Uber Technologies, Inc. (UBER.US): Can autonomous driving really kill the "American version of Didi"?

As one of the most favored long-term investment targets by Dolphin Research, Uber Technologies, Inc. (UBER.US) has lost its former market favor amidst concerns about the revolutionary changes in the entire automotive and transportation industry due to the rapid development of autonomous driving technology (especially when Tesla is flourishing and has loudly announced its Robotaxi plan). The stock price dropped from $87 to below $60 between October 14th and mid-December 24.10.14~24.12, representing a retreat of nearly 1/3 and clearly underperforming the rising performance of the S&P index during the same period. However, despite this, the market's long and short battles around Uber are still quite intense. Well-known investor Bill Ackman has loudly announced his bullish view on Uber, and in an investor survey conducted by JPM in early January, Uber was also listed as one of the most favored mid-cap targets. Disagreement is one of the sources of opportunity, creating opportunities to invest in good companies that are undervalued because they are not unanimously favored. Therefore, taking advantage of this opportunity, Dolphin Research will explore whether the post-correction Uber fully reflects the impact of autonomous driving technology. Is the current Uber offering significant value for money, or is it still standing on the "cliff edge," seemingly "valuable" but actually a "trap"? From the perspective of analysis, this article aims to explore the impact of autonomous driving becoming a daily technology on the transportation and ride-hailing industry, the actual development path and timing of autonomous driving technology, and which companies and technology paths will eventually prevail, which is not the subject of our discussion, so we make the following two assumptions as the basis for later analysis: 1. Autonomous driving technology can consistently achieve driving abilities not inferior to or even surpassing those of human drivers in terms of safety, speed, and other aspects. And the government does not restrict the widespread implementation of autonomous driving. 2. The autonomous driving industry will produce multiple suppliers (more than a few) with no significant technology gap, and some of them will provide these technologies to third parties. Below is a detailed analysis: I. In the era of autonomous driving, how will the demand for ride-hailing services change? Before discussing the impact of autonomous driving on specific companies and the ride-hailing business model, Dolphin Research believes that the first question to be answered is what kind of impact autonomous driving technology will have on the total volume and structure of (automobile or local) travel demand? First of all, we believe that the fundamental change brought about by autonomous driving technology is the liberation of the manpower and its labor costs required for driving. In other words, its innovation mainly manifests in cost reduction. Based on this judgment, we believe that: 1. The advantages of autonomous driving technology on the growth of travel demand should not be significant. The underlying logic is that travel is just a tool, not the purpose itself, and few people would "travel just for the sake of traveling." Therefore, compared to significant technological innovations that significantly expand the destinations people can choose or significantly reduce travel time, such as cars vs. horse-drawn carriages, airplanes vs. cars, we believe that the cost reduction of driving brought by autonomous driving technology should not show a clear advantage in the growth of travel demand. In the medium to long term, if autonomous driving can completely change the current form of cars or the operation of highway traffic, it is more likely to lead to a more significant increase in the total demand for car travel (not elaborated). 2. The impact of autonomous driving technology on the travel market should be reflected more in the changes in the structure of travel by car (the proportion of various avenues). First of all, from historical data, the changes in the proportions of various channels of car travel in China (by amount) from 2018 to 2023, the proportion of private car travel in China is highest and continues to rise (recently reaching about 85%); the proportion of public transportation continues to decline (from about 12% to about 5%); the overall proportion of ride-hailing + taxi travel has stabilized at about 7% to 10%, with the trend being that ride-hailing is eroding the market share of taxis. Similar to this, in the United States, the structure of commuting methods for residents, from structure and trends in change, the proportion of private car use is vast (over 90%), public transportation travel ranks second (accounting for between 3% and 5%), and taxi and other (including ride-hailing and motorcycles, etc.) have a very low proportion of only between 1.4% and 1.8%. In terms of trends, the most significant change is the significant increase in telecommuting (WFH) (5.7%/17.9%/13.8% in 19/21/23), resulting in reduced proportions of other travel modes. Excluding the influence of WFH, the main changes before and after the epidemic show that the proportions of private cars and taxis & others are increasing, while public transportation is the main loser in terms of market share. However, as the impact of the pandemic fades, by 2023 a significant portion of the increased shares of private cars, taxis, and others has already returned to public transportation. (It should be noted that the above data are only in the context of work commuting, which may underestimate the share of ride-hailing & taxis in all scenarios). From the analysis above, over the past few years, regardless of the changes in travel structures occurring in China and the United States, the decline in the proportion of public transportation, with an increase in the proportion of private cars and ride-hailing, has been observed. (Of course, the