Austin Burch, CEO, Above Robotics
Ezra Goldman, CEO, Upshift
Bibhrajit Halder, CEO, SafeAI
Daniel Hoffer, Managing Director, Autotech Ventures (moderator)
Read below for a full transcript of the conversation.
Andrew Holtz - Hello and welcome, my name is Andrew Holtz, I'm a Senior Relationship Manager based in Palo Alto working on our Tech Banking team. Thank you for dialing in for our latest webinar. We're excited to be presenting today's topic of revolutions in the mobility industry. While I'm not an expert in this field, we've assembled an incredible panel who are. They're specialized in various areas of the mobility industry and are here to discuss the current market insights and the future of transportation and mobility tech including connectivity and autonomous vehicles. To facilitate the conversation, we have my friend Dan Hoffer, Managing Director of Autotech Ventures, who will moderate the session, along with Austin Burch, Chairman and CEO of Above Robotics, Ezra Goldman Co-founder and CEO of Upshift. Bibhrajit Halder Co-founder and CEO of SafeAI. Before we get started just a quick housekeeping note. We welcome all questions and we'll be addressing them toward the end of the session. Feel free to submit them along the way using the Q&A feature at the bottom of your screen. And now I'll turn it over to Dan to tell us a bit about his background and introduce the rest of the panel. Okay.
Daniel Hoffer - Thank you, Andrew. Appreciate the introduction. My name is Daniel Hoffer and I'm a Managing Director at Autotech Ventures. And Autotech Ventures is based here in Silicon Valley. We are a early stage venture capital fund that focuses exclusively on the global ground transportation market. And I will share a few slides regarding autotech but also trends that we're seeing in the ground transportation market. And then we'll open up the introductions to to other folks around the table. Can everybody see my screen? I believe so. So we focus at, we focus on startups that are at the intersection of technology and ground transportation and our thesis is to focus on five different themes in the industry. And these are macro trends that we're seeing play out across the entire industry. And we believe that anytime there's rapid transformation and disruption in an industry that founders can make a lot of money and investors in those companies that are changing the world and riding this wave of innovation can also do very well for themselves. So our focus is we call this the CASED acronym CASED stands for connectivity including telematics and connecting vehicles to the cloud autonomy and autonomous driving that's going to be a big focus of our session today. Sharing the sharing economy, which includes the sharing of underutilized fixed assets, electrification the transition to electric vehicles, and finally digitization of enterprise, which basically move, basically means moving old school processes and pen and paper based and fax machine based approaches to online digital versions. Autotech Ventures has two funds under management. We've got approximately 275 million. Overall, our approach is to take an early stage focus on Seed and Series A deals.
We're often the lead investor, mostly focused on the U.S although we look at other countries and regions of the world as well. We do early stage Seed and Series A, we also do a mid stage Series B and C deals. Overall with initial checks of tends to be one to $6 million for the initial check. We have partnerships with many of the world's largest transportation corporations. We work with over 25 corporate LPs that are typically publicly traded multinationals representing almost a trillion dollars in market cap and operating in over a 100 countries. We connect our portfolio companies with our LPs whenever there's an opportunity for a win-win which is quite often. So we've facilitated many relationships between startups and our corporate LPs that have resulted in partnerships, investment opportunities, and customer relationships. Overall, we believe that this is an exciting industry to be in ground transportation. There is more disruption and transformation going on now in the industry than there's been in the last century. And you can see that in the increasing size of the exits for the industry, both acquisitions and IPOs obviously the last year has been dominated by the the SPAC market and all of the various SPAC exits that are, that have taken place and are continuing to take place. And so overall you can see a clear trend line towards larger and larger exits more activity taking place in the space also more venture capital focus on the industry than there was 10 years ago which is when we first started thinking about forming the fund. Overall in the VC market and in the mobility tech VC market in particular, you see a rising trend towards towards doing deals towards investments, both in terms of deal value and the number of deals.
The COVID pandemic has obviously had a big impact on all of us and certainly on our market as well. And so what we've seen is that COVID has created kings and pauper in some cases. And so it has helped enable structural winners to emerge from the industry. For example, ecommerce has grown substantially as a result of COVID. And so companies that have been focused on ecommerce or have benefited from ecommerce like for example various kinds of logistics companies have done very well. Whereas other transportation companies that have focused on for example, mass transit with lots of people in a crowded space have not done as well. We've seen investors all around the world doing an increased number of deal making over Zoom instead of face-to-face interactions ourselves at Autotech included. And overall valuations in the startup ecosystem have remained very high despite COVID. We were expecting to see a possibly a recession that mirrored the big drop in the public markets that never really took place in the private. And then startup markets valuations have remained very high and have continued to frankly decline over the last year or so. Meanwhile, we've seen a lot of interest in mobility and transportation-related SPAC IPOs. We are continuing to see those taking place it's cooled off a little bit in 2021 but there's still a lot of interest and a lot of focus on, on SPACs overall. So on that note I'd like to introduce the folks around the table and maybe everybody could give a brief background on themselves and the companies that they represent. Let's start with Ezra.
