Uber

IThe business

Uber runs a marketplace that connects people with a ride, a meal, a grocery order or a truckload, and keeps a fee on each. In the second quarter of 2026, 208 million consumers used it every month, up 16%, and made 3.9 billion trips, up 18%.⁷ It operates in over 70 countries and more than 15,000 cities.⁵

Uber measures itself in Gross Bookings: what customers spend on the platform before Uber pays drivers, couriers and merchants. In the twelve months to June 2026 they were $215.6B, up 23%.C (calculated)^5781011 Revenue was $55.2B, 25.6% of that, and it changes with Uber’s accounting. Uber recast it in 2020, changed its model in several countries in 2022, and from 2 January 2026 shows UK rides net of driver pay, which cut Q2 2026 revenue by $1.1B while Mobility Segment Operating Income still rose, helped by an $813M fall in driver payments and incentives.⁵⁶²⁹ Since 2016 Gross Bookings have grown 29.2% a year, and operating income has gone from −$3.0B to $5.6B.¹⁸⁵

Three segments are three different businesses. Mobility booked $107.9B and earned $8.1B of Segment Operating Income, 7.5% of its bookings: 69% of the segment profit on 50% of the bookings.C (calculated) Delivery booked $102.2B and earned $3.7B, 3.65% of its bookings. Freight booked $5.5B and lost $0.1B. Corporate G&A and platform R&D took $4.1B, 1.9% of all bookings, which leaves the $7.6B of Non-GAAP Operating Income.⁵⁶⁷⁹¹¹

How it makes money

• Mobility. Riders pay a fare; Uber keeps about 28% of Mobility bookings as revenue.C (calculated) The share fell in 2026 in the UK, where revenue is now net of driver pay. Uber carries the accident insurance for these rides, a cost its results are sensitive to and one I cover under Financials.⁵⁶²⁹
• Delivery. Restaurants, grocery and retail. Uber keeps about 19% of Delivery bookings.C (calculated) Advertising added $568M of revenue in 2025 and $362M in the first half of 2026; Uber does not disclose total advertising revenue.⁵⁶
• Freight. Connects shippers and carriers. It lost $112M over the last twelve months and carries $1.44B of goodwill after the 2021 purchase of Transplace for $2.3B.C (calculated)⁶⁷¹⁵
• Membership. Uber One had 50 million members in the first quarter of 2026, and the company says members drive half of Mobility and Delivery Gross Bookings.⁸

Three things changed in the last year

1. A stricter profit measure. On 12 January 2026 Uber said it would replace Adjusted EBITDA with Non-GAAP Operating Income, which counts depreciation and stock pay as costs.¹¹ For the last twelve months the new measure is $7.6B against $10.0B, 24% lower.C (calculated) I give management credit for it. Both measures still add back legal and regulatory charges: $3.6B from 2018 to 2025.C (calculated)^51213141516
2. A takeover. On 16 July 2026 Uber agreed to acquire Delivery Hero for EUR 41.50 a share in cash: EUR 13.0B for all the shares, $14.8B at Uber’s rate of 1.14 and $14.5B at the Q4 2026 rate of 1.1156.^23244757 Uber puts its all-in price at $13.7B because it bought its stake at prices below the offer.²⁴²⁵ It holds 24.77% of the shares and 11.74% through instruments, about 36.5% by my sum.C (calculated)⁴⁷ Closing is expected in the second half of 2027, and Uber owes a EUR 700M fee if regulators stop it.²³²⁴ The deal is outside the balance sheet and outside my three cases; I size it in Valuation.
3. Robotaxis. Uber says it is live in 7 autonomous-vehicle cities and on track for 15 by year-end, works with more than 25 partners, and plans about $10B of investment and about 120,000 vehicle commitments over several years.²⁹¹⁹ The CEO puts AV trips at hundreds of thousands a week, under 0.5% of Uber’s 300 million; he did not say whether that counts only trips on Uber.²⁹

