VC funding in mobility hit a new high in 2020 with $8.4 billion invested in the sector, up 33% year-over-year. Within the sector, autonomous vehicles have shown immense promise—attracting funding and exciting investors and entrepreneurs. This week on the "In Visible Capital" podcast, we talk with Joy Tuffield, partner at Generation Investment Management, and Luminar CEO Austin Russell about recent advancements in self-driving cars and the remaining barriers to ubiquity. Explore more of Season 2 and subscribe to get new episodes of "In Visible Capital" every Tuesday on Apple Podcasts, Spotify, Google Podcasts or wherever you listen. For inquiries, please contact us at firstname.lastname@example.org. Transcript Joy Tuffield: Since autonomous systems are really hard to build as it's kind of evidenced today, the companies that ultimately end up successfully building them will end up accruing a significant amount of value. They will be, likely, really high-margin businesses that will have really significant moats around them. Lee Gibbs: Welcome back to the show. I'm your host, Lee Gibbs. As someone whose job it is to stay up to date on industry trends, I work closely with PitchBook research analysts and industry professionals to understand market drivers across private and public capital markets. Recent years have shown significant investment in growth for the autonomous vehicle space, resulting in tech advancements that have helped propel the industry forward. Venture investment into autonomous vehicles hit a record high in 2020, with over $8.4B deployed into the sector. That's up 33% year over year. And Q4 of 2020 showed strong performance with over $2.5B invested. That's up 227% year over year. Today, we're welcoming Asad Hussain to the show. Asad is an emerging technology analyst at PitchBook who focuses on mobility as a pandemic has gone on. Asad Hussain: What's been interesting is that, yes, some subsectors within mobility have seen and continue to see pressure areas like public transportation and ride hailing, for example. But others are actually seeing a boost as a result of the pandemic. So areas like last-mile delivery, for example, continue to be really strong. Micromobility since it provides a sort of social distance form of transportation, and also electrification. We're seeing a lot of growth there. And as it relates to tying it all back to autonomous vehicles, I think we've continued to see really strong investment there. Joy: I think if you take a long term perspective, it's pretty clear what is driving excitement and investment in the autonomous vehicle space firstly has the potential to be a huge market. Lee: Joy Tuffield is a partner at Generation Investment Management. Joy: Generation is a pure play sustainable investment firm. So we have $30B roughly of assets under management. And I'm a partner in a growth equity fund. And here we're investing in mission-driven companies that are accelerating the transition towards a fully sustainable economy. Lee: In 2020, the total market size of the autonomous vehicle industry was around $4.2B, according to PitchBook, and the potential impacts are far reaching. Joy: It has the ability to transform pretty significant chunks of the economy by providing a low cost, 24/7 alternative to labor, essentially, and so on. The obvious end of the spectrum, that's things like fleets of autonomous vehicles transporting people around cities at a lower price point than an Uber, for example. And if you just think about how much additional demand Uber ended up providing by providing a lower cost alternative than traditional taxis, you kind of begin to start understanding the reason that people are excited. And then it's not just people, right? It's the transportation of goods, again, either in cities or in the middle mile, last mile long haul. Lee: And with the rapid shift to e-commerce brought about by the pandemic, the outlook for autonomous delivery is strong. According to PitchBook, global revenue for last mile delivery services reached approximately $347B in 2019 and is forecasted to grow to $579B by 2025. And this market expansion will drive opportunities to invest in automation, reducing delivery costs and expanding the market for delivery. Asad: We're seeing a lot of growth in last-mile delivery services and also autonomous delivery solutions. So companies like Nuro, which are actually looking to automate and automated drone companies like Flytrex and Matternet. And delivery services continue to see an outsized amount of funding going into the space. And I think it says a lot about the level of demand and the level of growth that these companies have seen due to an expansion of the market. Joy: There's also much more niche applications that are coming up. So there's autonomous vehicles that will likely be deployed in the construction industry in different parts of the supply chain from like managing trailers in yards through to managing and moving containers around ports. But not only does it have the potential to be a really big market, but since autonomous systems are really hard to build, as its kind of evidenced today, the companies that ultimately end up successfully building them will end up accruing a significant amount of value. So they will be likely really high margin businesses that will have really significant moats around them. Lee: This high value potential has automakers and technology companies investing heavily in autonomous technology. In July of 2019, Cruise Automation, a subsidiary of General Motors, received $3.4B of development capital from Softbank and GM. Volkswagen committed $2.6B into Ford-owned Argo AI, and leading developer Waymo, a subsidiary of Alphabet, received $3B in a late stage VC round in May of 2020. Joy: Waymo has taken a pretty capital-intensive route into developing an autonomous vehicle. Lee: Subsectors with an autonomy include autonomous software, autonomous hardware and full-stack solutions, which is the approach