Ezra Goldman - Sure, hi, can you see me? So I don't have any slides, but I'm the CEO and Co-founder of Upshift, we're a car membership service in San Francisco that makes it easy to get access to a car when you need one without having to own one. So we're focused on folks who live in cities who only need a car occasionally and they can subscribe to our service whenever they need a car, they can request one via our app and we'll deliver it to their door. So as a Toyota Prius hybrid or RAV4 Hybrid SUV. They can drive it wherever they would like to go and then when they're done, they can just drop it off anywhere in our San Francisco service on and we will gas it and clean it for the next time. So we take care of everything like long-term storage, insurance, maintenance, repairs et cetera. And just excited to talk about the future of mobility here and how we're building towards a autonomous connected, shared and electrified future fleet.
Daniel - Thank you, Ezra. And Bibhrajit maybe you can introduce yourself.
Bibhrajit Halder - Yes, my name is Bibhrajit Halder founder and CEO of SafeAI. So at SafeAI we are actually working on the retrofitting the existing vehicle for mining and construction industry. So think about the big truck, the dozer, the loader kind of make them fully autonomous taking the operator out of the harm so we do that in a safe office environment. And yeah, I think this is really exciting. One of the things we are looking the autonomy is a big trend is not only in the passenger but also broader ecosystem. So really looking forward to discussing that.
Daniel - Very good and Bibhrajit is the CEO of SafeAI which is one of our portfolio companies here at autotech. Austin could you please introduce yourself?
Austin Burch - Sure, thanks Dan. Hi everyone, my name is Austin Burch. I'm the Co-founder and CEO of Above Robotics which is a SAS platform for connected autonomous solutions. So Above is based in San Francisco. We really bring people, vehicles and smart devices together on one intelligent platform for automated operations. My background is technical and robotics and distributed autonomous systems. And I've done a lot of research on infrastructure enabled autonomy. And so that kind of transitioned into the remote autonomy and now edge and distributed solutions that we're building at Above.
Daniel - Very good, very good. Well, thank you. Thank you, panelists. You know, as a VC, I get asked frequently about the future of autonomous driving. And it's just one of the trends that we track as as I mentioned in the slide is there are, you know, five different things that we track connectivity autonomous sharing economy, electrification and digital enterprise. But autonomy is the one that on the surface seems to be the most seems to have the highest potential to impact everyday lives and societies. And frankly, it's a very sexy topic. Like everybody loves the idea of no longer having to drive having to worry about a stop and go traffic. They just want to wait in their house until the robo autonomous taxi shows up outside their door and whisks them off to wherever they're going. And then they don't need to worry about it. Note that I'm conflating in that example, two concepts autonomous driving, and also fleets. But in general the prospect of autonomous driving is very appealing. And so so let's start off our our group conversation talking some more about autonomous. And maybe Ezra, we can start with you. Ezra is there one specific day in the future when suddenly everyone will be riding in autonomous cars is going to be like, you know October 17th, 2024, or if not if it's not October 17th, you know, how will how will autonomous cars get rolled out in practice?
Ezra - Well, obviously there's lots of opinions on this but I like your initial framing here. I mean, I personally don't believe that there's one date on the calendar. I think that even asking that question is actually problematic because as you said a lot of people tend to think of it as sort of robo Uber's will show up one day. And that's one of many, many, many things that will happen. And so it's really a question of what can be done, on what timeline with what kinds of vehicles, with what kinds of services and what kinds of environments in each one of those is actually a completely different problem to solve. So typically I see it like you have to either constrain the environment and go all the way with some particular niche solutions, say mining for example, where you can basically be underground and not have any traffic to contend with and be very autonomous with your solution. And if you want to do something that's kind of in a in an urban environment, for example with passenger vehicles, then you have a lot of regulations, a lot of restrictions, and you either have to be highly regulated or you have to limit the technology that you're allowed to roll out. So, for example, with Upshift, we're just rolling out level two technology that comes with the cars right now. So adaptive cruise control, pre-collision avoidance that's already on all the cars today.
Daniel - Sorry Ezra. Ezra you said level two. What does level two mean?
Ezra - Yeah, well, so this is essentially you know, level one, level two. These are all different levels of autonomy on the vehicles today, right? So the tech that you see in cars today is typically, you know, L one and I think we're starting to see some L two technologies as well, where it's, the cars are starting to drive themselves you know, limited fashion on, you know so like a Tesla, for example, you know can keep itself in its own lane. It can keep distance with the car in front of it and cruise control in some circumstances, it might be able to pass other cars. This is the kind of tech that we're starting to see on on sort of manufactured cars today. And as that becomes more and more readily available a service like Upshift can put that on its platform give people access to that. But then there are also other elements of our service where for example, parking, like we could stack a bunch of cars into a parking garage with autonomous tech. You know, if we put sensors into a garage and built that a mobility hub, you could have a more fully autonomous highway driving experience for the member. You could even deliver a car through teleoperations at four o'clock in the morning at low speeds on low volume roads. And then maybe you have the member drive the car themselves the last mile essentially from home to highway. And so they just drive themselves that one mile that's really challenging to solve for which is the, the city driving, where there's pedestrians and kids and all kinds of different situations. And leave back to the human driver for, you know, for a while until that technology is really viable. So there are different edge cases you can try to solve for as you constrain your technology solution for those different kinds of environments. So that's kind of how we see the rollout is more, how do we you know, yes, we all would love to see the robo Uber's but what's kind of between here and there and what are the different aspects and elements that we can roll out over time as that technology becomes viable.
