Uber (Minutes): Deliberately slowing profit margin growth, AV will not be profitable for a long time

The following are the Minutes of the FY25Q3 earnings call for$Uber Tech(UBER.US) compiled by Dolphin Research. For earnings interpretation, please refer to "Uber: No issues with performance, but scared by Robotaxi?"

I. Review of Core Financial Data

Trips: Increased by 22%, marking the fastest growth since 2023.

User Growth: User base grew by 17%; user activity increased by 4%.

Financial Performance: Total bookings grew by 21%, with average fare remaining stable; adjusted EBITDA and free cash flow reached record highs.

Halloween Weekend Performance: Completed over 130 million trips (including delivery services); total bookings exceeded $2 billion.

Fourth Quarter Guidance: Total bookings are expected to maintain nearly 20% growth; adjusted EBITDA growth will reach the low 30% range.

II. Detailed Information from the Earnings Call

2.1 Key Information from Executive Statements

Strategic Focus:

Enhancing platform user engagement, cross-platform user spending is three times that of single-product users, with a retention rate 35% higher.

Seamlessly integrating human-driven and autonomous driving into a unified market to achieve unparalleled flexibility and efficiency.

Rapidly expanding the fresh retail sector, currently achieving an order total run rate of approximately $12 billion, with growth far exceeding the food delivery business.

Expanding income channels for 9.4 million drivers and couriers, including new digital tasks based on Uber AI solutions.

Supporting over 1.2 million merchant partners in achieving significant incremental sales through advertising, promotions, Uber Direct, and strategic partnerships.

Deeply integrating smart technology into the Uber system to enhance productivity, optimize operations, and provide a more personalized consumer experience.

Future Investments: Over the next few years, the company will continue to invest in product development, talent cultivation, and capital construction to deepen customer relationships, expand technological advantages, and sustain the established profit flywheel effect.

2.2 Q&A Session

Q: Can you discuss the path to increasing penetration in markets where 20% simultaneously use ride and delivery services? Also, can you elaborate on the recent collaboration with NVIDIA?

Both parties have invested in several technology providers and mentioned deploying 100,000 vehicles. Can you discuss the timeline and who will own the fleet in this scenario?

A: In every country where we operate both ride and delivery services, only 20% of users use both services simultaneously. For example, 30% of ride users have never tried Uber Eats, and 75% have never tried fresh retail. Markets with higher penetration rates are typically regions where both services have strong penetration, such as Australia.

We are now implementing specific plans to drive cross-platform behavior, such as adding convenient switch tabs in the app, creating personalized experiences for cross-selling based on scenarios (like "ride and dine"), and leveraging membership systems. Cross-platform users spend three times more on average than single-service users, which is our unique and significant advantage, and we are still in the very early stages of this.

We are very excited about the collaboration with NVIDIA. NVIDIA is creating a reference architecture for L4 autonomous driving through its "Hyperion," which will be available to any OEM manufacturer. Our strategic vision is that within the next decade, every new car sold will not only support L3 for personal use but also L4 for operation on shared mobility platforms like ours. This is very beneficial for our ecosystem, as L4 vehicles on the platform have been proven to achieve higher daily per-vehicle revenue.

We have also established a partnership with Stellantis, initially deploying 5,000 vehicles driven by NVIDIA technology, and expect to expand significantly in the future. NVIDIA not only provides the hardware platform but is also investing in building the L4 software stack.

Initially, we can intervene using the balance sheet to establish the economic model for these fleets. Ultimately, we believe these assets will be financialized, similar to how real estate investment trusts (REITs) own hotels, and there may be income-generating vehicles owned by private or public market funds to own these fleets.

Q: Regarding the delivery business, as you continue to expand the range of goods offered to users, can you discuss how much this has stimulated new user growth for Uber Eats or driven increased usage frequency across the platform?

Additionally, regarding autonomous driving, based on the previous question, in areas where you have already deployed autonomous vehicles, what have you learned about the impact of more road supply? How does it stimulate demand or affect pricing?

A: We are very excited about the accelerated growth of the delivery business. The third quarter marked its fastest growth in four years, accelerating by four percentage points. This growth spans multiple markets and stems from our investments in product improvements across various areas. Particularly in fresh groceries and retail, we are pleased to see it bringing users into the online food delivery space. This business has now reached a $12 billion run rate, growing significantly faster than our online food delivery business, and its variable contribution has turned positive, achieving high growth profitably, and complements the delivery business well. At the same time, this is part of our cross-platform strategy aimed at helping users flow between our three business lines (rides, delivery, retail).