influence of the preference for privacy in travel due to the COVID-19 pandemic is also at play). The impact of short to medium-term autonomous driving on the structure of travel is most likely: Private cars & ride-hailing services continue to erode the market share of public transportation, Ride-hailing services accelerate the replacement of human-driven taxis, However, due to the significant advantages of autonomous driving technology for private cars or ride-hailing services, it is difficult at present to say which of the two will benefit more after the widespread adoption of the technology, and we do not attempt to make a judgment on this. We believe that the reasons why autonomous driving ride-hailing services can replace public transportation are as follows: According to research, the two most common reasons people use ride-hailing services (as a substitute for other modes of travel) are: Ride-hailing services are cheaper than taxis & private cars, Ride-hailing services are more convenient than public transportation & private cars. In summary, these two key points--affordability & convenience. With the widespread adoption of autonomous driving, ride-hailing services are expected to be more convenient than public transportation.There is also hope to be close to the cost of public transportation in terms of usage cost. According to ARK's calculations, after the maturity and scale of the autonomous driving technology, the operating cost of autonomous ride-sharing vehicles can be reduced to about $0.25 per mile (we can obtain a similar number after our own calculations, but it is worth noting that this number only considers costs and has not yet considered the profit space for operators), which is close to the average price of $0.22 to $0.23 per mile for subways and buses, with almost no difference.4. How much incremental market space can autonomous driving technology bring to the ride-hailing industry from a quantitative perspective? Incremental market space gained by eroding public transportation market share: According to previous data, the usage rate of public transportation in commuting scenarios in the United States is approximately around 4%. Considering that the usage rate of public transportation in other scenarios should be even lower, let's assume an overall usage rate of around 3% for public transportation (around 70%-80% of 4%). After the maturity of autonomous driving, we believe that it is possible to aim for a 50% share of the current public transportation share taken by private cars and ride-hailing. (However, due to congestion, it is also unlikely for autonomous vehicles to completely replace buses or subways). We assume that 50% of the usage scenarios of public transportation will be replaced by autonomous vehicles, with private cars and commercial vehicles taking them over in a 50%/50% ratio. Based on the calculations in the graph below, the usage scenarios taken from public transportation could increase the equivalent of 40% of the current market size of ride-hailing and taxi services (in terms of volume). Within the operational vehicles, ride-hailing continues to seize market share from traditional taxi services: According to official data from New York City, traditional taxis are expected to maintain around 20% of the market share between 2023 and 2024. Based on domestic experience, the market share of taxis nationwide in the United States should be lower than that of New York City (many small cities may not have taxi operations, but the concept of ride-hailing is more prevalent). Considering that there's a high probability that human-driven taxis will not be completely replaced, we assume that after the popularization of autonomous driving technology, the share of ride-hailing within operational vehicles will increase from over 80% to over 95%. Combining the external grab of public transportation share and the internal replacement of traditional taxis, without considering industry natural growth, in the example of the United States, autonomous driving technology could potentially increase the market size of ride-hailing by approximately 66% in terms of volume. II. In the era of autonomous driving, how will the competitive landscape and business models of ride-hailing change? 1. Different pricing = Different platform revenue According to the analysis above, autonomous driving technology is expected to increase usage in the ride-hailing industry in the United States by around 2/3, but this growth in volume does not necessarily mean an increase in the total profit space of the ride-hailing business model, provided that ride-hailing pricing needs to noticeably decrease. Therefore, the key question is how the UE model of the ride-hailing business model will change in the era of autonomous driving. In a nutshell, this is the balance between platforms retaining more net income for themselves (more profit) or adopting lower prices to attract users (lower pricing). We provide a simplified UE model in the table below, containing only four factors: labor costs (those that can be replaced by autonomous driving technology), non-labor costs (costs that cannot be replaced by autonomous driving such as depreciation and energy), platform retained revenue, and prices paid by end consumers. Looking at different scenarios under different pricing standards for autonomous driving ride-hailing, what changes can be expected in the platform's retained income compared to the current situation: US market: According to ARK's calculations, in the mature state of Robotaxi (referring to autonomous driving ride-hailing), the cost per kilometer is $0.16 (equivalent to $0.25 per mile). If, in the era of autonomous driving, the platform's net income (after deducting driver shares) remains at the current rate of about $0.35 per kilometer for Uber, the end pricing would be $0.51 per kilometer, which is around 41% of the current ride-hailing prices, a significant discount. However, it is worth noting that this pricing ($0.51) is still higher than the $0.43 per kilometer cost of using a private