Daniel - And Ezra you said, level one and two, you know, would you call the Ls, like how many Ls are there and what are the differences? And just at a basic level.
Ezra - I'm probably not the best one to explain all this different levels but there's basically a, I believe there's four or five different levels depending on the different systems that I've seen ranging from very, very lightweight, you know, essentially assisted driving technologies that you see on cars today, all the way up to sort of what you would typically think of as a fully autonomous vehicle with the level five.
Austin - If I could chime in, I think that's a good distinction. And I think something else to say is that the gap between level two and level three is pretty large. I think level three is when you start relying on automated systems to do the monitoring and management. And so, you know, right now all of these level two cars have some level of the driver needs to pay attention or you'll be alerted or they'll shut off the autonomous driving. So I think really, I think there's two pieces. I think it's will the technology get there? And the public adoption. And I actually think level two is really good at exposing the public to what the car is doing and how it might react in situations. And so I think that that's really laying the ground floor for the public to move into these higher levels of autonomy when we're ready. 'Cause I mean, technology will get there at some point there's a lot of ways to solve this problem. And there's a lot of people putting, you know mind share into that. And so it's just, when is it ready for the public? And when can it be integrated with, you know, everything that's already ongoing?
Daniel - So if there are five levels of autonomous, which I think is the standard that the U.S Government has come up with, where does Tesla fit into that? Because you know, they're talking about self-driving autonomous, you know, full self-driving capability. Does that mean that the Tesla's is level five?
Austin - Nope, level two.
Daniel - Level two, yeah.
Austin - They're even as I think starting to make the decision to watch the driver a little more which was initially a privacy discussion, but you know when safety is involved, I think they kind of went with it.
Daniel - Yeah. Yeah, I think it's really, it's really interesting how how Tesla, you know, is a level two and yet the way they talk consumers often assume that they're level five, that they're, you know that they offer the maximum amount of self-driving capability available.
Ezra - Marketing is a lot easier than technology is at itself.
Daniel - It's true, it's true. I tell my kids, it's a lot easier to draw a picture of a rocket than to actually build a functional rocket. So thank you for that overview. I'm curious what other vehicles might use autonomous technology besides cars, because there are other vehicles out there in the world besides cars, maybe regular consumers, don't run into that many of them, but Bibhrajit maybe you can answer them. What other vehicles might use autonomous technology besides cars.
Bibhrajit - Absolutely. And maybe kind of give you a little macro view right? When you think about it, is that, you know is autonomy going to happen? I think the broader question is absolutely yes. So if you really zoom the heck out of it, you know, we created the wheel, then we put the engine, then we start doing and eventually it doesn't make sense for human to drive anything. It's not a human job to drive. So I think that's a macro we, you know, is whether 10, 20, 30 you know, 10, 20, 30 years who cares if they can take the time. So any technology, you pick any technology the way it introduces that it introduces in a more constrained use cases. And even internet it was only used by the defense at the beginning. It was not available to everybody. You talk any particular so far autonomy I think one of the thing will happen already happen is that anything constraint as Ezra's mentioning, right? Where there's a construction industry, mining industry indoor robotics and whatnot already been autonomous. I think what we are focusing on mining and construction because as you can imagine this big truck, big dozer, loader working in a new more mine or construction area, same thing repeating over and over safety is a big deal, your productivity is a big deal. So those are area you'll see autonomy is going to be already ramp up is going to accelerate significantly. Again, it's come from the you know, constraint use cases. While you do that, we are going to actually learn a lot how we are going to make it actually scale up. I think that's the, there's the two different that technology is one thing. How do you really scale up? So, yeah, to your answer mining construction I think the number one use case is defense which we are not going to touch here but it's really, you know, defense is number one.
Then you're talking about any of those constraints use case mining, construction, agriculture, indoor, all that. Going back to maybe just to touch up on the levels I think, you know, it is more of it, you know, it needs a not cool not it's confused the heck out of population on that different levels. There's really is two level. Anyway if you look at the use cases that do I need to pay attention when I'm driving or can I get the heck out of that gap right? And from user point of view they don't care either way. Either you give me the vehicle that I can sleep or you give me a vehicle I can't sleep. So I think that if you think about what industry gets or mining, construction, agriculture, they can sleep. They don't even have to be in the cab and as already there we are ramping that up. In the other word is that no, you cannot sleep. That's really, the distinction rest is basically us creating for our own technology geekiness right? The user doesn't care. But I hope that answered your question.
Daniel - I like that distinction. I like the two levels and it certainly makes a lot of sense that, you know, in constrained environments like a construction site or a mining site, there are, it's still complicated to navigate that. But there are not as many variables as there are for, for a a car that's trying to navigate between highway driving and residential driving and potholes and construction and everything else that the people run into on on the road, kids chasing balls across the street and deer and so on. So yeah, makes a lot of sense that that autonomous will come faster to these specific industries. And you mentioned defense as, as the number one industry, the terminator was obviously autonomous. So I can see how there's there's really a lot of applications there.