Waymo has taken. Joy: And that means they are building the entire suite of software themselves internally—from the sensing and mapping, through to the perception and the decision-making controls, middleware. There are merits to that kind of luxury, and that capital-intensive approach, because it gives you a time-to-market advantage. It's quicker to do. You're not relying on external providers to provide you kind of elements of it. It's fully within your control, but it's also very capital intensive and it requires a significant amount of very highly specialized niche people in a bunch of different areas coming together and collaborating. Joy: If you think about Waymo, it essentially had the luxury of being able to go down the really capital-intensive route because it was born out of arguably the most cash generative business of all time. And so it had a stream of capital that was available to it in order to execute on this potential. And what's slightly taken us by surprise, I would say, is that the rest of the industry has followed suit. Lee: Other large tech companies that have invested in full-stack solutions like Waymo include Amazon with their acquisition of Zoox for $1.3B in 2020, and Intel, who acquired Mobileye for $14.9B in 2018. Asad: The companies that are larger and have access to a greater amount of capital, whether it's through a larger balance sheets or greater ability to invest billions of dollars not only today but over the next five to 10 years, those are going to be the ones that really dominate this space. And as I look out, I see the landscape of tech companies, strategic corporates, transportation companies, automakers, and I really think the tech companies are in the best position to really shape the space going forward and invest in self driving. And so by tech companies, I mean companies like Google with their Waymo unit, Amazon with Zoox, Apple, Intel and others that I think are really going to be instrumental to driving the adoption of this technology. So I really I'm most optimistic about tech companies solving this broader robotaxi the opportunity. Lee: Although Asad is optimistic that these large tech companies' big bets will pay off, Joy has reservations. Joy: We found at least that it's a bit of a head scratcher from a private investor standpoint. And it's a bit of a head scratcher because it's kind of unclear how much capital is required. You know that it's in the order of billions, but is it a billion? Is it 10? Is it 50? Who knows? And on top of that, you're still taking on tech risk and you still don't quite know if that technology is going to materialize with the kind of group of people that you're investing in. And if it does, when it does. Asad: I think there's been a bit of a reset in a lot of people's expectations as far as the timeline is concerned, but also a bit of a reset as it comes to the sheer amount of capital it's going to take to bring these self-driving cars to market. And the complexity of the technological problem is a lot larger than I think many people thought it would be a few years ago. Joy: So let's start with the hardest problem, which is a fully autonomous vehicle that's carrying passengers and is let loose in a city among the kind of general population. So that's like the hardest problem to solve. And the real challenge there is very much a tech challenge still. We need to be able to build an autonomous vehicle that senses things accurately and can make decisions off of that effectively in a safe manner, and that needs to be able to be done in all different types of weather, in different geographies, and as well as incorporating a huge number of cases of weird and wonderful scenarios that will likely happen at some point during the kind of journey of that vehicle. Austin Russell: The key distinction here when it comes to autonomous vehicles is building something that can achieve the performance and safety that's needed, something that you can put your life into the hands of. Lee: Austin Russell is the founder and CEO of Luminar, an autonomous vehicle sensor and software company. Austin founded the company eight years ago with the goal of building a new type of LiDAR sensing system capable of solving the problem of how autonomous vehicles make sense of their unpredictable, real-world surroundings. Austin: From a LiDAR standpoint, it stands for light detection and ranging as the most fundamental level. You send out pulses of light into the environment from a laser, hits an object, comes back, and we're measuring effectively the speed of light, you know, for how long it took to get that object come back. And then doing that millions of times, you know, every second, so to say, or per minute. For most industries, 99% accuracy of something, 99% detection accuracy, that's great. If you have an iPhone LiDAR, you know, for facial recognition, as long as it works most of time, you're pretty good. For this, 99% is completely unacceptable. You need 10 nines with a reliability to achieve the right level of safety. Lee: And just to clarify, by ten nines, he means a one in ten billion chance of failure. That's where the level of performance that comes into place. You can accurately understand what's going on around you, safely interpret the environment and be able to navigate accordingly. And that's how you can get to a point where these aren't just assisted driving systems, they actually evolve into autonomous trading systems where you don't have to be constantly paying attention, ready to take over the wheel of the system at any given moment? Whenever it makes a mistake, you need to have that threshold work substantially safer than a human level capability. And that's what we're finally enabling. And there's really no reason why people in this day and age should be getting into accidents in their vehicles. Pretty much like 95% of them are caused by human error and it's totally preventable. And if we can if we can solve this problem, it means a lot. Lee: Although historically, a significant portion of VC investment has gone towards startups developing full-stack, autonomous solutions, this dynamic is changing as investors concentrate more capital and companies focus on single aspects of autonomy—a trend that has translated to the public markets, with Luminar's share price nearly doubling since it went public through SPAC in December of 2020. Austin: We got a lot ahead of us, but what it means for the industry is really just everyone's always talked about autonomy. There's been a lot of hype around it. And I think it led to some level of disappointment, like, you know, where are we? Like, you know, we were promised these level-five robotaxis driving around everywhere, you know, picking people up, dropping off, doing all the stuff. That totally didn't happen. And this is where we made a strong bet on the existing industry of production cars and trucks, and working with the right ecosystem partners in that domain to be able to see this through. And that's where it's certainly been paying off, because what you can do is you can start to solve more constrained applications. We don't have to have our cars drive around everywhere all the time. You know, in every urban environment globally, like we're starting out, for example, with more constrained highway autonomy use cases of the product for for freeways, and this is where you can take your hands off, eyes off, read a book, use your phone, work on your laptop, watch a movie, take a nap, etc, exit to exit. And it makes a huge difference. Lee: And Luminar leveraging its partnerships with companies like Daimler Trucks, Mobileye and Volvo to achieve this state. Austin: With Volvo, these guys are, of course, historically associated and certainly the leader in safety. They've taken big technological leaps and spearheaded it for the industry, everything from the introduction of the modern seatbelt all the way to the introduction of, talk about Mobileye, their systems originally on their cars. And that's what what really makes a difference, because you see these kinds of technologies and ultimately they get standardized throughout the broader industry and achieve a new level of safety standard. So that's what you can drive towards. Lee: Of course, Luminar isn't the only company taking the partnership approach. One of Generation's portfolio companies is developing a complementary solution to the problem of navigating the unpredictable world by giving autonomous vehicles a baseline of what they should expect. Generation was a lead partner on a $60M Series B for Palo Alto-based DeepMap in twenty eighteen. DeepMap provides highly detailed maps to a centimeter level of precision to ease the processing burden for autonomous vehicles. Joy: We essentially had a thesis a Generation, which was that if you look at history essentially and you think about other software stacks out there, they ultimately end up rationalizing and the long term approach doesn't seem to be the full-stack approach. Over time, bits of the software stack end up getting divvied up, and best-in-class third-party providers end up providing best-in-breed solutions. And we have a view that over the long term, the same thing will happen in the autonomous vehicle trend, and that there will be certain elements of the software stack that are more likely or deserve to be serviced by inbreed providers, either because there's network effects at play or the specific bit of functionality that they're providing is really niche and not necessarily in the comfort zone of the autonomous vehicle developers. And mapping was certainly a part of that. There's mapping and a whole host of other elements of the software stack from middleware, through to simulation, through to verification testing and a whole host of other things. The mapping was that really a crucial part of the autonomous stack that we got excited about. And that's why we ended up investing in DeepMap. Lee: DeepMap offers their partners what Joy says is a crucial component of autonomous vehicle technology. Joy: They end up embedding their software into autonomous vehicle developers and the vehicles of those autonomous vehicle developers, and they end up building a real-time, scalable and economically feasible HD map, which again is kind of critical in the development of autonomous vehicles. The other reason that we're really excited about DeepMap, was that the incumbents were pretty much nowhere on this front. So if you think about the likes of HERE or Tom-Tom, these large companies which have a choke hold, particularly in the automotive industry, of mapping, had really not invested in this capability and did not really have the technical mass to be able to achieve it. And on top of that, DeepMap is a really capital-efficient business, so it doesn't require billions of dollars in order to be able to be commercially viable at the end. But it is getting revenues right now from autonomous vehicle developers out there, helping them with their engineering services, and over time will end up being a high-margin, very sticky software business. And so we thought it was the perfect risk-return for us. Lee: Even with companies like DeepMap and Luminar helping to advance autonomous vehicle technology, there are still some very challenging navigation scenarios the industry will need to solve. Imagine a situation in which a delivery truck is double parked, narrowing the street while a pedestrian peeks their head out for an opportunity to cross, but not at a crosswalk. Joy: The way that a human would end up reacting to that would be to essentially look into the person's eyes and and somebody would nod either like the driver would nod and say, "You can cross." Or the passenger would nod or wave you on and say, "Feel free to go." And then the delivery man is coming and he's waving like it's totally fine. And you maneuver your way through that via human interactions. Now, in the point of a kind of autonomous vehicle, like, you