Regarding autonomous driving, it is still very early. Our largest operation is with Waymo in Austin and Atlanta. We have observed that these markets are growing faster than other U.S. markets. For example, in Phoenix and Atlanta, their growth rate is more than twice that of other U.S. regions, which is undoubtedly a positive signal. It also makes driver hourly earnings in these markets very healthy.

For example, in Austin, where the most autonomous vehicles are deployed, driver hourly earnings actually exceed those in other U.S. regions. Whether this growth is correlation or causation is still too early to judge, but the market does indeed look healthy. Our collaboration with Waymo is excellent operationally, and Waymo vehicle utilization remains very high. We see a generally healthy market, which is a very good signal as we transition to a mixed network of autonomous and human drivers.

Q: Could you discuss the margin transmission for this quarter and whether there were additional investments?

Secondly, regarding the investments needed for scaling the autonomous driving (AV) business over the next 12-18 months, will this impact the profitability of the ride business?

A: Adjusted EBITDA for the third quarter grew by 33% year-over-year, with a margin of 4.5% of total bookings (GB), an increase of about 40 basis points year-over-year, reaching a historical high. The fourth quarter outlook is consistent with this. We are precisely tracking the three-year framework target given in February 2024: mid-teens growth in total bookings, with a compound annual growth rate of EBITDA in the high 30% to 40% range.

We are proud of achieving large-scale profit growth, with both the ride and delivery businesses accelerating growth. We have deliberately slowed the pace of margin expansion, now asking investors to measure us by dollar growth in total profit, and we are committed to achieving annual profit growth in the foreseeable years. But we will balance investments between the two product lines and quarterly because we have many exciting investment opportunities, such as cross-platform, affordability and low-cost products, fresh groceries, and retail.

Regarding autonomous driving (AV) investment: The autonomous driving business is currently not profitable. Any new product we launch will initially be in a loss-making state, with the same model: introduce new products -> invest in building supply -> establish ecosystem liquidity -> demand grows and improves payment willingness.

We adopt a "barbell strategy" in the ride business: one end is basic services like UberX, and the other end is premium products with premium profits like Uber for Business, Black, and Reserve. We use the profits generated by these premium products to invest in new growth areas (such as previous taxi, two-wheeler/three-wheeler business).

Autonomous driving is the same, losing money while building the supply base, and is expected to remain unprofitable for the next few years. But as business volume grows, we can improve its profit margin. We can balance the overall profit margin of the ride business through the "barbell strategy," using the profits of premium products to feed investment products like AV.

Additionally, we will also use the balance sheet to invest in the AV ecosystem, establish a global fleet network to maintain vehicles, and collaborate with NVIDIA to invest in AV data collection, collecting ride-hailing specific data to provide to partners. This is a model we have successfully practiced multiple times, and we expect to execute this strategy again in the AV business.

Q: Regarding the investment theme, you mentioned some short-term investments for gaining loyalty in the letter. Can you help us understand more about the loyalty benefits brought by Uber One?

Additionally, regarding the acceleration of U.S. ride volume, please discuss the impact of insurance rates and the benefits from new users (such as seniors and teenagers).

A: We are very satisfied with the progress of Uber One. Membership has reached 36 million, and continues to grow healthily. This membership program contributes about two-thirds of our delivery business's total bookings, and its penetration rate in the ride business is also continuously improving. The benefits it offers are the best in the industry. For us, the good news is that despite the rapid growth in membership numbers, the retention rate of the membership group is actually still improving, especially when we transition more users from monthly to annual subscriptions.

At the same time, we continue to expand the program geographically, now covering 42 countries (compared to 28 a year ago). In the initial months of membership, it is a negative contribution to profit, because the discounts offered exceed the incremental usage or the value of retaining them. But when the membership matures (over 6 months), it becomes profitable.

Regarding the insurance strategy, 2025 has been a year of good progress for us. Our insurance strategy consists of three elements, all of which have contributed this year:

Legislatively, we have achieved victories in multiple states, particularly in California, successfully reducing the coverage limit for uninsured/underinsured motorist insurance applicable to us from $1 million to $60,000 per person, $300,000 per accident, which is very beneficial for us.