What else Uber owns

Uber carries $14.0B of stakes at their 30 June values: Delivery Hero $5.0B, Grab $2.0B, Didi $1.9B, Aurora $1.8B and smaller holdings. That is 10.5% of its market value, or $6.87 a share, before any Delivery Hero deal.C (calculated)⁶ The marks have not been updated to Q4 2026. Grab and Aurora, $3.8B of the stakes, closed 27% and 35% lower in Q4 2026 than on 30 June, a gross $1.2B or $0.57 a share, before the offset from the Aurora exchangeable notes that sit in debt.C (calculated)⁶⁵⁸⁵⁹ Uber has since sold 101.4M of its 258.5M Aurora shares in block trades, 72.0M at $6.55 on 17 August and 29.4M at $6.205 on 15 September, about $654M in all. On those sales the gross fall is closer to $1.0B, or $0.47 a share.C (calculated)⁸¹⁸²

Of every $100 Uber books, about $26 is its revenue and about $3.50 is profit after stock pay.C (calculated)

IIThe moat

I think Uber has a local network moat, helped by habit, and that it is widening. The 10-K says the industries it competes in have “low barriers to entry, low switching costs”.⁵ A rider or a driver can open two apps. What holds the network together is density, and density is local.

How the network pays

The network improves as it fills. More riders shorten the wait for a driver, and shorter waits bring more riders. In delivery, more orders keep couriers busy, which keeps delivery times short. The test is whether that reaches profit, and it does. Non-GAAP Operating Income rose from 2.8% of Gross Bookings in Q3 2024 to 3.7% in Q2 2026, while platform costs fell from 2.13% to 1.90% of bookings.C (calculated)¹¹⁷

Habit across products

About one in five eligible consumers, in countries where Uber offers both, is active each month in rides and delivery, and in Q4 2025 those consumers booked over three times as much as single-product consumers.⁵²⁹ The CEO says that group is growing 1.5 times faster than single-product users.²⁹ The three-times figure is Uber’s own comparison of groups that chose differently, so it shows habit, not the effect of cross-selling.

About 58% of first-time Delivery consumers in Q4 2025 were new to Uber, so delivery brings new riders into the network.⁵ Uber One had 46 million members at the end of 2025, up 55%, and 50 million by the first quarter of 2026.¹⁹⁸

Room to grow outside the biggest cities

The CFO says that in the United States fewer than 10% of eligible consumers in sparse markets used Uber in the past year, against over 50% in dense markets, and that the top 20 U.S. cities produce only 30% of U.S. bookings and 25% of U.S. profit.²⁹ These are management statements, not audited figures. They bear on robotaxis: the places where robotaxis operate today are not where most of the profit is.

The moat needs almost no capital to hold. Capital spending was $0.3B in the last twelve months, 0.14% of Gross Bookings, and the robotaxi plan, about $10B over several years, would be its first large call on capital in years.C (calculated)⁶⁷²⁹

Widening or narrowing

The test: who keeps the rider when the car drives itself

Uber says its demand keeps robotaxis busy: partners reach mid to high 20s and low 30s trips per vehicle per day on Uber’s demand.²⁹ That is a company claim I cannot check. If it holds, an operator with thousands of cars needs somebody to fill them, and Uber is a way to do it. The counter is Waymo, which runs its own platform in cities as well as supplying cars to Uber’s, and Tesla, which runs its own robotaxi service.⁶³⁹ Alphabet’s CEO said in April that Waymo had passed 500,000 fully autonomous rides a week.⁵⁰ Where Waymo has its own app, Yipit’s estimate puts it at 15% of rider spending in San Francisco in June, from 16% in January, while Uber’s CEO says its trips there accelerated.⁵¹²⁹ The CEO’s answer is to diversify:

We want to make sure that we’re not dependent on one partner.

My base case gives Uber no moat credit for robotaxis beyond holding its share, and it assumes the AV partnerships, Waymo in Austin and Atlanta among them, stay on Uber’s platform. My three cases do not remove riders from Uber; the robotaxi check is a separate table in Valuation.

The moat is local and real in rides and delivery. The party that can go around it is a supplier that sells straight to the rider.

