Autonomous driving is something that is often disregarded as science fiction and for the distant future. What you may learn in this article is that this isn’t quite true and self-driving vehicles may be coming to streets near you quite soon. With the emergence of autonomous driving will come a dramatic shift in how the ride-hailing market, on which companies like Uber (UBER) and Lyft (LYFT) have been working so hard to capture as much as possible, will change dramatically. Companies that don’t currently compete in the sector, like General Motors (GM), Tesla (TSLA), and Alphabet (NASDAQ:GOOGL) could become some of the biggest players in the market. One of my previous articles details an important component of the road to autonomy, the usage of LiDAR, and, if this is something you deem to be important, I would recommend reading it before this article. Through this article, I will discuss the various new players in the changing market and how their differing approaches and positions within the market will lead to the ultimate success or failure of their endeavors.
Now, it’s time to meet all of the players. I’ve already mentioned Uber, Lyft, General Motors, and Tesla, but Alphabet’s Waymo, founded in 2009, is also putting in a lot of work to make autonomous driving a reality. In fact, the company received approval back in July 2019 to operate its small fleet of autonomous vehicles in the Californian South Bay area. However, this approval still requires an employee to sit in the front of the vehicle ready to take over if something goes awry and is exclusive to Waymo’s employees. Yet, curiously, Waymo had also received permission to test its vehicles without a test driver in Santa Clara County. Shortly before this, Waymo partnered with Lyft in order to provide customers with access to their self-driving vans just outside of Phoenix, Arizona. However, Waymo now also operates independently in the greater Metro Phoenix area, providing rides to a select group of clientele. Quite recently, Waymo actually committed to 100% driverless rides in Phoenix to all members of the public, eliminating the backup driver from all offered rides and fully utilizing their permit. In the Metro Phoenix area, as well as Santa Clara County, Waymo continues to be the only autonomous vehicle company that is permitted to offer rides without a backup driver present.
Waymo’s autonomous vehicle uses a diverse sensor suite, comprising five LiDARs, five cameras, and five radars. Each of these components is meant to create the clearest possible image of the vehicle’s surroundings, from depth to what direction a pedestrian is walking. The brain behind turning all of this data into driving control is an Intel chip (NASDAQ:INTC), designed specifically for autonomous driving. The computer chip is actually incredibly important, as it must be powerful enough to process up to five gigabytes of data per second with calculations taking mere fractions of a second.
However, there is one more thing that Waymo is doing on the software side in order to operate their vehicles. The company creates an incredibly detailed, three-dimensional, map of every area that its cars will drive in before they hit the roads. This, paired with highly accurate GPS, allows for a greater source of driving precision, as the maps act as more of a guide than the sensor suite. This pairing makes the sensors play more of a support role, as they account for any changes in the environment, such as other cars, pedestrians, or traffic lights, that aren’t on the maps. Unfortunately, this also geofences the locations that Waymo can operate its fleet based on the creation of such detailed maps. Waymo is also utilizing deep learning software in order to try and train their autonomous software, simulating as if it were a car driving on real roads. In their simulations, Waymo has “driven” over 15 billion miles total with 20 million simulated miles every day. This high level of data training is incredibly important to develop the effectiveness of Waymo’s software in order to make sure that the vehicles can make the most out of all the data they collect.
Tesla, which first unveiled its semi-autonomous driving system in 2015, has a pretty different approach than Waymo. Beyond the aversion to LiDAR, which I believe to be quite wise, the company does not use any precise mapping. Instead, the car relies solely on the information taken from its hardware and processed by its software. Tesla’s sensor suite uses eight cameras, twelve ultrasonic sensors, and one radar. The company’s proprietary computer chip is also a large part of the working of the company’s autonomous driving and boasts some pretty impressive specs as an industry leader, making 36 trillion operations per second with HD footage. To apply all of this hardware, Tesla began offering Autopilot. Autopilot began as a mere driver-assist system but started to grow to semi-autonomous highway driving and steering assist on “tighter, more complex roads.” Tesla began on highways in order to train their system with more simple driving. Compared to city driving, which adds a multitude of complexity in terms of traffic laws (stop signs, traffic lights, crosswalks, bikers, etc.) and road density, highway driving was the easiest place to start.