Bibhrajit - Just to top it on that I mean, defense I mean the whole autonomy started with defense, right? The DARPA Grand Challenge too you know, that's really, kick-started the whole thing. So yeah. I mean, obviously they saw the great technology they're like, why should we only use it? Let's give it to the, give it to everybody. And that's the journey we are on since 2002, 2003. We are on a great journey I think so. Yeah, just to add that comments.
Daniel - And how...
Ezra - I was just going to say, we actually already see autonomous tech in trains and planes for a long time.
Daniel - Yeah, autopilot.
Ezra - Yep.
Daniel - Yeah, for sure. Yeah, and so, so it seems that, you know autonomous has a lot of, a lot of autonomous technologies in vehicles need to take into account a lot of things that they are environment. And, you know in some cases they can do that on their own as isolated standalone computers in motion. But one of the trends that, that, you know we're sensitive to in the industry is connectivity. And so you can connect to a lot of things. I mean, we're all, we're all connected every day through our cell phones to do a whole massive ecosystem of infrastructure with cell phone towers and satellites and all sorts of things. What, from an autonomous application perspective what infrastructure and connectivity needs to exist in order to enable autonomous striving for either on road like, you know passenger cars driving around or off-road like construction and mining sites. Maybe Austin, you can speak to them.
Austin - Absolutely. So I think first, you know the important thing to note is that this path to autonomy is always going to include humans. And we're slowly going to kind of taper that off. We'll have a mixed environment for a long time. Fortunately, a lot of people are putting work towards this connected autonomy environment which also includes humans sharing the road. But, you know, you mentioned infrastructure, you know, imagine humans driving on highways without you know any kind of lane lines or signs, you know that's the infrastructure humans need to, you know share the roads with all the other humans safely. In the same way with these connected autonomous systems we need an infrastructure to support them. You know, and in the public at scale, you really have both digital and physical infrastructure requirements. So the physical, you can imagine replacing like these stoplights with, you know, the high precision localization or the sensor fusion needed for the autonomous systems in that environment more simply just the networking like 5G. But the more important piece is we need some systems of systems architecture that shares data of these vehicles that shares the perceptions and kind of enables this grander understanding of traffic where people are trying to go and driving as a whole so that we can really unlock kind of the next level of safety and efficiency. 'Cause people aren't going to adopt autonomous systems if there's no added benefit. You know some people might like just sleeping, but, you know if we can provide added benefit to safety and efficiency, you know, we can quantify those and then people will really be motivated to transition. So Above is really kind of working with some groups to do this. The Texas Autonomy Institute is a nonprofit trying to assemble the city, government and industry resources to deploy the hardware for this. And then like NASA, I know just released more information on their DRF project, which is Data Reasoning Fabric.
And it's this like marketplace of consolidating different data sources, so that automated systems can kind of subscribe to them. And this is, you know goes back to like your digitization question. You know, we need a higher level of localization and how are we going to get that it's not going to be from one sensor or one vehicle system it's going to be from this collaborative environment to share and understand the environment. And then Above Robotics, you know, we're focusing on this ecosystem of services microservices to support these mobility operations and applications at scale, and, you know beyond just the single vehicle. So I think it's an important problem to address, you know if we do want these autonomous systems integrated into our public areas.
Bibhrajit - So then let me just add a little bit of collect to that how, you know, kind of take that question a little bit on. And I'm going to use the construction industry as a use case. So, you know, we are deploying the autonomy and we are saying that it's going to save you 40% of your body cost. Boom, they're sold, right? That's your ROI. No question as literally the next construction industry doesn't use autonomy they are going to be out of business. Pretty straight out away right? Now, for the autonomy now we need infrastructure. We need the connectivity, right? So yesterday they were like, no I don't want to use connectivity, but tomorrow they say oh I to do deploy autonomy, I need connectivity. So you want to look at autonomy almost a trigger point for this other technology. So they are going to say I'm doing autonomy anyway, let me put the infrastructure for communication in a network. And that actually can be useful for all other teams. We all know connectivity is not just for carrying the vehicle. We need the data for all the... So it's almost like a trigger point autonomy is triggering the things right? Now you, you know big construction company never thought about connectivity you need to think about it. And they're all coming on board, right? And you want to think about that this is kind of turning around upside down. It will help the whole industry. We see that in this all industry if you will mining and construction they are really jumping on board to digitalizing adding infrastructure to make sure they are, you know ready for autonomy deployment if you will
Ezra - It'd be interesting to hear your guys who are who are more heavily in the AV space right now talk about kind of the scale of that connectivity because we have telematics in all of our fleet right now and we get, for example, GPS position off of the car, but, you know, we're pinging the car over a 2G network like once every few minutes. And maybe if it's driving more frequently than that you know, or even like once an hour or something it was just sitting there parked. And that's a completely different level than needing like, you know every single centimeter of every city mapped in real time. And you know, all the data uploaded to the cloud, like, you know, within, with zero latency it's just like the scale is just phenomenally different than what's currently. And we have third party telematics integrated into our car. So this isn't even stuff that comes with the car. And so I think that that's something that we need to bear in mind is that we're just talking about a massive infrastructure lift here that needs to happen.