IIIThe competitive advantage

Uber’s advantage over Lyft and DoorDash is reach. It books across rides, delivery and freight in more than 70 countries on one app and one membership. Its lead is large in rides, narrower in delivery and untested in robotaxis.⁵⁷

• Rides. Uber’s Mobility Gross Bookings were $29.0B in Q2 2026, 5.3 times Lyft’s total of $5.5B.C (calculated)⁷³⁴ Lyft’s bookings grew 23% and its rides 12%, against 22% for Uber’s Mobility Gross Bookings; Lyft bought Freenow on 31 July 2025 and TBR in October 2025 and does not disclose their contribution, so neither Lyft figure is organic.C (calculated)⁷³⁴⁵² Lyft’s Adjusted EBITDA of 3.2% of Gross Bookings is 1.8% after stock pay and the payroll tax on it, and Uber earned 3.7% as Non-GAAP Operating Income after stock pay and depreciation.C (calculated)⁷³⁴
• Delivery. DoorDash is the bigger delivery business by volume: $33.1B of Marketplace GOV in Q2 2026 against Uber Delivery’s $27.5B. It grew faster, 36% or 23% without Deliveroo, against 26% for Uber and 25% in constant currency.³⁵⁷ Uber’s CFO says its organic Delivery growth is accelerating as it laps the Trendyol GO purchase.²⁹ DoorDash earned Adjusted EBITDA of 2.8% of GOV and free cash flow of 2.2%.C (calculated)³⁵
• What DoorDash does not have. Rides. Uber’s cross-use between rides and delivery is its answer to DoorDash’s scale in food, and Delivery Hero would roughly double the markets where Uber offers both.²⁹
• Brazil. Didi Food and Meituan have entered food delivery and raised the cost of two-wheeler supply, which also serves Mobility. Uber says it held share and moved incentives to Delivery.²⁹

Where Uber leads

Liquidity and reach come first: 208 million consumers a month against Lyft’s 30.5 million active riders, in more than 15,000 cities.⁷³⁴⁵ Next is one membership across products, with 50 million members. The third is profit conversion: every dollar booked left 3.7 cents of Non-GAAP Operating Income after stock pay in Q2 2026, and the figure has risen in six of the last seven quarters.C (calculated)⁷⁸¹¹

Where it trails

DoorDash is larger and faster in U.S. delivery. Waymo and Tesla run their own robotaxi services. And Uber does not own the car or the driver’s loyalty: drivers are contractors who can work for a rival, and the 10-K lists switching costs as low.⁵ The table shows what robotaxis are today against Uber’s volume.

IVManagement

Dara Khosrowshahi has been chief executive since September 2017. He ran Expedia from 2005 to 2017.¹⁹ Under him Gross Bookings rose from $34.4B in 2017 to $193.5B in 2025, operating income went from a $4.1B loss to a $5.6B profit, and free cash flow from −$2.2B to $9.8B.C (calculated)¹⁷⁵⁴ The filings do not split that growth into organic and bought. Mobility, which has had fewer purchases than Delivery, grew from $49.7B in 2019 to $97.5B in 2025, 11.9% a year, the closest organic proxy the filings allow; Delivery’s growth includes Postmates, Drizly, Cornershop and Trendyol GO.C (calculated)¹⁶⁵

Shareholders have done less well than the business. Uber priced its IPO at $45.00 and the offering closed on 14 May 2019. The close in Q4 2026 is $65.74, up 46% on the IPO price, or 5.2% a year; Uber has never paid a dividend, so that is the whole return.C (calculated)¹⁷⁵⁵⁵ The proxy’s pay-versus-performance table, a total-return measure, puts $100 invested at the end of 2020 at $160.22 at the end of 2025, against $258.38 for the S&P 500 Information Technology index.¹⁹

Ronald Sugar, 77, has been the independent chair since July 2018. Khosrowshahi is the only director the board does not call independent among ten. Directors are elected every year by majority vote, and there is one class of stock.¹⁹ Say-on-pay passed with 93.8% on 4 May 2026.²²

The finance chief changed in February: Balaji Krishnamurthy, 41, a former Goldman Sachs equity analyst who ran investor relations at Uber, replaced Prashanth Mahendra-Rajah, who had held the job since November 2023. The Chief People Officer left in May.²⁰²¹¹⁹ On 2 September Khosrowshahi told staff in a memo, as reported in the press, that Uber would cut about 3,300 jobs to remove management layers, and that the savings would be reinvested.⁷⁵⁷⁶ Reports call that about 10%. The 10-Q counted about 36,600 employees at 30 June, which makes it about 9%.C (calculated)⁶ I found no 8-K and no figure for the cost or the savings. Uber told Al Jazeera the same week that it had left Nigeria and Uganda.⁷⁶