car and the less than $0.25 per kilometer cost of public transportation. Therefore, looking at it from a different perspective, if the pricing of autonomous driving ride-hailing is raised by 30% over the cost of using public transportation, meaning $0.30 per kilometer (equivalent to 70% of the current private car cost), then the retained net income for the platform and operators would only be $0.14 per kilometer, about 40% of the current levels. Chinese market: According to calculations from Dolphin Research (also referring to other institutions), in a completely mature technological state, the operating cost of ride-hailing in China can be as low as $0.1 per kilometer. The key assumptions behind this include the cost of a regular Robotaxi dropping to around 150,000 yuan, no longer needing any safety personnel, and maintenance costs including insurance and repair costing around 20,000 yuan per vehicle per year (55 yuan per day). Under the above assumptions, if the net income of Robotaxi remains around $0.04 per kilometer, the end pricing for the passengers would be $0.14 per kilometer, equivalent to 40% of the current ride-hailing prices and 61% of the cost of using a private car. Common characteristics seen in the markets of China and the US are that after the technology matures and is operated on a large scale, even if the pricing of autonomous driving ride-hailing is about 60% cheaper than the current prices, the operators of Robotaxi can maintain comparable average income per ride to the current ride-hailing platforms, which is sufficient to provide attractive prices. 2. In the new era, how will the current market landscape change? All the analyses above have been from the perspective of viewing the ride-hailing industry as a whole. However, with the popularization of autonomous driving technology, more industry participants (including but not limited to providers of autonomous driving technology and operators of driverless ride-sharing) will inevitably enter the market, bringing significant changes to the relatively stable competitive landscape. The primary influencing factor will be how providers of autonomous driving technology position themselves and enter the ride-hailing industry, which can be roughly categorized into the following scenarios: Autonomous driving technology being monopolized by a few tech giants and completely closed-source, such as Tesla and a few other car manufacturers or autonomous driving unicorns. Although this possibility cannot be completely ruled out, Dolphin Research believes it is unlikely. If this were to happen, there is a considerable chance that the ride-hailing industry could be monopolized by these few tech giants, but this scenario will be omitted in this discussion. Autonomous driving technology providers building their own ride-hailing platforms and competing directly: Autonomous driving technology providers like the current leading Tesla or Waymo could enter the ride-hailing industry in this way.By leveraging technological or hardware advantages, compete with existing ride-hailing platforms in a direct or exclusive manner through building their own ride-hailing matching platform or large-scale self-operated fleet (for example, not allowing their own self-driving vehicles to operate on third-party ride-hailing platforms). Car manufacturers or other Robotaxi providers cooperate with ride-hailing platforms rather than engage in direct competition: autonomous driving technology becomes a common technology. Autonomous driving capacity enters the ride-hailing industry through various channels, including official cooperation with existing ride-hailing platforms, third-party operators purchasing (leasing) autonomous driving vehicles to join ride-hailing platforms, and some autonomous driving companies also attempting to directly serve consumers by building their own platforms. Although it is difficult to predict the exact changes in the market landscape of the ride-hailing industry in the era of autonomous driving, Dolphin Research speculates that in the future, there will likely be a main collaboration between autonomous driving technology providers and existing platforms (scenario ) with autonomous driving technology providers building their own platforms as a secondary option (scenario ). In terms of pace, we believe that current leaders - Waymo and Tesla have the opportunity to accumulate enough user awareness and traffic in the closed-loop ride-hailing business model before autonomous ride-hailing becomes widespread, by leveraging their first-mover advantage (Waymo) and sufficient consumer awareness (Tesla) to establish new platforms that can operate independently. However, future unicorns in autonomous driving are more likely to enter the ride-hailing industry mainly as providers of transport capacity, needing to collaborate with existing ride-hailing platforms that have the ability to aggregate users. Depending on the positioning of new players in the industry, the impact on existing ride-hailing platforms such as Uber will be reflected in two main aspects: First, unicorns or car manufacturers that operate independently directly participate in the competition for consumer sources and market share, affecting the market share of existing ride-hailing platforms. As autonomous driving vehicles replace individual drivers, there is a likelihood of the emergence of group and corporate-operated fleets of autonomous driving capacity. The transition from independent and dispersed individual drivers to a scaled fleet of autonomous ride-hailing vehicles could weaken the negotiating power of ride-hailing platforms upstream, potentially reducing the share of revenue and profits they retain in the ride-hailing business. In other words, this could affect the average income and profit margins retained by the ride-hailing platform. In simple logic, with the increasing number of advancements in autonomous driving technology, insignificant technological differences, and the lack of industry-leading companies, the impact on existing ride-hailing platforms like Uber will likely be minimal. How much impact will new players in the ride-hailing industry have on existing players? In this regard, we will focus on the two perspectives mentioned at the end of the previous section: Can the leading unicorns in autonomous driving seize a significant market share from current ride-hailing giants? After the entry of autonomous driving capacity into the market, what will be the impact on platform profitability? 