By introducing the system at a greater level of simplicity, the company was able to iterate on its basic driving capabilities without worrying about other complications. On the city streets, though, Tesla has been training its neural network to recognize all of the necessary markings in something they call “shadow mode”. Shadow mode allows Tesla to train its fleet as if Autopilot was actually active, collecting real-world data without jeopardizing road-dwellers with a juvenile driving system. Tesla’s use of its large fleet as the main training source of its neural network is an advantage that other companies just don’t have access to. This is perhaps Tesla’s greatest advantage. While the company does also utilize its own simulations to train its software, the advantage of a tremendous amount of real-world data is unrivaled and hugely important, even at Waymo CEO, John Krafcik’s, own admission.
Pictured above is how Tesla’s Autopilot processes its environment on a typical road. Footage courtesy of Tesla
General Motors is pursuing autonomous driving through its subsidiary, Cruise Automation, a company founded in 2013 by Kyle Vogt and Dan Kan. The company uses five LiDAR sensors, 16 cameras, and 21 various types of radar sensors. This is far more than any of the other companies as Cruise seems to be a proponent of the idea that the more data available to the computer, the easier the software’s job is. While this may seem intuitive, more data can overtax the computer chip and raises costs with expensive sensors that some may seem as unnecessary. Cruise, just like Waymo, uses Intel as its chip supplier to act as the brain for all of the information that is gathered by the various sensors.
Cruise also operates on the California streets and is racking up its test miles. To clarify, Cruise has been testing in California for some time, but this permit, granted early in 2020, now allows them to carry passengers, as long as a backup driver is present. A more recent permit allows for incredibly limited testing for autonomous rides without a backup driver present. The limitations include permission for just five test vehicles, no driving during heavy fog or rain, and speeds not exceeding 30 miles per hour, all within San Francisco. Additionally, Cruise will still be limiting their passengers, driverless or with a driver, and be unable to charge them. Cruise also launched a very limited program called ‘Cruise Anywhere’ back in 2017 that offered free rides to employees of the company in San Francisco.
But inside the simulation, Cruise is driving a one-hundredth of the number of miles that Waymo’s doing daily. At 200,000 miles a day, Cruise is doing good work, but there is no denying the advantage that being a sub-company of Google can do for your machine learning capabilities. While being owned by General Motors doesn’t provide Cruise the same advantage that Tesla has with its fleet (GM cars aren’t equipped with fully autonomous driving hardware as Tesla vehicles are), it does provide an advantage for designing their sensor suite. The company is provided with a clear blueprint to follow and can develop the sensors to fit with the vehicles manufactured by General Motors.
Now that I’ve got their competition out of the way, it’s time to move on to the main subjects of this article, Uber and Lyft. Beginning with Lyft, the company began its autonomous endeavors in 2017 and is taking a bit of a mixed approach to their autonomous development. Lyft has partnered with Aptiv (APTV) to develop their autonomous driving capabilities in Las Vegas, while also developing their own technology independently in Palo Alto, California. Lyft’s partnership with Aptiv, however, seems to be exclusively in providing the ride-hailing software while Aptiv develops its autonomous capabilities independently and doesn’t look like it will expand outside of the Las Vegas pilot program. So, where do Lyft’s own autonomous capabilities stand?