Turqi Alnowaiser, a deputy governor of Saudi Arabia’s Public Investment Fund, is attributed the fund’s 3.57% stake, the largest among directors, and the board lists him as independent. He chairs Lucid, whose vehicles Uber has committed to purchase: the 10-K records a minimum of 20,000, and Lucid’s filing puts it at 35,000 since April 2026, when Uber agreed to invest another $200M in Lucid and a fund affiliate $550M. The proxy, dated before that deal, lists no related-person transaction with Lucid or the fund.¹⁹⁵⁷⁷

Capital allocation

I grade capital allocation on the deals and on the share count. The exits worked: Uber sold Uber China, Southeast Asia and its autonomous unit for stock that it carries today at $5.7B.C (calculated) Freight did not: Uber paid $2.3B for Transplace and the segment lost $112M over the last twelve months. Postmates and Drizly built out Delivery, and Delivery Hero is unresolved.⁶¹⁵

Since 2024 Uber has spent $11.2B on buybacks at an average of $77.56 a share, 9.8% below my base value of $86 and 18% above the current price; the filings round each amount to $0.1B, so the average is good to about $1.C (calculated)⁵⁶ Net of issuance the share count fell 1.4% for that money, from 2,071.1M at the end of 2023 to 2,042.6M on 31 July 2026; it fell 3.1% from the 2,108.0M peak at the end of 2024. Buybacks retired 144.4M shares, so about 116M were issued, 45M a year, and the money was 8% of today’s market value.C (calculated)⁴⁵⁶ The $20B authorization has $15.7B left.⁶ The CFO says the aim is to put about half of free cash flow into buybacks, and that they will rebuild after the roughly $4B Uber put into Delivery Hero stock in the second quarter.²⁹

Incentives

The CEO’s 2025 pay was $35.6M, 96% of target pay variable.¹⁹ Under the 2025 plan the annual bonus paid 60% on Gross Bookings, Adjusted EBITDA and Adjusted EBITDA less stock pay, and 40% on five strategic priorities; the result was 129% of target. Performance share units paid 40% on Adjusted EBITDA as a share of Gross Bookings, 40% on Gross Bookings growth and 20% on strategic goals, then moved by 0.7 to 1.3 times with relative shareholder return.¹⁹ The plan bit: the 2025 margin tranche paid nothing (4.51% against a 4.65% target), while Gross Bookings growth averaged 18.8% against a 15% target.¹⁹

For 2026 the bonus pays 30% on Non-GAAP EPS and 30% on Gross Bookings, and the performance share units pay 45% on Non-GAAP Operating Income growth, 45% on Gross Bookings growth and 10% on safety.¹⁹ Non-GAAP EPS is per share, and both profit measures count stock pay, so the gap I would have flagged is closed.¹¹ Nothing pays on free cash flow or return on capital. Khosrowshahi held 1,367,100 shares directly after his 10 September purchase, worth $89.9M at $65.74, about 2.5 years of his 2025 pay. The proxy’s count on 2 March, with options and awards that vest within 60 days, was 2,380,203, 0.12% of the company.C (calculated)¹⁹⁶³

Bookings growth is 30% of the 2026 bonus and 45% of the performance shares, and nothing pays on return on capital. So on bookings, bought growth scores like built growth. Their own money is in: in September Khosrowshahi paid $10.0M for 141,000 shares at about $71 and Macdonald $5.3M at about $76, both above today’s price.¹⁹⁶³⁶⁴ Freight, bought for $2.3B, lost $112M in the last twelve months.C (calculated)¹⁵⁶⁷ Since 2024, $11.2B of repurchases retired 144.4M shares while about 116M were issued.C (calculated)⁵⁶

Management took Uber from losses to cash and adopted a stricter profit measure. On capital allocation the exits were good, Freight was poor, the buybacks were priced below my base value but barely moved the share count, and Delivery Hero is unresolved.

VFinancials

Uber has gone from burning cash to earning it. Free cash flow was −$2.2B in 2017 and $10.1B in the last twelve months.⁴⁵⁶⁷ Each headline measure flatters in a different way. Below I count the same year eight ways and value the strictest.

Eight ways to count what Uber earned

The spread between the most generous count, $10.1B, and the strictest, $4.7B, is the whole debate about Uber’s cash flow. Stock pay is $1.9B, the build in insurance reserves $2.0B, and taxes at a normal rate instead of $458M of cash taxes $1.0B more; net interest and depreciation above capital spending make up the remaining $0.5B.C (calculated)⁵⁶⁷

What an owner keeps

I start from normalized owner cash flow of $4.7B: operating cash flow before the insurance reserve build, less capital spending, less stock pay, with tax at 24% in place of the cash taxes Uber paid.C (calculated)⁵⁶⁷ The table reaches it from both ends.