1. Developing and operating autonomous vehicles is not an easy road Firstly, from the perspective of autonomous driving developers independently operating ride-hailing platforms, although there are quite a large number of enterprises currently independently developing autonomous driving technology (mainly domestic car manufacturers conducting independent research and development, while overseas third-party unicorn organizations dominate), a significant portion of them have faced challenges in their development journey, with only a few providing autonomous ride-hailing services to the public. Specifically, in the European and American markets, among the top autonomous driving technology developers (excluding Tesla), only Waymo and Zoox have achieved some success in operating autonomous ride-hailing services, with the former gradually expanding its operations and the latter only conducting limited trials. In contrast, Cruise, Motional, and Aurora have generally paused or terminated their operations of autonomous ride-hailing, either cutting expenses and returning to a pure focus on technical development or shifting their focus towards the development of autonomous driving truck technology for freight transport. Even large enterprises like Apple and Uber that have independently developed autonomous driving vehicles have faced failures and project cancellations. Therefore, because of the significant capital investment required for independently operating ride-hailing platforms and fleets, it is not a common choice for autonomous driving developers to pursue independent operation (consistent with our view). Thus, in the following sections, we will focus on Waymo to observe the current state of autonomous ride-hailing operations. 2. The lone wolf Waymo opts for a 1P/3P approach The following diagram briefly outlines Waymo's development from its establishment in 2009 as a project department under Alphabet Inc. Class C, to its first public launch of autonomous ride-hailing services in Phoenix in 2020, and its expansion to providing public services in four markets, which we will not delve into. What we are more interested in is Waymo's exploration and testing of different operational models. It can be seen that Waymo has explored four different models based on the combination of the roles of "platform provider (responsible for acquiring customers and dispatching orders)" and "fleet management provider (responsible for managing and maintaining capacity)": Pure 1P model where both the platform and fleet are managed by Waymo; Pure 3P model where Uber is responsible for the platform acquisition and fleet management (Waymo only provides autonomous driving technology and vehicles); Hybrid 1P/3P model where Waymo manages the fleet but can acquire customers from both its own platform and Uber's platform; Waymo acquires customers through its own platform but outsources fleet management to a third party, similar to a light asset model like model . Therefore, even if Waymo chooses to build its platform and operate independently while leading the industry, it still opts to collaborate with third-party platforms or fleet managers. Hence, we believe that even after the widespread adoption of autonomous ride-hailing, a mixed 1P/3P operating model is likely to prevail. 3. The rapidly growing Waymo is not far from Uber/Lyft So how big is Waymo in terms of its current operating status and scale? In summary, Waymo mainly operates in the four regions listed in the table above; in terms of fleet size, according to news reports, Waymo currently operates a total of several hundred autonomous ride-hailing vehicles externally (possibly around 700-1000 vehicles), and in terms of the volume of rides, according to reports 25At the end of March, Waymo's total number of orders in four operating regions has reached 200,000 per week.Although its absolute volume compared to the daily tens of thousands of orders in the California region for Uber and Lyft, there is still a difference in scale. However, when considering constraints such as limited operating vehicles and city numbers, there may be no significant difference in operating efficiency between autonomous ride-hailing vehicles and Uber and Lyft. From the current situation, as long as autonomous ride-hailing vehicles expand their fleet size and operating market, there is a good chance of reaching a similar scale to leading companies like Uber and Lyft. According to official California data, by the end of 2024, Waymo's autonomous vehicles in California were averaging close to 24 orders per vehicle per day. Assuming each order takes 30 minutes (which may be short), plus about 20% to 30% idle waiting time, it means that the "effective" operating time of autonomous ride-hailing vehicles in California has reached 15 hours per day. In addition, according to news reports, the number of daily orders for human drivers working long hours is only around 25-30. This shows that the usage rate of autonomous ride-hailing vehicles deployed in California is quite high, and the demand for autonomous ride-hailing vehicles from local residents can be said to be relatively strong (at least given the current supply of capacity). Furthermore, research shows that the strong demand for autonomous ride-hailing vehicles in California is not based on significant price discounts. Waymo's usage price in the Los Angeles area is about $6 per mile (or $3.5 per kilometer), which is comparable to Lyft's standard pricing and only about 10% lower than Uber's standard pricing. This means that autonomous ride-hailing vehicles in California have already established an