Lyft, as I mentioned earlier, began developing its own autonomous driving capabilities three years ago, in 2017, a good amount later than most of its competitors. However, in that time, the company has made large strides in advancing its technology and has piloted a rapid rate of development. 2019 was when Lyft’s first autonomous vehicles hit the California roads, and, while the company still lags behind companies such as Waymo and Cruise by a significant margin, in terms of total autonomous miles driven, the progress is promising to watch. To observe its surroundings, Lyft vehicles are equipped with six cameras and three LiDAR sensors. In its simulation, Lyft is using data gathered from its ride-share service, leveraging its massive fleet in a manner similar to Tesla. Lyft is also creating highly detailed 3D maps, again by leveraging data from its massive ride-hailing fleet. These 3D maps are incredibly helpful sources of base navigation for autonomous vehicles and Lyft’s decision to begin to incorporate them will likely help expedite the development of their vehicles. Even so, the company is still quite far behind its competitors and will be stuck playing catchup for the foreseeable future as a consequence of their late entry.
Uber had the most controversial involvement in autonomous development thus far. A fatal collision in 2018 caused Uber to halt all testing in Arizona, the area of the fatality, where they had been testing since 2016. Uber also decided not to renew its California testing permit after its expiration in 2018. Neither area allowed passengers to be present in the vehicles. Uber announced in 2015 that it would be pursuing autonomous driving, just a year before beginning tests in Arizona. This obviously caused massive delays in their program as Uber’s work in the field was incredibly limited. But their hiatus only lasted for nine months. Resuming tests in Pittsburg, Uber was back on the streets, working on their autonomous vehicles. However, even though the company had returned to the roads, testing was still severely limited. Operating only during the day in optimal weather conditions, without exceeding 25 miles per hour, the cars were working with one hand tied behind their backs. It wasn’t until more recently that Uber has been able to really resume their testing.
To date, Uber is still only operating autonomous vehicles in San Francisco and Pittsburg, with test drivers behind the wheel ready to take over if need be. Uber’s return to San Francisco is still rather fresh, beginning earlier this year, and, much like in Pittsburg, is still quite limited. Daylight tests in ideal weather conditions continue to restrict Uber’s accumulation of meaningful data as their system is not pushed in the slightest. This makes the simulation the only place for Uber to train, and analyze, its software in any form of challenging conditions. Unfortunately, Uber’s behind the curve here too, as simulation testing only began in mid-2018, following their fatal accident. Since it was made public that they began operating simulations, Uber has been very quiet about what exactly goes on within the simulations, which leads me to believe that they have been unable to produce anything remarkable and their late entry into the technology has hurt them quite a bit.
Just because a company believes that they have the capability to operate an autonomous driving vehicle, doesn’t mean that they will be permitted to do so. That’s what we’re seeing with Tesla right now. The company’s tumultuous CEO, Elon Musk, has often been a bit overzealous when predicting his company’s ability to bring autonomous driving to their vehicles – and that’s putting it kindly. Touting their semi-autonomous driving system, Autopilot, as full self-driving has been criticized by many as dangerous for its users. And while Autopilot remains a driving-assist feature, it is still a very capable one. As Tesla begins to rack up more miles on its semi-autonomous driving system, without driver input, the company will be able to demonstrate the effectiveness of their system. Having done this for years, Tesla hopes to be able to demonstrate the safety improvements of autonomy over human drivers. Tesla has consistently demonstrated that their system makes a car around 50% less likely to be involved in an accident than a vehicle with regular active safety features. Posting similar numbers to this for city and regular street driving would make approval seem like a no-brainer.
However, the NHTSA would likely want to see at least a year’s worth of data before even considering the company’s application. So, while Tesla should have an easier time proving the capabilities of its highway driving system, its application for street driving will likely face greater scrutiny. I also expect an extensive testing process to be done of Tesla’s Autopilot system by the NHTSA to give a final seal of approval. Because of this, I expect Tesla’s complete permitting process to take at least one and a half years after becoming ‘feature complete’, though it will likely take closer to around two more years. While Musk seems pretty convinced that his company’s cars will be feature complete by the end of the year, I wouldn’t expect to see truly ‘feature complete’ until the end of 2021. Though the company just released an early beta of their ‘feature complete’ full-self-driving software, the roll-out will be incredibly slow and issues are likely to arise, as with most new Autopilot capabilities. As such, this development seems to be in line with my expectation of a feature complete release for the end of 2021. This would put Tesla’s regulatory approval at some point in 2023 or, perhaps more likely, in 2024.