Cash costs outside Non-GAAP Operating Income were $1.4B, 0.67% of Gross Bookings, in the last twelve months: $0.66B of legal, regulatory and other charges that the measure adds back, and a remainder of $0.78B that the filings do not split. The cash flow statement shows where much of it went: receivables grew $0.63B and prepaid and other assets, the line that holds UK VAT paid under protest, grew $0.66B, while payables and accrued liabilities gave back $0.80B.C (calculated)⁵⁶ Before interest and tax, Adjusted EBITDA did not turn into operating cash by $1.3B to $2.7B a year from 2022 to 2025, and by $1.6B in the last twelve months; interest income and cash taxes take that to the $1.4B above.C (calculated)^56121314 I model the cost falling from 0.67% to 0.50% of bookings in my base case.

Uber paid $458M of cash tax in the last twelve months. At the 24% rate it uses for its non-GAAP earnings it would have paid $1.5B.C (calculated)⁵⁶⁷ The difference is a $10.2B deferred tax asset, backed by $4.1B of unlimited federal operating losses, $7.0B of state losses, $1.2B of federal research credits, capitalized research and fixed-asset timing.⁵⁶ I tax every year at 24% and value the asset on its own: $7.9B, or $3.85 a share, on the assumption that it is used evenly over six years.E (estimate) The 10-K reports $5.6B of gross unrecognized tax benefits at the end of 2025, $5.1B of which would affect the tax rate, and does not say how much is netted against the asset. I show the asset gross; each $1B off it costs $0.38 a share.E (estimate)⁵

The insurance float

Reserves are a liability on the scale of the debt: $13.3B at 30 June 2026, up from $7.0B at the end of 2023, against $12.6B of debt principal.⁵⁶ Uber carries the accident risk on its rides. In 2025 it added $4.9B to reserves and paid out $2.4B of claims. Additions were 2.5% of Gross Bookings, against Non-GAAP Operating Income of 3.3%.C (calculated)⁵ A move of 0.2 points in that ratio changes profit by about 6%.C (calculated) Mobility insurance expense rose $851M in 2025.⁵

Reserves have held so far: prior-year development was +$158M, −$78M and −$21M in 2023, 2024 and 2025.⁵

The 2025 build in reserves was $2.7B, 26% of operating cash flow. Of that, $2.5B is additions less claims paid and $0.2B is reserve covered by third-party insurance with a matching recoverable; the cash flow statement shows $2.66B. The last-twelve-months build, $2.0B, is the figure in the ladder above. The cash behind the build goes into restricted investments that the company’s free cash flow definition does not deduct.C (calculated)⁵⁶

Restricted assets were $1.5B below reserves at 30 June, an upper bound on the funded gap, and I treat that gap as debt.C (calculated)⁶ California’s SB 371, signed on 3 October 2025, cut the required uninsured-motorist cover for rideshare from $1M to $60,000 per person and $300,000 per accident. The CFO calls insurance a tailwind this year and says Uber is reinvesting the savings in prices, especially in California.⁴⁴²⁹ I treat the reserve build as neither a source nor a use of cash and give no credit for the income on the float.

The balance sheet

Uber carried $12.7B of debt at 30 June 2026 and $5.4B of cash and short-term investments, so net debt before the stakes was $7.3B.C (calculated)⁶ Principal was $12.6B; the difference is mostly the Aurora exchangeable notes, carried at their $1.3B fair value against $1.15B of principal. The $12.7B is the debt in the enterprise value and in the model. Since the quarter end Uber sold EUR 4.5B of senior notes at coupons of 3.75% to 5.25% for general corporate purposes and signed a EUR 4.0B term loan, which cut the EUR 14.2B bridge for Delivery Hero to about EUR 10.2B.²⁶²⁷ The company says gross leverage will stay under 2 times and that it will keep its investment-grade rating; the revolver covenant needs Adjusted EBITDA of at least 3 times interest.²⁴²⁵²⁶