need to figure out those kind of nuanced ways where there's not rules of the road really being abided by. And that's just an example I can think of. I'm sure there's like thousands of those that will ultimately end up taking place. Lee: And there are other human interactions that are crucial to autonomous vehicles being ubiquitous—the kind that involve lobbyists and lawmakers. Joy: I think in terms of regulation, the area where I suspect is going to be the most interesting to look at is what the testing and verification is going to look like, because as you all know, vehicles in general are heavily tested and regulated and there is a huge industry around it, which is mostly around the structural safety of those vehicles themselves. And a whole new industry is going to have to be created to verify the safety of the control systems that are dictating how it's going to be driven. So I think there's probably going to be a whole host of really interesting opportunities there in terms of figuring out what a standard safety protocol will end up looking like, and companies out there, which will end up verifying and testing those and delivering safety standards associated with autonomous vehicles. We're obviously not there yet, but that might be a really interesting area to take a look at. Lee: One area of mobility that has a lot of experience with regulation is ridesharing. And giants like Uber and Lyft are also looking ahead to the future of autonomy. Lyft announced partnerships with autonomous companies Waymo and Motional for its self-driving initiative Level 5. And Uber's autonomous arm Uber Advanced Technology Group was acquired by Aurora Automotive for $4B this past December, a move by Uber that has Asad feeling cautiously optimistic about the company's future. Asad: I think it will allow them to focus a bit more and refocus on their their North Star, which is their network. And going forward, I think what makes the most sense for them is as they think about self-driving and the evolution of the industry, self-driving is going to happen and it's going to be an existential risk to incumbent ride hailing providers. And so the question becomes, how do you adapt to that? If you're not at a place where you can build it in-house, what do you do then? Well, what we've been saying is that they should seek to establish partnerships with the leading companies in the space, kind of what DoorDash is doing with GM Cruise. I think Uber should be working as hard as they can to get Waymo vehicles, to get Zoox vehicles, to get Cruise vehicles on their network so that when the time comes, when these self-driving vehicles are coming to market, Uber's network is the go-to spot for them to deploy. Lee: And there's definitely incentive for ridesharing companies to win in the self-driving space. As leaders and the robotaxi industry have the potential to radically reshape the world of transportation. PitchBook reports of the space could represent a multitrillion dollar market opportunity. By not requiring a driver, these vehicles will be able to lower the cost of transportation well below a dollar per mile in major urban cities, thereby drawing consumers away from personal vehicle ownership, ride sharing, car rental and even public transportation. Joy: You now have a fairly limited set, I would say, of autonomous vehicle projects out there that are well funded and are in the hands, either of the OEM's or Waymo, and there's probably two or three that are very well funded independents. The high likelihood is that a set or kind of subset of these companies will end up being successful and will end up being really highly valuable companies. It's just hard to say which set is going to succeed and again, how much capital is required to get them there and when they will ultimately get there. And so it's quite a difficult thing to wrap your head around as a private investor looking to put capital in, particularly when you essentially have a fund, and you are competing that investment with a number of other investments out there, which you can more concretely understand the risk-return perspective and frankly, which can be maybe not just as transformational, but very significantly transformative in the industries in which they're operating. Lee: And while the timeline for wide scale adoption for self-driving vehicles that can transport people is still a question mark, we could be much closer to scaling these vehicles for the transportation of goods. Asad: So we think middle-mile and logistics-focused autonomous vehicle solutions will come to market faster than those serving consumers. So basically, if you're moving goods, you're probably going to go to market faster than if you're moving people. And there's a few reasons for that. I think there's a less regulatory overhead, less scrutiny around that space, because if things go wrong, people's safety isn't at risk. So I think the hurdles that you need to hit, as far as, is this a product that works 99% of the time or 99.9% of the time, or 99.99% of the time? That sort of hurdle, it comes down a little bit when you're moving things instead of people. So that's the number one thing. I think passenger experience is a huge factor, and that's not really a factor when you're moving boxes around. If people get into a self-driving car and they have a bad experience, they might not be very likely to use that service going forward. They might go to a competitor. And I think that's something that's top of mind for many of these providers as they look to go to market. And then beyond that, I think the technological problem. So not only is the hurdle rate lower for these sort of pragmatic reasons, but also it's a much more structured use case environment, so the technological hurdles are lower if you're operating in a warehouse with trained workers or a logistics or an autonomous dockyard or something, yes, there might be people walking around, but they're probably trained workers that know how to