Something to note is that when, or if, Tesla gains its regulatory approval, it will be nationwide. This is due to the approach that Tesla has taken with its autonomous software and sensor suite, as the company wants their cars to operate the same on any road, regardless of whether or not they’ve seen it before. This approach to autonomy is clearly influenced by the mindset of a car company, working to bring this service to all customers at the same time and provide the greatest amount of utility.
Waymo’s got a bit of a different regulatory approach to take. Without a Tesla-sized fleet to work with, Waymo is at a bit more of a disadvantage when trying to prove its capabilities to regulators. However, with more than 20 million real-world miles and impressive safety statistics, Waymo’s making a strong case for their real-world success. While Waymo’s real-world miles are limited in breadth, based on where the company tests, this isn’t necessarily a bad thing as it fits their approach for solving autonomy. The denser testing areas will also create more real-world miles in areas that they’re looking to operate in and are, therefore, perhaps equally as significant as Tesla’s nationwide figures.
Instead of looking to gain national approval all at once, Waymo will gain approval town by town, city by city, as their reliance on mapping every new area and testing their vehicles accordingly will require this. While likely expediting the review process for each area the company looks to enter, this will likely be the most time-consuming aspect of Waymo’s mass deployment. Essentially, Waymo opted for an easier, likely safer as well, but much more laborious process that seems more suitable for a ride-hailing company, instead of a car company. As a result, regulatory approval will be quite staggered, but it is already the first to market as a result. Perhaps, once Waymo has gained supreme confidence in its sensor suite, it will no longer rely on this more cautious approach and will apply for mass approval as well.
The aforementioned Phoenix operations are an incredibly big deal. While Waymo had been offering rides in the Phoenix area previously, committing to fully autonomous rides to the public was a monumental step. Previous customers were incredibly limited, just 1,500 were eligible, and had to be approved by Waymo to have access to their ride-hailing service, which was only permitted after they had signed a non-disclosure agreement. To have the confidence to make their service commercially available in a form, similar to how it will exist for the foreseeable future, is a very powerful statement from the company. This is also a major first step in demonstrating that Waymo’s autonomous taxi is already capable of receiving the regulatory approval that is so sought after. As such, I would expect Waymo to start seeing regulatory approval in more areas as soon as the coming six months year and they will continue to expand their service as time progresses.
Waymo and Tesla are at essentially opposite ends of the spectrum when it comes to their approach to autonomy, while Cruise, Lyft, and Uber all fall into similar categories, or somewhere in between. Cruise Automation is closer than either Lyft or Uber bringing their product to market, so it makes more sense to discuss their path to regulatory approval first. Cruise is quite fortunate, in the sense that they’re able to reap the rewards of Waymo’s dirty work and follow in their footsteps. After Waymo blazed the path to gaining regulatory approval, reducing the stigma behind autonomous vehicles, and improving public opinion of the technology, Cruise will have a bit of an easier time with their endeavors.
Continuing the idea of Cruise following in Waymo’s footsteps, it makes sense to look at what footsteps Cruise would be following exactly. In April 2017, Waymo launched its first early rider program. The program approved 400 riders during the first year of operation and service remained pretty limited. It wasn’t until November 2019 that Waymo started offering some rides without a backup driver at the wheel and rides became meaningful. It took Waymo less than a year to get from that stage to a fully autonomous ride-hailing service available to the general public.
As discussed above, Cruise has been ramping up its test capabilities of late, putting them in a situation that is similar to where Waymo was around this time two years ago. However, Cruise’s rapid rate of progress is not something to overlook. While perhaps not as popular or as well known as the other companies discussed in this article, Cruise is more than capable of being a leader in the field. The company registered a 134.8% increase in the number of miles that their autonomous vehicles drove without human interference from 2018 to 2019. With 12,221 miles driven without human intervention, they are still 998 miles behind Waymo’s 13,219 miles, but Waymo’s improvement came at just 20%. To be fair, Waymo had less room to grow, but Cruise’s growth is still quite impressive and not something to overlook.
