CT6 (Source: Seeing Machines)
Neonode (NEON) Neonode have three primary application areas — touch sensing, gesture sensing and driver & cabin monitoring. The stock price is up over 4x YTD, particularly due to the application of its tech in a post-Covid19 world — touchless interactivity (i.e. elevators, menus, airport check-ins etc). The below clip as an example of the holographic interface which they will soon be rolling out in China. Although Neonode are mentioned here under the category of ‘autonomous vehicle’ the clear short-term application of their solution is in the commercial and retail building space.
It’s hard to write about Semiconductors in the mobility / autonomous space without mentioning NVIDIA. As you can see in the ecosystem map at the top they are THE company that ties everything together (next to Intel’s MobilEye unit), however, auto is just ~7% of their revenues which, at present, is dominated by data centres, gaming and other ‘supercomputer’ applications. Despite those being incredible application areas, this piece is about AVs, so (at 7% contribution) it doesn’t really fit the remit. They’re also trading at ~38x 2022e P/E. If they maintain an annual revenue growth rate of close to 40%, then that may be reasonable; but it’s still a heft price to pay — even for the market leader.
Ambarella (AMBA) is another which is well worth keeping on your watchlist as a pure-play semiconductor company focussed on AVs and AI camera applications, however, despite a healthy balance sheet, it has considerable R&D outlays and is loss-making. Even if that were to turn around in 2022 it’s likely still trading at >100x fwd P/E. Semi companies like Texas Instruments (TXN) and Xilinx (XLNX), also heavily focussed on the AV market, are trading at much more sensible multiple of ~25x fwd earnings. Both considerably different risk/reward profiles to AMBA though.
So, let’s take a look at some other stocks in the semiconductor space that are trading at slightly healthier multiples, and with a more core focus on AV (in terms of their revenue split).
Marvell Technologies (MRVL) If NVIDIA are the master in the semiconductor space, Marvell would be (one of) the protegees. In their words, “our [Marvell’s] products are essential to the required transformations the carrier, enterprise, data centre and automotive data infrastructure market must achieve to advance the emerging services and applications of the proliferating data economy”. Within auto, their ethernet and wireless solutions support HD video and sensor data in AV applications as well as in-vehicle networking (between sensors, ADAS, infotainment, body electronics etc). They are current partners of NVIDIA, as well as VW, Continental and Bosch, as part of the NAV Alliance (Networking for Autonomous Vehicles).
Renesas (6723.J) Japan’s Renesas is 9% owned by world-leading auto part manufacturer Denso (which in turn is ~33% owned by Toyota Motors & Industries). Their semiconductor solutions cover a wide range of industries including auto, building automation, energy, healthcare and network infrastructure (i.e. data centres). Within auto in particular, their technology is being applied to multi-camera sensing, radar, LiDAR, connectivity, sensor fusion, cognitive computing, security as well as their traditional areas of powertrains, safety, steering etc. They are also collaborating with one of the industry leaders in LiDAR, LeddarTech, on solid-state LiDAR solutions.
Synopsys (SNPS) If you’re not in the game of picking the likely winners of the semiconductor manufacturers, then a stock like Synopsys may be more your preference. Synopsys is a leading Electronic Design Automation (EDA) company and are well-positioned to take full advantage of the advancements in AV, AI/ML, cloud, 5G and other such sectors/applications. In fact, they are the world’s leading provider of solutions for designing and verifying advanced silicon chips, and for designing next-generation processes and models required to make those chips. They also have a broad portfolio of silicon IP, including pre-designed blocks of logic, memory, interfaces, analogue, security and embedded processors. Some of their partners/customers include Infineon, NXP, Xilinx, NVIDIA, Ambarella, Texas Instruments and Intel. From a financial perspective, they're trading at ~31x fwd P/E, have a healthy net cash position and profit margin. They will likely be key beneficiaries of the substantial growth in advanced semiconductors (in particular, as we see a greater shift of manufacturing back to the US).
The investment landscape for autonomous vehicles has incredible promise, however, as with any sector with high promise, has considerable risks — particularly during what is one of the most challenging macroeconomic times in our lifetime.
Despite Level 4 and Level 5 automation being some time away from mass adoption, there is no doubt that cars will increasingly become more intelligent with an incredible stack of hardware and software being adopted progressively over the years. A look at every car rolling off the market (GM, Volvo, Audi, Mercedes) confirms this.
The companies mentioned above are some of the many companies in this space who are contributing valuable technology to the space (and in many instances, technology which is also applicable across multiple high-growth areas including 5G, cloud and smart cities).
Although I won’t give specific recommendations here — I will advise that at this time it is particularly important to weigh up the hype of the sector vs the logic. By all means, get excited about the possibilities here, but it’s essential to look at cash burn, R&D spend, net debt, fwd earnings and the overall risk profile. For example, as exciting as NVIDIA is as a company, it’s trading at a considerably high multiple. Further, Uber and Lyft have huge promise, but for ride-sharing companies to get the margin improvement from AVs (i.e. to ‘retire’ drivers) it will take a huge amount of R&D and time which will need to be supported by other revenue streams (or ongoing capital contributions).
Happy (Cautious!) investing :)