Bibhrajit - And if Dan, just to add I think not only massive infrastructure is actually going to be a massive opportunity. That's two thing is coming late. You're getting huge amount of data at a huge amount of quantity. The second most important thing is that it is going to be a highly accurate. In a telematics can lie about engine quality, right? Because they detect in way of but the autonomy can not lie about his position interact with... So think about a data which is thousand times more volume or maybe even more and actually accurate. What can you do with it? Daniel figured out where to invest but there are going to be a multi-billion dollar company will come out who will handle the data. So we are really sitting on the tip you know, tip of the iceberg and we see that in our industry sitting on the tip of the iceberg on that.
Austin - Yeah, I think we're also fortunate to have new technologies that we can deploy. You mentioned, you know, kind of pinging vehicles from the cloud with some latency or, you know now we have more peer to peer tools, whether it's, you know we can host our own networks and it can be decentralized or we can have low latency high throughput communications on peer-to-peer networks. I mean, both of these one, one addresses the time dependent issues and high throughput for data and the other one addresses the security and, you know relying that these, these autonomous systems are validated from some, you know, reliable source. So I think if we bring in those new technologies that you know, yes it's a huge, massive undertaking, but we have new tools that we haven't been able to use before.
Daniel - But what I'm hearing is that, that many different pieces of infrastructure could be getting built and potentially coming together in order to facilitate autonomous driving. And so there's the 5G networks which can and have a higher bandwidth and less latency of course, 5G is not very widely deployed at this point. And I think maybe I got a 5G microchip when I got my COVID vaccine. So, you know, maybe I'm walking around with one but aside from that, there's this not very well well deployed. And the coverage across the U.S is not great, which makes it harder to enable autonomous driving across the U.S if the coverage isn't there. But then meanwhile, there are other opportunities with vehicle to vehicle as you said, peer to peer connectivity and communication. And then of course there's also the role of the cities and public infrastructure in providing beacons and landmarks and connectivity points as well. And potentially those can also be used to facilitate autonomous driving, but there's a lot of pieces to it sounds like that all need to kind of come together, not even including you know, regulatory support for for some of these initiatives and geographic regions.
Ezra - I think a lot of people tend to think like they're just kind of looking for that silver bullet. They're like, well, is it Uber, or is it Tesla? Or is it some other company who's just going to do it and you're like, well, like, there's so many different pieces and so many different industries and so many different use cases. And I mean there's literally you know, thousands of different opportunities here. And, you know, just robo Uber's is just one of one of many, many opportunities.
Daniel - And so, you know, one of the questions is what layers of a technology is stack enable autonomous driving to occur? You know, is it just a chip made by a single company? I think Austin you mentioned the term sensor fusion before, you know maybe you can, you can define that. And I'd love to understand how that fits in and Bibhrajit maybe you can also talk about the layers of the technology stack that are needed to enable autonomous.
Austin - Sure, so, you know, sensor fusion can occur locally within one vehicle system or multiple.
Daniel - Austin what is sensor fusion
Austin - It's, you know, overlaying and combining different sensor different sensors to get a better resilience and understanding of you know, what's out there.
Daniel - So, a sensor like on the edge of the car that interprets its surroundings, like a kind of like a camera or something like that?
Austin - Sure, so probably the most common are cameras, LiDAR and radars. Well, we can have algorithms where we can overlay color data with distance status so that we can have a colorized point cloud. And that's one example of fusing sensor data. And then you can match up your, you know visual classification and detection algorithms with your LiDAR point cloud. And, you know, you can you know, have some uncertainties and kind of minimize those there. To address your stack question Above looks at autonomous systems kind of at a break from local to global. And this is also an robotics topic, kind of global control and local control. Local relies on all of its internal perception and those sensors to fuse and understand where it's at and not make any catastrophic decisions. And then the global control, you know looks at the high level where it's trying to go, what task is trying to complete. And so Above kind of breaks it there and what we see is there's a lot of companies building an autonomous solution to solve a task. And, you know, these are necessary because there's a lot of difficult tasks to solve out there but we don't want to per se solve every single task. So we want to sit on top of those vehicles and enable them to operate outside themselves and not just work on one task. You know can these robots or autonomous vehicles be integrated into larger workflows that include multiple systems or humans and unmanned systems. And I think that's kind of the break that we're looking at. I'm sure Bibhrajit is looking at other ones.
Bibhrajit - Yeah, maybe just, I mean, from the stack point of view really looking at three main component, right? The sensors to you know, pull the data in, it could be LiDAR, radar, camera, GNSS system. It could be also be V2X, right? We do V2X, V2V also a sensor to the vehicle. Its just bringing the...
Daniel - Bibhrajit what is that, what is V2X?
