Analysis
A Better Business Does Not Automatically Mean a Better Investment
Uber's investment story is easy to tell: more people use its services, and it earns more from the activity on its platform. The harder question is how much of that improvement reaches shareholders.
Our buying case rests on two things happening together. First, Uber must keep more profit from its rides and deliveries while still offering customers a service they want to use. Second, the cost of running the wider company must grow slowly enough to leave something over.
The first part has recent evidence behind it. The second requires a judgment about what happens next. That distinction is the heart of our rating. We see enough to consider buying near $70, but not enough to call Uber an obvious bargain.
What Uber Has Already Shown
In both of the first two quarters of 2026, rides and delivery produced more operating profit for each dollar of bookings than a year earlier. Bookings measure the value of transactions passing through the platform; they are not the same as Uber's revenue.
In the second quarter, for example, the rides business earned about $7.64 of segment operating profit per $100 of bookings, up from $7.28 a year earlier. Delivery earned about $3.84, up from $3.52. These measures already include the expense of paying employees in shares. They are not the amount left for shareholders after all company costs and taxes. [S01], segment note; [S03]; Evidence: operating evidence
Our measure of segment operating profit after corporate costs adds the results from rides, deliveries and Freight, then subtracts shared corporate costs. Employee stock compensation is already counted as an expense. This is our analytical subtotal, not Uber's reported net income or cash flow. It includes ordinary legal costs within corporate expenses, but not the separate legal-cost allowance used in our valuation model.
This subtotal rose from $1.534 billion to $2.143 billion in the second quarter of 2026 compared with the same quarter a year earlier. The increase was 39.7%, rounded to roughly 40%—growth in profit, not a 40% profit margin. See the Quarterly Dollar Reconciliation in Evidence & Assumptions for the component amounts and calculation. [S01], Note 10; Evidence: operating evidence
But two quarterly comparisons do not establish a lasting trend. They support the possibility of further improvement, not a promise that it will continue.
The Expense Line That Could Spoil the Story
Corporate costs include shared administration and platform research and development. Those costs recently grew about 18%. A company can improve the economics of each ride and delivery and still spend much of the benefit elsewhere.
Our preferred scenario, called “continuation,” assumes further growth in rides and deliveries and modest improvement in how much profit they earn per dollar of bookings. It also assumes that corporate-cost growth slows to 8% a year. These are our assumptions, not Uber's guidance. Management's workforce reduction makes slower spending plausible, but a smaller workforce does not tell us the eventual savings. Rehiring, technology investment and acquisition integration could absorb them. [S45]; Evidence: operating evidence
We do not need costs to land at exactly 8% for buying to make sense in our model. With the other preferred assumptions unchanged, cost growth of 12% still produces an attractive illustration. At 18%, however, the advantage over SPY largely disappears.
| Annual Growth in Corporate Costs | Illustrated Annual Uber Return |
|---|---|
| 8% | About 16% |
| 12% | About 13.5% |
| 18% | About 8.6% |
For comparison, our separate SPY illustration is about 8.7% a year. These are answers to “what if,” not forecasts of the return investors should expect. The Scenario Tests and Strategic Risks section of Evidence & Assumptions explains the operating and valuation assumptions. We have not assigned probabilities to them. Evidence: scenario tests
The Share Price Can Disappoint Even If the Business Improves
Investors also have to decide how much they will pay for Uber's future profits. If they demand a better return for taking the risk, or expect less growth far into the future, the same business can command a lower price.
That is a real weakness in our buying case. Keeping our preferred operating assumptions but applying both tougher valuation assumptions reduces the illustrated return from about 16% to about 8.2%—slightly below our SPY example. Strong growth alone does not protect the investment.
The reverse is possible too. More favorable valuation approaches produce annualized illustrations around 22% to 23%. Those examples help explain the upside, but we do not adopt them as expected returns. Our Uber and SPY calculations also use different measures of profit and different valuation treatments; this is a comparison of conditional investment cases, not a perfectly matched experiment. Evidence: scenario tests
Three Ways the Downside Could Be Worse
The first is a combination of weaker growth and weaker profit margins. Our severe operating examples show annualized losses of about 14% and 27%. Adding disappointing project spending to the harsher case takes the loss to about 31%, uses all the borrowing capacity assumed in that scenario and requires new shares. Issuing shares spreads ownership across more shares and can reduce the value belonging to each existing shareholder. None of these examples sets a limit on possible losses. Evidence: scenario tests
The second is driverless cars. Uber could remain a valuable marketplace connecting passengers with vehicle operators. But developers could reach customers directly or demand a larger share of each fare. More rides would not necessarily mean better economics for Uber. Our tests include Uber accepting less favorable fees, but cannot establish the worst possible outcome. Much of the value we assign to the business still depends on it remaining profitable for many years. [S33], [S35]; Evidence: scenario tests
The third is the pending Delivery Hero acquisition. Our model assumes roughly $15.3 billion of cash is needed at a December 2027 closing. That is an assumed funding calculation, not a confirmed final bill. It is large enough that the purchase deserves attention alongside the rides business: cash must be found before the hoped-for benefits are realized. We assign value to the acquired business, but that value depends on future earnings and the price it could command. Paying for an asset does not guarantee recovering the money. [S21], [S22], [S23], [S37], [S40]; Evidence: scenario tests
Why We Choose 6, Rather Than 5 or 7
The case for 5—Consider other opportunities—is credible. Corporate costs could keep rising near their recent pace. Profit improvement could stall, or investors could pay less for future earnings. Any of these could make SPY the better choice.
We choose 6—Consider buying because Uber has already improved operating profit while investing in its service and absorbing rising corporate expenses. Our preferred case allows slower growth and modest further improvement; it does not require the recent pace to continue unchanged. There is also room for some disappointment in costs before the modeled advantage disappears.
That is a judgment, not something the spreadsheet proves. The evidence for improving operations is stronger than the evidence for the cost slowdown we assume. We would move toward 5 if costs stayed near 18% growth while margins flattened, or if the tougher valuation assumptions became our preferred view.
The case for 7—Buy—is that better operations and a more favorable valuation could produce substantially better returns. We are not choosing it because ordinary combinations of disappointment can already erase the advantage. We would need stronger evidence that operating gains reach shareholders and that the investment remains attractive under reasonable setbacks, assessed at the price then available.
What We Will Hold This Thesis Against
Our corporate-cost forecast is explicit: no more than 12% comparable growth in each of 2027 and 2028. That is a test of our judgment about slowing costs, separate from the 8% assumption in the preferred valuation scenario. Missing it would count as a forecast miss even if better results elsewhere helped the shares.
We also retain our forecast that the rides business earns at least $7.80 of operating profit per $100 of bookings over the year ending September 2029. Neither test alone proves the investment case. We also need to examine cash after employee compensation and ordinary investment, acquisition spending, debt and new shares. Evidence: forecast tests
Acquisitions complicate those comparisons. If Uber does not disclose enough to compare the same business across years, we will call the test unobservable—not quietly record a success or a failure from mismatched figures.
Our conclusion remains selective: near the historical $70 reference price, the operating evidence justifies considering Uber. The thesis fails if improvement in rides and deliveries is repeatedly swallowed by corporate spending, investment needs or weaker economics. A rising share price would not excuse getting that reasoning wrong.
Disclosures & Limitations
Research Cutoff: September 23, 2026 at 2:44 p.m. MDT
The Publisher reports no interests, holdings or partnerships relevant to Uber or this report. Prepared with AI assistance.