Because of this rapid rate of growth, as well as Waymo’s success being able to improve public opinion of the technology and set precedents to follow, Cruise is well on its way to success. Receiving a permit, similar to that which Waymo recently received in Arizona, is likely to happen by the end of 2021. Further approval in greater areas is unlikely to be seen until 2022.
Lyft and Uber are, undeniably, far behind their future competitors. While Uber has the most still to prove, Lyft, as the latest entry to the competition, is also lagging in the sector. Uber’s progress has been at a near standstill for around two years now and that’s bound to take a toll on the development of their autonomous vehicle. With little to no information publicly available about their simulator, it stands to reason that the company has nothing worth bragging about on that front either. With continuously dismal results, I wouldn’t expect Uber to launch an autonomous ride-hailing service anywhere until mid-2024, accompanied by a slower build-out than its competitors.
Lyft is making strong progress, able to skip the earlier mistakes of rivals who had nothing to base their development on and is making a case to come out of this better than Uber. Still, they are noticeably behind even Cruise Automation and their late entry to the field has, obviously, put them at a disadvantage. With current progress and their rate of innovation, Lyft will likely begin to offer a service at the end of 2022, or early 2023.
The New Market
The arrival of autonomous driving will bring about a huge shift in how the ride-hailing market operates. Waymo, Tesla, and General Motors aren’t even players in the ride-hailing market, but they could soon be some of the biggest names. The first-to-market advantage for autonomous ride-hailing is huge, due to the advantages that come with offering a driverless cab. Currently, about 75% of fares paid to Uber for a ride, go towards the driver’s payment and 80% goes to a Lyft driver. Essentially, getting rid of the driver can reduce the cost of hailing a ride by between 75% and 80%. Now, companies entering the space may decide to cut rates by closer to 50% or 60% in order to further boost margins, especially before much competition is introduced. By offering such dramatically lower rates, the first company to be able to offer an autonomous service will be able to quickly topple person-driven ride-hailing services.
Speaking of first to market advantage, Waymo will likely be the first out of the gate. Well, they already are the first out of the gate. This first-to-market advantage will prove invaluable as they are able to build brand recognition before their competitors. Additionally, when dealing with something like driving, building trust is incredibly important. The ability of Waymo to prove their safety capabilities before anyone else and build a strong reputation will likely create greater trust among consumers and serve as an invaluable marketing tool. The advantage of being first to market will also allow consumers to get more accustomed to a given experience and make them less inclined to change. There is also the possibility to introduce rewards programs that may reward continued usage, further inclining users to stick with their first experience.
Large urban areas are the largest areas of demand for ride-hailing services, and likely will be for quite a while, which explains the prioritization of them by companies like Waymo and Cruise. By ensuring that they will be the first to these markets, they have positioned themselves as well as they can to enter the market. Therefore, upon first glance, Tesla shot themselves in the foot a bit with their ride-hailing service. There is no doubt that they took the best approach from an automotive perspective and gaining nationwide approval in one go will be a pretty incredible feat, but they will do so much later than their competitors. Tesla opted for a solution that looks sleeker, costs less, and has mass marketability, but is far more difficult to develop to a similar degree of safety as their competitors. Their late entry into the market, which I predicted for 2024, will see them start behind companies like Waymo and General Motors. However, their mass-approval may allow them to come to some cities before their competitors do.
In many ways, Tesla is the Apple of auto-manufacturers, and with that comes an incredibly loyal fanbase, which can often be quite toxic. However, these fans of the company will likely wish to ride on Tesla’s autonomous ride-hailing service over any others and will play a meaningful role in making up for lost time. Additionally, with nationwide approval, Tesla will likely be able to enter some markets before its competitors as they progress to cover as many people as they can, prioritizing population hubs across the nation. Tesla, as a result, will likely dominate suburban markets upon approval and will draw enough of a crowd in cities, based on brand alone, to be successful. As such, Tesla’s later entry will not be as damning as Uber or Lyft.