Bibhrajit - The V2X is basically vehicle to vehicle. So you put two different type of, you know, maybe a pod if you will, into different vehicle and they communicate. So you get information about the other vehicle through their communication. So that's the first component, right? You add a sensor in the vehicle to bring that data in. Second thing, is that, okay now you've got the data. What do you do with it? You need a heavy processing unit. This is where NVIDIA's and Intel's of the world is all excited about autonomy 'cause they want to sell more chip. So really what they're doing they're bringing the whole sidebar into the vehicle. So that's your second component. That NVIDIA, GPU, Intel's and all that. Third is the algorithm Now you've got all the data, you've got a compute platform what do you do with that? And this is where, you know, whether the Tesla, whether is Waymo, whether is Apple that's where their dream power is going. How do I make that and the better the algorithm better the output is. So this is your three main layer, the hardware, the sensor the compute platform, and the algorithm itself that makes your end you know, a sensor fusion is a part of that algorithm well that makes it decision making, right? What do I do with it? That's your brain basically.
Daniel - Great, that's very helpful. And so, so is there one company that does all those layers or they're multiple companies or does it vary?
Bibhrajit - Yeah, I mean, as, I mean, both I think Ezra and Austin has mentioned, right? There is no one company maybe I usually say beside Apple and Google you got to give them exception. There is no one company in the world has all the expertise under one roof. 'Cause you are talking about the hardware you're talking about the algorithm you're talking about the chip. Now you're going outside the vehicle as Austin is talking or is not just a vehicle now you're talking of the cloud. So yes, other than Apple and Google nobody has all the expertise in one roof. So you'll see a lot of partnership a lot of collaboration, everybody will bring together. And that's really how the industry will move. It's definitely not just one company.
Ezra - Well, even Apple and Google don't know how to build cars, at least not yet. I mean, you know, Talk about you know, how to building a car. Yeah, that's a whole I mean, crash testing. And there's a whole bunch of stuff there that they can't do either at the moment.
Daniel - Well, thank you. That's very helpful overview. So how advanced is passenger car autonomy today? We touched on it a little bit earlier, but you know we've talked about five levels and Bibhrajit said well, "Either you can sleep or you can't sleep." You know,
Ezra - You can't sleep.
Daniel - You can't sleep you know, even though every now and then you see like I've been on a highway 280 and I've passed Tesla's where there's where there's someone sleeping. But I think you can also get arrested for that. So...
Bibhrajit - Let me take that as a soundbite and then I think I'll have them go to the detail, right? I think it is, you know, there was you know, back in the 2014, 2015, 2016 and back then was a lot of hype and people thought always about to come now we've kind of gone down, right? People are say oh my God, it will take another five years I think the soundbite is that its actually lot ahead than today people think. I mean, you know most of people have, 'cause it's been such a hype and gone down. I mean, why Google and Apple are not going to use this as an example. They are actually further ahead than people think they are. And if you look at their kitchen side, you'll see that... So I think technology wise we are further ahead than people are thinking now. And it will come. I think Google I know for sure is way ahead than the perception we are having that it will take another five or 10 years. Or I'll have the you know, Ezra and Austin maybe we can get into the detail.
Ezra - Yeah, and I think the passenger cars, if you look at that market, it's still very much early days. I mean, our cars, you know even our newest Toyota vehicles can do a lot of the same stuff that, that a Tesla can do. You know, they can stay at the they can steer the car and it's laying on the highway. They can speed up and slow down relative to the traffic ahead of them. They can apply the brakes, if someone in front of them applies the brakes You know, there are some other think they might hit there's blind spot monitoring detection real traffic collision avoidance, you know, there's actually quite a bit that they can do today but you know, you still have to have your eyes on the road and hands on the wheel. And what I like to tell our members is to think of them like a student driver, right? Like if you had a 15 year old kid in the... You know, your son or daughter, like driving the car you might still want to be paying attention and, you know be ready to stick your hand out on the wheel or hit the brakes or something if you had to, right. Because you just don't know exactly what they're going to do. And yes, 80% of the time it's probably going to do what you would expect. And it's probably going to, you know, take a lot of the burden just the mental burden off of steering the car or driving the car. But getting the problem is that getting that last even 1% from 99% to a 100% that's typically where, it's kind of an exponential curve, right? And that's where it's really the hardest is all of these unique edge cases, right? So if I'm driving one of our cars it does great in rush hour traffic. But if somebody cuts you off, it's kind of slow to respond. If there's a motorcycle, it's a little slow to respond. If there's some weird issue with the lighting while you're driving across the bay bridge its can get a little out of whack and lose the lane marking. Yeah, there's all these kinds of edge cases where it's really hard to solve for. And I think that's where an automaker, for example say a Toyota now, Elon Musk, Tesla I think he's a little willing to take more risks than most automakers, but your typical FORD, GM, Toyota et cetera, like they're very risk averse.
And they're putting millions of cars out there every year and they don't want to just, you know, they're looking like moving fast and breaking things is not a good strategy when you're talking about going 70 miles an hour down the highway. I don't think anybody wants to break things certainly not their bones. So there's a really high risk. And I know some of the companies, Google, Uber, et cetera, have kind of had a bit of a mentality of like, well if no one's dying, you're not running fast enough. And it's, I think we're starting to see a little bit of a pushback on that and say, well, maybe that's not exactly the right metric that we need to be going on. So, but that's kind of the state of play. I mean, it's very slow it's very kind of, you know, whatever you can legally get out on the road, which tends to be nowhere near what you can put into say a mine or a university campus or an airport, or somewhere like that.