Technologically, we have launched the "Driver Insight Dashboard," providing drivers with feedback on driving behavior (such as rapid acceleration, hard braking, sharp turns). Through this feedback, we have seen drivers spontaneously improve their driving behavior, and the proportion of mileage driven by high-score driver groups has increased. Additionally, we have launched "Advantage Mode" in some cities, rewarding drivers who improve their driving scores.

Commercially, through professional team negotiations, maintaining stable partnerships, and exerting influence on risk-taking and profit-sharing to control costs. The result of these efforts is that we expect to save hundreds of millions of dollars in expenses, and plan to pass these savings to customers next year by lowering the unit price nationwide.

Q: Regarding the new diversified gig plan, what new work areas are you exploring? How does this impact driver retention and company profitability?

A: We are positioning Uber as a "work platform," not just a logistics platform. Through Uber AI solutions, we are providing AI-related job opportunities globally, such as training AI models, annotating audio and video data, and evaluating AI responses. These jobs support multiple languages, and some high-barrier positions (such as requiring a PhD in physics) offer higher pay. This provides additional income opportunities for existing and new earners and can enhance their engagement and retention on the platform. Uber AI solutions have already gained many customers, and although the business is still in its early stages, we believe its potential is huge and may become another profitable business line in the future.

Q: Firstly, can you discuss the key capabilities provided by the collaboration with Toast? How does it fit into your broader framework of leveraging partnerships to drive growth and profitability?

Can you discuss the reasons for adjusting some non-GAAP metrics? Specifically, the shift from adjusted EBITDA to adjusted operating profit, does this reflect increased capital investment in the autonomous driving field?

A: Regarding the collaboration with Toast, we are very pleased with it. Toast is a leading POS system provider in the industry and is our strategic partner. The key to this collaboration is that restaurants using the Toast system will be able to automatically activate Uber Eats services. The integration will be seamless, with menu and image uploads automatically completed. This greatly simplifies the setup and operation process for restaurants on Eats, saving our restaurant ecosystem partners a lot of time.

This will bring stronger control, greater flexibility to restaurants, and support them to go online immediately. At the same time, we also hope to leverage our more mature market network outside the U.S. to help Toast expand its international business. We believe this is a win-win collaboration: it expands our merchant ecosystem and helps Toast achieve international growth.

Regarding the adjustment of non-GAAP metrics, the reason reflects that as the company grows in scale, size, and maturity, we want to provide investors with metrics that can better compare with their investment choices. This also reflects our management team's recognition that depreciation, software amortization, equity incentives, etc., are real costs of operating the enterprise, and also reflects the benefits brought by returning cash to shareholders through share repurchases, thereby reducing the number of shares.

This is essentially a journey that all companies experience as they expand in scale and increase their proportion in investor portfolios.

Additionally, as CFO, I personally prefer to hold business line leaders accountable for adjusted operating profit because their decisions on talent and office locations will affect real costs such as equity compensation and depreciation. Including this in their considerations helps them drive revenue growth while also remembering our commitment to continuously improving profitability. This is just the appropriate evolution as the company reaches its current stage.

Q: Firstly, is the strong ride performance mainly due to relief from insurance pressure or other network improvements? Secondly, regarding autonomous driving, what is the scale and quality of real-world data you can contribute?

A: The growth in ride performance is primarily led by trip volume, which is the healthiest form of growth. Our monthly active ride users reached a historical high of nearly 150 million, with strong growth in usage frequency. This is thanks to our "barbell strategy": the core UberX business accounts for about two-thirds of ride volume, while our innovations in low-cost products (such as Moto, wait and save, carpool) help us attract a broader user base;

At the same time, investments in premium products (such as Comfort, Black, and the upcoming Elite) balance overall profitability. Growth in international markets (Latin America, Asia-Pacific) and European summer travel also boosted third-quarter performance. We believe this is not a one-time phenomenon and are confident in continued growth in 2026 and beyond.

Regarding autonomous driving data, we are leveraging the existing ride-hailing network to collect real-world road data, especially in pick-up and drop-off hotspots (such as stadiums, airports). The collaboration with NVIDIA will help us expand this effort, for example, by building a stronger sensor suite to obtain higher quality camera and lidar data. We are not limited by scale and can deploy sensors on partner fleets or independent driver vehicles. Combined with NVIDIA's Hyperion hardware platform, our multiple partnerships, and strong simulation capabilities, we believe this can bring autonomous driving technology to market the fastest and operate on the Uber platform.

Risk Disclosure and Statement of this Article:Dolphin Research Disclaimer and General Disclosure