If the whole EUR 14.2B of committed financing were drawn as new debt, gross debt would be $28.6B, 2.9 times the last twelve months’ Adjusted EBITDA of $10.0B, or 2.6 times with Delivery Hero’s $1.1B added. Under 2 times needs $14.3B of Adjusted EBITDA or less debt. Uber also has cash, and its Q3 guide of $2.86B to $2.96B a quarter is about $11.6B a year, so I cannot reproduce the under-2-times claim from filed numbers.C (calculated)⁵⁶⁷²⁶

Accounting to watch

1. The profit measure changed. The new one counts stock pay and depreciation; both still exclude legal and regulatory charges, which were $548M in the last twelve months.⁷¹¹ Adjusted EBITDA is shown during the transition but is no longer a key measure, and the 2026 pay plan uses the new measures.⁸¹⁹
2. Net income. GAAP net income of $9.6B in the last twelve months includes tax valuation-allowance releases, $6.4B in 2024 and $5.0B in 2025, and swings in the value of equity stakes: +$1.8B in 2024, −$1.5B in Q1 2026 and +$1.6B in Q2 2026, before tax.⁵⁹⁸⁷ Without the releases net income was $3.5B in 2024 and $5.1B in 2025.C (calculated) I use non-GAAP net income, taxed at 24%, which was $6.0B over the last four quarters.⁷⁸⁹¹¹
3. Revenue. It changes with Uber’s accounting; I use Gross Bookings.
4. UK VAT. HMRC assessed about $1.8B on VAT for March 2022 to September 2024. Uber paid under protest and carries it as a receivable, and expects more assessments for 2023 to 2025. I count it neither as an asset nor as a loss; as an asset it would add up to $0.88 a share.C (calculated)⁶
5. Stake marks. The $14.0B of stakes is at 30 June values; Didi is valued at an over-the-counter price.⁶

VIValuation

First, how this goes badly wrong

I asked first what would most likely hurt an owner, and worked back. Three things can.

• Operators sell to the rider themselves. If they take 29% of Mobility bookings, my base falls to today’s price. If they take half, it falls to $50.E (estimate)⁵⁶⁷²⁹
• Driver rules and accident claims. Accident costs accrued were 2.5% of bookings in 2025, against Non-GAAP Operating Income of 3.3%.⁵⁶ Each 0.2 points of bookings in extra cash cost takes $3.94 off my base.E (estimate)⁵⁶⁷
• Capital in the wrong places. Delivery Hero costs $14.8B and lost EUR 359M in the first half.²³⁴⁷ A 20% overpayment is $1.86 a share, less than the other two, but management cannot undo it.E (estimate)

Only the second is inside my bear case. The first and third are separate checks, because I cannot say how likely they are.

What an owner can take out is smaller than what is reported. Free cash flow was $10.1B, before stock pay. After stock pay, the reserve build and a normal 24% tax, it is $4.7B: under half, and 78% of non-GAAP net income of $6.0B. That is 3.5% of the market value, below the 5.00% a ten-year Treasury paid in Q4 2026.C (calculated)⁵⁶⁷³⁰

What I cannot know is the right rate. Uber’s beta has a standard error of 0.25, which puts my base at $67 to $120 before any view on the business.³⁰³¹³² No filing shows what Uber earns on a robotaxi trip. So $86 is the middle of a wide range, not a point.²⁹

What the price implies

Uber’s market value is $134,281M on 2,042.6M shares, and its enterprise value excluding the stakes is $127.6B.C (calculated)⁵⁵⁶ That is 22.3 times non-GAAP net income of $6.0B for the last four quarters, which is $2.95 a share on 2,042.6M shares (Uber’s reported $2.89 is on diluted weighted shares and gives 22.7 times), and 16.8 times Non-GAAP Operating Income.C (calculated)⁷⁸⁹¹¹ The number most people quote is a 7.5% free cash flow yield on $10.1B; stock pay takes it to 6.1%, the reserve build to 4.6%, and a normal tax rate to 3.5%, or 29 times $4.7B.C (calculated)⁵⁶⁷

Run my model backwards. Hold my base margins, platform costs, tax and AV spending, and scale down only Gross Bookings growth. The price then needs Gross Bookings to grow 6.0% a year for ten years, to $386B in the twelve months to June 2036, against 9.4% and $528B in my base case.C (calculated) That path is my base path times 0.63: 11.2% growth in year 1 and 2.9% in year 10. Gross Bookings grew 23% in the last twelve months in reported dollars and 22% in constant currency in Q2 2026, are guided to 18% to 22% in constant currency for Q3 2026, and grew 21% a year from 2021 to 2025.⁵⁷¹⁵ At my margins the market is already pricing a sharp slowdown, and whether it is the right one depends on who keeps the rider when the car drives itself.