Daniel - Yeah, one of the interesting models for a potential approach to this industry and this technology evolution is the Federal Aviation Administration, the FAA where they are extremely conservative and they don't want, you know, 99% of flights to make it to their destination. And only 1% of crashes, they really want a 100% of the flights to land safely. And so, meanwhile, you've got you know, conservative governmental bodies trying to find the right posture and position with regard to this self-driving technology because it's also possible that, you know, the technology can reduce accidents by 80% compared to human drivers. And yet there are still going to be a bunch of accidents because 80% is not a 100%. And then there's the risk of public fallout and perception and fear of the technology because there were any accidents at all. So in some ways the public demands a 100% safety which is, which is an aggressive target for...
Ezra - Well, if you think about airplanes and trains right now it's like, they are way safer than driving a car but every time there's a plane crash or a train wreck it's like front page news for, you know, a week on end. And everybody wants to know what happened. And it's like, hey, you don't twice as many people died on the road in the last hour just driving down the road to get to work. Like nobody is talking about that.
Daniel - Yeah, yeah definitely.
Austin - Yeah, it's interesting like we really are already at that point where these vehicles do a great job. I don't know the statistics on number of miles, but you know if they were a human they'd have amazing interests. But it is kind of interesting, you know there's companies doing it differently. You know, you have the Waymo approach, Tesla approach there's even open-source approach. And they're all solving the same things in different ways. And they're all kind of reaching this ceiling at level two and they can't get the adoption whether it be from the public or from the government. And so, you know, what do you do? And I think that's when we bring in the digitization we bring in the infrastructure and that really lays that foundation that we can kind of transition collectively as a whole.
Daniel - Yeah. And actually one of our questions in the Q&A section is how long will manual overrides and safety systems need to remain in place before the public fully accepts autonomous vehicles as part of daily life.
Bibhrajit - Let me start out with a high level again otherwise we'll go down the detail. So there are two side of that question, right? One is a safety standard another is some, you know when do we take that up? Let me start with the story and then I'll answer why I said, you know when we had the elevator and this is one of my favorite stories, right? When you have the elevator it took humanity 40 years to understand that elevators are safe. And the reason we took the elevator man out of elevator not because technology was not ready because they had a strike and during the strike, no elevator man worked. And then they're like, oh, assuredly they are not needed. So as a, again, broader view humanity collectively is always anywhere from 10, 20, 30 years behind technology. So, you know, it's not like, you know human will accept the safety and then technology will come. That never happen. That's not the way technology works, right? We are not taking, you know we are not taking this manual override out not because of perception because technology is not ready. And as I mentioned, if you take Waymo with the example Waymo is not ready yet. I mean, using webinars example, right? The moment Waymo think we are ready they will take it off. It's nothing to do with perception. Humanity will catch up over the next 20, 30 years. Usually that gap is reducing. If you think about elevator like took 40 years. But Uber when people say, oh, I don't want to get into strange it took them seven years to get over that they are whatever they had. So human is doing better of reducing the gap of catching up with the technology. So to your question, when we will take over manual override? As soon as Waymo is ready, again when you use example technology is ready. It is not a regulation problem. Is it, you know, people are going to accept, they're going to catch up. It's, you know, it will take next 10, 15 years. There will be always a portion of population they will never catch up. That means there are people in this world they don't want to take Uber because it's not safe. So there will be a, you know, remain that there but eventually they will catch up if that broad sense makes sense.
Daniel - Yeah. As a venture capitalist I find it very interesting to reflect on the evolution of trust in systems. And in the 1990s, people were afraid to buy a book on Amazon because hackers might steal your credit card. And then later people were afraid to stay in the home of a stranger through couchsurfing or Airbnb, because, you know there might be problems and online dating as well. You know, you might you might meet a serial killer and so on, right. So there is a, in general
Austin - Digital currencies.
Daniel - And digital currencies, yes. That's a whole topic in and of itself. So yeah I agree it's interesting to reflect on this. We've been very focused on autonomous driving in this, in this session I just want to highlight, you know for a moment that there are very exciting trends besides autonomous driving, you know, autonomous is very sexy. I'm personally excited about things that are also not sexy like buses. I think that any any company that can increase the utilization of under utilized buses can actually do very well. And so I'm an investor in three bus companies and one of my companies, you know, bus app helps commuters get to get to work every day. And kind of like the Google Shuttle here in and make sure you're in the Bay Area. And so, you know, it doesn't, there's, there are still huge businesses to be made, you know Uber and Lyft, for example, you know do not have they have sophisticated technology but not at the level of a Waymo, you know, and yet these, you know, Uber and Lyft and Airbnb are some of the largest IPOs in the last century. And so so there are huge businesses to be made through autonomous technology, but also in other areas as well.