Looking at all of the information that I’ve laid out thus far, it seems reasonable to expect me to label Google’s Waymo as the new king of ride-hailing, but I don’t believe it will be. Instead, Detroit-based General Motors will take the crown. The reason for this bold prediction is simple – manufacturing. The benefit of such a capability may not be immediately obvious, but consider a city a peak demand. Averaging 500,000 rides per day in New York (taxi and other ride-hailing services combined), a massive fleet is required to supplement this. Again, in New York City, the required fleet far surpasses 100,000 to attend to the needs of the residents. While not all 100,000 vehicles are required at all times, periods during rush hour will see more demand and the greater number of available rides will allow for lower wait times. Even beyond rush hour, the more cars there are available in a fleet, the lower the average wait time will be because there is a greater chance that a vehicle will be located nearby.
More vehicles will also allow for faster data collection. As all companies fight to enter the most areas the fastest, they must do so by gaining approval from one city at a time. With the advantage of a larger fleet, Cruise will be able to collect data faster than Waymo, or its other competitors, and will likely be able to enter more new areas faster. Therefore, backed by the manufacturing powerhouse that is General Motors, with annual production nearing 8 million units, Cruise Automation is poised to challenge, and defeat, Waymo with the combined utility of faster data collection and greater availability of rides.
General Motors has committed manufacturing resources at its Lake Orion manufacturing plant, located in Michigan, USA. With 40% of the components being unique from the typical production model Chevy Bolt, the direct integration of the manufacturing process allows for a more streamlined product. This provides even greater of an advantage than just sheer manufacturing power. Waymo, on the other hand, has already ruled out manufacturing its own vehicles. Striking deals with Jaguar (TTM) and Chrysler (FCAU) for their I-Pace and Pacifica Hybrid models, Waymo is regulated to outfitting existing production cars with their technology. Beyond limiting their fleet size to the number of vehicles promised in a prearranged agreement, it is significantly more expensive. While Jaguar and Chrysler are looking to profit off of their deals, Tesla and General Motors aren’t selling their vehicles to anyone and are able to more efficiently equip their vehicles with the required hardware during the manufacturing process. This economy of scale dramatically reduces the cost and time spent on each vehicle. The I-Pace, starting at nearly $70,000, will be an especially pricey option for Waymo. Uber and Lyft employ a similar strategy, with Volvo (OTCPK:VOLAF) or Ford (F) and Chrysler, and suffer from the same consequences.
Moving on from individual companies, the ride-hailing market as a whole will begin to look dramatically different. This goes back to the price of ride-hailing being able to drop dramatically with the removal of the driver. This is where an incredibly intriguing idea begins to take form – Mobility-as-a-Service (“MaaS”). MaaS, sometimes referred to Transportation-as-a-Service (“TaaS”), is a shift from private ownership of transportation devices, cars, to mobility as an on-demand service. The basic principle behind this is that, eventually, it would be less costly to call an autonomous vehicle to pick you up whenever you need to drive somewhere, than buying and maintaining a vehicle of your own. While this ignores the cultural significance and undeniable freedom expressed by owning your own car, this principle holds much less ground for people living in cities. It is already more common for people in cities to not own a car than those in less densely populated areas, meaning that there isn’t much cultural significance behind owning a car to overcome in cities. This is likely due to the lack of reasons to own a car in a city. I have serious doubts that we’ll ever truly see the mass adoption of MaaS in suburban America, but the likelihood for city residents to adopt this seems incredibly high. Owning cars in cities, as a result of autonomous driving, may become a thing of the past. Within this generation, it will begin to make tremendous financial sense to adopt MaaS in cities and, without the cultural barrier to overcome, car ownership in urban areas will continue to drop. As a result of everything I’ve discussed, ride-hailing usage will rise dramatically – especially in urban centers, as analysts expect this market to jump to $230.4 billion in 2025, from $43.44 billion in 2019.