effective customer acquisition capability (without relying solely on third-party platforms like Uber to acquire customers). According to Yipit data, by the end of 2024, Waymo's market share in the ride-hailing market in the San Francisco area (based on order amount) had caught up with Lyft, with a market share slightly above 20%. In comparison, Uber and Lyft's market shares have decreased by around 10% from their peak in 2023. This indicates that at least in the San Francisco area, Waymo has significantly competed with existing ride-hailing platforms and has reached a comparable scale. Therefore, given future technological developments in the autonomous ride-hailing sector, there is a possibility that one (or even 2-3) autonomous ride-hailing platforms could emerge with market shares comparable to current ride-hailing platforms. Additionally, with autonomous driving technology becoming a fundamental infrastructure rather than being monopolized by a few companies, existing ride-hailing platforms may also leverage autonomous driving capabilities. Thus, the probability of new platforms completely disrupting existing leaders like Uber is not high. (Continued in next comment)"Stand guard" to keep up with the evolution of technology and adapt accordingly.2. How does pricing affect self-driving cars? Obviously, estimating the value of Uber in the era of self-driving cars is a question that currently has no precise answer, to the point where the market has basically given up trying to answer it (perhaps abandoning it is a rational choice). However, the value of this article lies in giving everyone a rough reference, and Dolphin Research will provide a way to answer this question. It should be noted that the calculations below are not so much about predicting what will actually happen in the future, but rather about observing where the company's valuation should be based on logical and conservative assumptions under this hypothesis. First, our forecast period will start in 2027 and last for 5 years until 2031, assuming that the impact of self-driving cars on the ride-hailing industry will gradually be fully realized during these 5 years. So, the first step in solving this problem - what impact will self-driving cars have on Uber's ride-hailing volume? Based on the analysis above, we make the following assumptions: 1. In the era of self-driving cars, ride-hailing companies are expected to gain a maximum incremental volume of 66% as they take over all of the taxi market and half of the public transportation market, assuming that self-driving ride-hailing cars and self-driving private cars do not compete with each other. However, for this situation to occur, the pricing of autonomous ride-hailing vehicles must be close to the cost of using public transportation. Although this pricing, theoretically possible (due to the very low operating costs of autonomous vehicles), can be achieved, the profit margin left for the operators will be relatively limited. Taking this into consideration, we estimate that the maximum positive effect may reach 33% by 2031. 2. In the era of self-driving cars, the ride-hailing market is expected to see the emergence of two new autonomous ride-hailing platforms, in addition to the existing two major players. Although the ride-hailing business model, such as C2C and low economies of scale, typically only supports two major players, advancements in autonomous driving technology will reduce vehicle operating costs and may lead to a shift from C2C to B2C operations. This could potentially support three to five players in the market, with market share shifting to a distribution more like 50%:20%:15%:15% or 40%:30%:15%:15%. Therefore, Uber's market share by 2031 is estimated to be around 60%, maintaining its position as the industry leader. Taking into account the combined effects of the above points, we estimate that by 2031, under the influence of self-driving cars, Uber's ride-hailing orders will be 80% of the original forecast. The second step in solving the problem - how will the average revenue/profit of Uber's ride-hailing business change in the self-driving era? This question may be even harder to answer than the first. It involves two key questions: how much will ride-hailing prices decrease in the self-driving era compared to the current prices, and what net income can be retained by the platform based on the new pricing? Although it is difficult to predict the actual pricing of autonomous ride-hailing vehicles, based on previous conclusions, it is clear that in an ideal scenario, autonomous ride-hailing vehicles could reduce end-user pricing by up to 60% while ensuring that the net income of operators remains unchanged. Additionally, the share of income that the platform retains from vehicle operators may increase in the self-driving era, possibly leading to a decrease in the profit retained by the platform. Therefore, for conservative estimates, assuming a 60% decrease in pricing of autonomous ride-hailing vehicles, we predict that under this scenario, the net income retained by operators (including platform operators and fleet managers) will remain consistent. We also assume that the platform, currently enjoying 100% of retained income, will need to allocate 20% of this income to fleet managers in the self-driving era. Taking into account the 20% decrease in Uber's average revenue over the forecast period compared to the original expectations, while the average costs of Uber (such as marketing, customer service, and management) theoretically may not decrease directly due to the impact of self-driving cars, we conservatively estimate a 10% decrease in costs over the forecast period. Therefore, under these conservative assumptions, the average profit of Uber (represented by adj.EBITDA) in the self-driving era could decrease by approximately 40%. After considering the above impacts, Uber's adjusted EBITDA by 2031 would be approximately 46% of the original forecast. This article was sourced from the "Dolphin Research" WeChat public account; GMTEight Editor: Chen Xiaoyi.