navigate around these robots and this machinery and won't take any unexpected risk. That's a much different problem to trying to solve self driving in a dense, crowded urban center where you have pedestrians, jaywalking and doing things in a very unpredictable way, and also dealing with other drivers on the road as well. That's a that's a big part of what makes this technological problem so complex on on public roads is dealing with pedestrians and other drivers. And so if you don't have that with a off-road environment, I think you'll be able to solve this problem much quicker. Lee: While Asad reports that he expects long-haul highway applications of autonomy to arrive in the mid 2020s time frame, Joy is a little less certain. Joy: People are still going to be driving at incredibly high speeds on those highways. And on top of that, you will have a very large vehicle that will be traveling at incredibly high speeds. And so actually, the risk level of a small thing going wrong having kind of catastrophic consequences could be pretty significant on the highway. Lee: Bringing it back to our focus on emerging technologies. Much of the innovation we've discussed on prior episodes has been realized by startup and early-stage companies. And despite his assertion that the big tech companies in this space will ultimately win out, Asad still sees opportunity for smaller players. Is there a space for startups? Absolutely. I think, you know, the sort of full-stack, full self-driving problem is probably going to end up being solved by the tech companies. But I think there's a lot of room for startups to play, and even startups solving this problem like Aurora and Nuro, for example, which are still independently backed and don't have a major corporate backer or haven't been acquired as of yet. But some of the more earlier stage startups, I think, will have an important role to play in terms of some of the sensing technology, for example, LiDAR, some of the augmenting technology, for example, like simulation software and and perception tagging and things like that to help sort of augment this broader ecosystem of self driving. And also, I think startups will play a key role in bringing off-road applications of autonomy to market, which I think is really important. So looking at things like construction and mining, automating logistics, automating airports, for example, those are really nice spaces. But there's also a faster sort of passive commercialization. And so, we're optimistic about companies like SafeAI and Built Robotics, actually retrofitting heavy machinery and bringing those automated solutions to market. I think a lot of investors are really looking through that and they're recognizing that the existing drivers for next generation mobility, technology and self driving, the need for cost effective, affordable, convenient transportation solutions, low emissions transportation, those will be persistent, and we expect them to be persistent going forward. Lee: Regardless of which technologies win out, we're well on our way toward having autonomous vehicles impacting the way we move. And beyond technology, there is implementation. Join us next episode for a closer look at the role cities play as stakeholders in implementing emerging technologies in the mobility space. In this episode Joy Tuffield Growth Equity Partner at Generation Joy Tuffield joined Generation in 2013 and is a member of the Growth Equity strategy, focusing on enabling technology investment without planetary and societal trade-offs. Tuffield has led the firm's industry roadmaps on everything connected to transportation, from electrification to autonomous vehicles to supply chain software, and spends a lot of time talking to truckers about how their lives are changing with technology. She led Generation’s investments in DeepMap, Asana, Gogoro and Convoy, where she also sits on the board. Tuffield was previously an investment banking analyst at Jefferies International, where she focused on the Oil & Gas industry and before that, she conducted research on resource and climate security in the Energy, Environment and Resources Department of Chatham House (The Royal Institute of International Affairs). Tuffield graduated with honors from St. Hilda's College, Oxford University with a Masters in Physics and Philosophy. Austin Russell Chief Executive Officer at Luminar Austin Russell is an engineer, entrepreneur, and LiDAR industry pioneer for Autonomous Vehicles. He founded Luminar at the age of 17 while pursuing research in optical technologies and exploring applications across the medical, augmented reality, robotics and automotive industries. After a brief stint in the Stanford applied physics department, Russell accepted the Thiel Fellowship to pursue his vision of developing a new kind of sensor technology with a mission to make autonomous vehicles both safe and ubiquitous. In 2017, Russell was recognized as one of MIT Technology Review's Innovators Under 35 and Forbes' 30 under 30. Asad Hussain Emerging Technology Analyst at PitchBook Asad Hussain is an emerging technology analyst at PitchBook, where he contributes to the company's emerging technology research covering the mobility tech and supply chain tech verticals. Hussain leads PitchBook's mobility and transportation tech coverage, which includes sectors such as autonomous vehicles, ridesharing, car-sharing, micromobility, last-mile delivery and fleet management, among others. His expertise has been regularly featured in top media outlets including CNBC, CNN Business, Financial Times, Forbes, Fortune and Reuters. Hussain has also appeared on Bloomberg TV's "Balance of Power" and "Bloomberg Technology" segments to discuss emerging trends in ridesharing, autonomous vehicles and food delivery. Prior to joining PitchBook, Hussain was an equity research associate at Westwood Holdings, where he supported the coverage of 100+ publicly listed US companies within the technology and industrial sectors.