Individual Performance in the Market
This growth will benefit all players involved in the race to autonomy, but those who can capture the greatest market shares stand to reap the biggest rewards. As the market emerges, it may not surprise you when I say that I expect Waymo to capture the largest share of the market. In the early stages of the market, there will be very little direct competition among each company. Instead, competition will play out in the sense that each company is fighting to see who can enter the most areas the fastest. This will create effective monopolies in cities across the country until it becomes more economically viable to begin competing with another ride-hailing adversary than to try and enter a market with far fewer potential customers.
This is where the first to market advantage comes in and where Waymo’s going to see its early lead. However, once direct competition begins, Waymo will likely begin to falter. With the manufacturing advantage discussed above, General Motors and Tesla will begin to induce a market shift as the companies begin to fight over the same customers. This will lead to a market separation of “the auto manufacturers vs the rest” with General Motors ultimately claiming the lion’s share.
Waymo may actually be able to compete against Tesla a bit more effectively than it can with General Motors as I expect Tesla to have the lowest trust rating among its competitors due to its late entry and past history. I expect this to outweigh the excitement around the brand, which primarily consists of younger individuals. Tesla’s far lower manufacturing capability, compared to General Motors, will also limit the pace at which Tesla can build out its fleet. Additionally, Waymo’s first to market status, and backing by Google, will likely lead to the greatest trust of any other brand. As a result, I anticipate Waymo to act as a strong competitor to Tesla, hindered only by its fleet size.
While many analysts may point to Uber and Lyft’s current dominance of the ride-hailing market as an indication of its future success, the two do not correlate. While both Uber and Lyft offer rewards programs, only Uber directly rewards customer loyalty. Lyft’s benefits program offers the same rewards to new riders as it does a constant customer. While Uber’s tiered loyalty program does promote more consistent use, the rewards aren’t nearly strong enough to retain customers when a cheaper autonomous vehicle enters the area. While the brand recognition of the two companies may also help a bit, autonomous vehicles introduce an entirely new element to the ride-hailing market. Trust remains an incredibly important factor for autonomous vehicles and both companies have built their reputation around person-driven ride-hailing. Uber’s reputation when it comes to safety in their autonomous vehicles isn’t the greatest and, whether or not it’s fair, the negative perception will likely cloud the brand for years to come. The autonomous ride-hailing market will likely evolve to favor General Motors and Tesla, with Waymo following close behind the latter. Unfortunately, Uber and Lyft will likely be taking the scraps of the other companies in a bid to maintain some form of relevance.
Uber, unsurprisingly, has the most to lose. The first thing to look at is Uber’s expenditures on their autonomous driving development. From 2016 up to March 2019, Uber Spent $20 million on the development of autonomous vehicles, which totals $1 billion during the four years and three-month period. Since then, under the category of “ATG” on Uber’s balance sheets, autonomous vehicles have continued to weigh on the unprofitable company’s expenses. ATG expenses, since Uber released its first report of ATG expenses report for the first quarter of 2019, have averaged $116.3 million (Q3 2019, Q4 2019, Q1 2020, Q2 2020). With six quarters of data, expenses total $698 million, though it should be noted that the ATG expenses also cover the expenses of some other ventures from the company, such as Uber Elevate. With this data, it is fair to assume that Uber is continuing to spend around $20 million per month on its autonomous vehicle development.
It’s not necessarily a bad thing for Uber to be spending money on the tech, after all, they need some way of developing it. However, due to a delayed entry into the market, Uber won’t be able to decrease spending for some time. This means that autonomous vehicle development will continue to weigh on the company. While Uber’s valuation clearly isn’t reliant on its current profitability, the bet on the company to turn large profits in the future seems over-blown.