Holding my base growth instead, the price needs only 23% of my margin gains: Mobility at 7.7% instead of 8.0% in year 10, Delivery at 4.1% instead of 5.5%, and Non-GAAP Operating Income at 3.9% of Gross Bookings against 3.5% now and 5.1% in my base case.C (calculated) If margins never rise from today’s levels, my base growth is worth $60 a share, 9% below the price.C (calculated)⁵⁶⁷¹¹

The price pays for Gross Bookings growth that falls to 11.2% next year and 2.9% by year 10 at my margins, or for margins that rise only 23% of the way to mine.

The base-case inputs are my estimates; each needed value comes from solving my model for the price with every other input unchanged.

How the model works

The model runs ten years from the twelve months to June 2026. Each year’s cash flow is Gross Bookings by segment times that segment’s profit margin, less platform costs (corporate G&A and platform R&D), less the cash costs outside Uber’s profit measure, less tax at 24%, less AV investment. Non-GAAP Operating Income already counts stock pay and depreciation as costs. I add what the cash flows leave out: cash of $5.4B, stakes of $14.0B and the tax assets of $7.9B. I subtract debt of $12.7B, the $1.5B by which insurance reserves exceed restricted assets, and minorities of $1.1B.⁵⁶⁷⁹¹¹

1. AV investment. The CFO describes the $10B as two kinds of spending over several years: equity in AV software partners, with milestones, and balance-sheet support for fleet operations, real estate and vehicle commitments. Uber is also looking for outside financing, and has not sized the profit-and-loss or cash effect.²⁹ I deduct all $10B, $2B a year for five years, in every case and give no credit for what it buys, except through each case’s Mobility growth and margin. The filings do not size the vehicle commitments: the 10-K gives a minimum of 20,000 Lucid vehicles and says the cash requirement is not determined, Lucid’s filings raise that minimum to 35,000 from April 2026, and Rivian’s release expects 10,000 more. The Q2 10-Q reports no material change in contractual obligations beyond a $1.1B cloud commitment and names neither company.⁵⁶⁷⁸⁷⁹ Uber has put $500M into Lucid and agreed to put up to $1.25B into Rivian, tied to milestones, with $300M paid in May.⁷⁸⁸⁰ At the $70,000 to $80,000 the CEO gives for the Lucid car, 120,000 vehicles would cost $8.4B to $9.6B if Uber bought them all, so $10B is not obviously a ceiling.C (calculated)²⁹ My base is $86 at $10B, $84 at $15B and $82 at $20B; if partners or outside sponsors paid for part of it, my values are low by up to $3.97 a share.E (estimate)
2. The discount rate. I use 10.6%: the 5.29% ten-year Treasury yield against which Damodaran measures his 4.20% implied equity risk premium for 1 October 2026 (the 30 September close), plus Uber’s beta of 1.262 against the S&P 500 since its IPO, from 87 monthly returns, times that premium.C (calculated) The Q4 2026 yield was 5.00%; pairing it with the October premium would mix two dates and give 10.3%. The 60-month beta of 1.26 gives 10.6%, the 24-month beta of 0.72 gives 8.3%, and the Blume-adjusted beta of 1.18 gives 10.2% and $91.^3031332 The beta’s standard error is 0.25, so a 95% range of 0.78 to 1.75 and a rate of 8.5% to 12.6%, which puts my base at $120 to $67. One point lower, 9.6%, takes my base to $100; one point higher, 11.6%, to $75. The rate is a cost of equity applied to flows before interest. A blended rate of 10.0%, with debt at an 8.5% weight and 4.5% before tax (my estimate, inside the 3.75% to 5.25% coupons on the September notes), would give $94. I keep the cost of equity, the higher rate.E (estimate)²⁷
3. Timing and shares. Values are as of Q4 2026. Cash flows arrive through the year, which adds 5.2% to every present value, and the 30 June balance sheet is rolled forward to Q4 2026, which adds 3.2%; cash, debt, stakes, the reserve gap and minorities stay at their 30 June marks. Per-share figures use 2,042.6M shares outstanding. Counting every unvested award, 2,115.8M, lowers the base by $2.98.E (estimate)⁶