Ezra - Yeah we actually filtered our service to be shared, connected, electrified and autonomous, but in that order. And so, yes we want to do all of those things, but we started out focusing on just fractional car leasing, what is the new business model for a fractional lease? And then we said, okay, we need to connect it so that we don't need to meet people in person so we can get more data off the car so we can unlock and drive the cars without needing a key. We've just integrated all of our fleet with telematics this spring. And, you know, the next level is like, okay now electrification for the next couple of years is a huge push for the automotive industry. We see tons of new electric cars, so more infrastructure more subsidies prices are coming down. The range is coming up, consumer adoption is there. So, hey, let's put some electric cars onto the platform and see how that performs. So you can get a hybrid or an electric car, depending on where it is you need to go and what you need to do. And then as we start to see more, you know, and right now we're just doing sort of standard issue, AV technology you know, pre-collision avoidance, adaptive cruise all the kind of stuff that comes with the car. And that's, as I think Austin mentioned earlier that's sort of getting people, an initial flavor for what this might be like to, to kind of have the car start to drive itself. And then as we start to see teleoperations and other kinds of AV tech come online, that we can integrate similar to the telematics, we can start to integrate that into various aspects of the fleet. So, you know, we see that as sort of a piecemeal approach. And I totally agree There's a lot of different opportunities here not just AV
Daniel - Cool. Well, we've got about five minutes left. Let's let's open it up to any questions from the attendees. Andrew, do you have any questions that you're tracking?
Andrew - Yeah, we have one that came in earlier here. Let's see here. What specific tactical advice do you have for startup founders that are in early stages in the sector? Given current, given the current climate?
Daniel - It's a very broad question.
Ezra - Well, what are you interested in? What do you know about? There's so many opportunities.
Bibhrajit - From the business point I can chime in one go have that ROI. If you don't find an ROI, you will be doing research then it's better to do a PhD date see you gotta go after that ROI and say am I actually giving back? Is there dollar add up? If your dollar doesn't add up you don't have a business. You have a PhD. So I think that's in a summary without knowing what you're good at. But if you're talking about startups, you got to be a ROI. And that's why you see like most of the startups in autonomy you know, either falling or dropping off because they haven't find that ROI yet. So yeah. Let me stop there.
Daniel - ROI is for business to business startups. It's less of a factor for consumer startups.
Andrew - Are there, as Dan mentioned earlier, we spend a lot of time on autonomous vehicles and things of this so. Are there any other non-sexy sort of sectors in the mobility world that you're excited about or that you see trending over the next few years?
Daniel - We have several logistics investments. And so, and started just to wrap up one of the comment to the previous question regarding tactical advice. It's kind of a generality, maybe even a cliche, but I would say, make something that people want, which sounds silly and obvious, but is actually not that silly and obvious and ties into ROI because if there's an ROI then people probably want it. But, you know, make sure you're making people something that people want. Later on, you can figure out how to make it profitably, but start out with with making sure that the people actually want it. Just a general piece of advice. And then, and then Andrew, regarding the question of other sectors, you know, I think, you know, logistics is not sexy but it's a trillion dollar industry moving goods from, from point A to point B there are lots of, lots of un-sexy industries and transportation, sub industries, like, you know I'd love to see a startup tackle, automotive recalls huge pain point for OEMs and dealers and repair shops as one example. But figuring out new battery technology or extending the range of electric vehicles is more likely to get people's at least the attention of the press.
Ezra - Well I would say that as a fleet operator, you know, I went into this because I had the pain point it's a pain in the butt to have a car in the city and I don't want to do any of the existing solutions that are out there and there should be a better way to do it that doesn't have all this friction and headache. And, you know, as I've been basically taking on all the burden of buying cars, selling cars integrating telematics, figuring out insurance, I mean it's literally like every single thing I tried to touch with the business is broken. Like, when you say, well, what's... It's like anything, just pick one of these problems and fix it because, you know, buying a car, selling a car, the telematics piece, the insurance piece. I mean, it's all completely broken. And like, there's a lot of innovative stuff going on but there's also just like the sky is the limit on how much better it could be on any one of those pieces.
Austin - I think the digitization around mobility is really interesting. We didn't talk too much about it, but there's like a lot of automation to be done for end to end supply chain. And if we can piecemeal autonomous systems into those operations and kind of surround it with, you know digital assets and digital processing, I think there's a lot of advantages to be heard there. And that's called all kind of under the hood and definitely not sexy for most people.
Daniel - Andrew, any other questions?
Andrew - No, that, that pretty much wraps it up on the Q&A frame. I just really would like to thank everybody for attending and be respectful of everyone's time here. We're here at two, o'clock on the dot a very special thank you to our panel for this insightful discussion. We really appreciate your time and expertise. We'll follow up with everyone. And with an email that includes details about the replay of today's webinar. We really hope you enjoyed the webinar on the revolutions in mobility industry. Here at First Republic Bank, we're proud to serve you as part of the innovation community across the U.S. Please feel free to reach out to us for any of your financial needs. Thanks again for spending time with us. And we look forward to seeing you at our next event. Thanks again every one.
Bibhrajit - Thanks everyone.
Daniel - Thank you.
Austin - Thank you Andrew.
The guest speaker(s) is neither an employee nor affiliated with First Republic. Opinions expressed by the guest speaker(s) are solely their own and do not necessarily reflect those of First Republic. This information is governed by our Terms and Conditions of Use.