While Lyft doesn’t release any data to specify the costs of developing their autonomous vehicles, it’s safe to assume that the development isn’t cheap. Their quicker entry into the market will allow the company to reduce this financial burden before Uber is able to, but investors should continue to expect high expenses through 2022. However, one expense that hasn’t been addressed yet, is the creation of an effective autonomous vehicle fleet.
Without the deep pockets of a company like Google, or the manufacturing capability of General Motors, Uber and Lyft will almost certainly need to turn to outside sources to fund their fleet build-out and upkeep. Some analysts expect the companies to turn to a similar strategy of REIT’s, instead creating FLEITs (car fleet investment trusts) in order to finance the operation. This strategy makes quite a bit of sense, but will contribute to lower profitability as the company wouldn’t own the vehicles in its fleet and fleet owners would have to be compensated. This will make both Uber and Lyft even less competitive in the market as lower profitability will contribute to less investment in the market and put them at a further disadvantage.
If neither company were to utilize this strategy, they would be forced to incur massive amounts of debt or dramatically reduce the pace at which they can deploy their fleet. Essentially, no matter which strategy Uber and Lyft choose to employ, they will suffer from dramatically reduced profitability compared to their competitors. At the end of the day, each company has a best-case scenario where their valuations can no longer be justified as their market share is encroached upon by new competitors and profitability is significantly lower than investors anticipate.
How to Trade
There is an important distinction here to make about the information provided about Uber and Lyft. This analysis concerns the long-term implications of their changing market and how their current positions make them vulnerable to incoming threats. As such, I would recommend taking a short position in Uber, but not Lyft.
Looking at both companies’ performance recently, and since their IPOs, Lyft has done incredibly poorly and currently sits at a valuation that seems fair for where their future lies in the market. However, this isn’t to say that I would recommend initiating a long position in the company as the company will still struggle in the new market. While Uber has fallen from its listing debut, it has not seen close to the levels of failure as Lyft has and still retains most of its value from its initial listing. Additionally, the company is positive on the year and is showing further signs of value growth in the future.
As I stated above, Uber’s valuation is based on its ability to profit off of its future business. What this should mean is that its current position in the ride-hailing market is not as important as where it will be five or ten years down the road, when autonomous vehicles are roaming the streets. So, looking at Waymo’s $30 billion valuation and Cruise’ $19 billion valuation makes Uber’s value of nearly $65 billion appear quite irrational. This holds even more true when considering the idea that Uber’s market share will only decrease in time, while Waymo and Cruise will do the opposite. Lyft’s $8 billion valuation in comparison seems much more reasonable.
With short volume at around 19%, it’s fair to be concerned about borrowing fees and the ability to cover the position if need be. Luckily, driven by high volume trading, Uber maintains a low short interest, or days to cover ratio, of a mere 3 (author’s calculations using average volume [3 month] and shares short). Beyond lowering the risk of initiating a short position in the stock, this allows for a quite attractive borrowing fee of just 3.75%. For these reasons, Uber’s short is the most appealing investment to make for the autonomous vehicle revolution.
In terms of when I would recommend initiating this position, now seems to be as good a time as any, especially as I don’t anticipate Uber’s value to increase significantly in the coming months, if at all. There is also no guarantee that the low borrowing fees will last, though it seems likely, and, with Waymo’s recent approval in Arizona, it appears that the autonomous revolution is starting to pick up. As it becomes increasingly apparent that Uber is far behind its competition, its future position in its market will be the subject of increased scrutiny as investors grow more skeptical of its ability to compete in an autonomous market. This higher level of skepticism around the company’s place in the ride-hailing market will be the catalyst for Uber’s fall from current highs, while weakened market performance in the coming years will solidify it.
With Waymo, Tesla, and General Motors, all of this is clearly positive news. However, evaluating each of their futures as well would turn this, already long, article into something far too long. As such, my analysis of these three companies has been limited. All three have major business elsewhere that could dramatically affect their future trading values, regardless of, or in addition to, the positives coming from their autonomous pursuits. As such, if looking to invest in any of these three companies, this is useful information to consider, but make sure to also consider the various other operations that they all undertake.