Bear, base and bull

My base year 1 sits at the low end of the Q3 guide: Gross Bookings up 17.7% to $254B, against 18% to 22% in constant currency guided for the quarter, which is 17% to 21% in reported dollars. Non-GAAP Operating Income reaches $27.2B in year 10, 3.6 times today’s, and owner cash flow rises from $3.9B in year 1, after $2B of AV spending, to $18.6B in year 10.E (estimate)⁷

For context, Nasdaq lists consensus earnings per share of $3.20 for 2026, $4.56 for 2027 and $5.72 for 2028 on a basis it does not define. On non-GAAP terms my base case earns about $3.59 a share in year 1, $4.29 in year 2 and $5.01 in year 3, so it sits at or below that line.E (estimate)³³

The bear case is margin compression with the rider kept. Gross Bookings still grow 13.7% in year 1 and slow to 3.0% by year 8, Mobility’s margin falls to 5.0%, Non-GAAP Operating Income reaches only $9.4B in year 10 (2.5% of bookings), and cash costs outside the profit measure rise to 0.85% of bookings. Of the $56 fall from my base case (steps rounded; the bridge table has the cents), slower growth is $22, the Mobility margin $17, Delivery and Freight margins $9, platform and other costs $7 and growth after year 10 $2.E (estimate)⁵⁶⁷ Operators taking the rider is the check in the next table, not this case.

The bull case is Uber as the demand layer for robotaxis, with ads and membership compounding. Gross Bookings grow 19.7% in year 1 and 11.1% a year over ten years, to $618B; Non-GAAP Operating Income reaches $42.9B (6.9% of bookings). Of the $52 rise from my base case (steps rounded), growth is $13, the Mobility margin $14, Delivery and Freight margins $13, platform and other costs $6 and growth after year 10 $6.E (estimate)⁷²⁹

The value after year 10 is 61% of my base operating value ($49 of $80 a share), 46% of the bear’s and 66% of the bull’s.C (calculated)³⁰³¹ The three values average to the price if the bear case carries 36% and the base case 64% with nothing on the bull; at one third each they average $85. That is arithmetic from the three values, not a forecast.C (calculated)

Checks outside the three cases

The bear case does not remove riders from Uber, so I test that separately. The CFO says the top 20 U.S. cities are 30% of U.S. bookings and 25% of U.S. profit, so even a heavy loss there is a small part of the total: half of that profit pool, which I size at 8% of Mobility bookings, takes my base to $80. The larger rows test operators winning more cities, and countries, than robotaxis serve today.²⁹⁵

The price is what my base case is worth if operators take 29% of Mobility bookings and platform costs stay where they are. Half of Mobility bookings gone takes the value to $50, 23% below the price.E (estimate)⁵⁶⁷²⁹ These rows are tests, not a rating, and none of them is a fourth case.

What moves the value

The discount rate moves my base by $14 one way and $11 the other for each point. Gross Bookings growth two points a year lower in every year costs $12, and two points higher adds $14. The Mobility margin is the sharpest operating input: $8 for each point lost in year 10, and $23 if it settles at 5.0%, where the bear case puts it. A year-10 Delivery margin one point lower costs $7.E (estimate)⁵⁶⁷

On my base case the price is reached at 12.7% with 3.0% growth after year 10, or at 10.6% if growth after year 10 is −1.8%.C (calculated)³⁰³¹

What a share is made of

Mobility’s cash flow is worth $79 a share and Delivery’s $48; platform costs take $32 and the cash costs outside the profit measure $11. Mobility is 62% of the value of the two businesses against 51% of their bookings, and its $79 alone is close to the whole operating value of $80. That is why the Mobility margin is the sharpest input. At $65.74 the market pays $60 a share for the operations after crediting the stakes, the tax assets and the balance sheet, against my $80.C (calculated)⁵⁶⁷⁹¹¹

The stakes and the tax assets together are $10.72 of the $86, 12%.C (calculated)⁶ The stakes are 30 June marks and the tax asset is my estimate of when Uber uses it; with the asset at nothing, the base falls by $3.85.

Delivery Hero, outside the cases

I leave the takeover out of the three cases and carry Uber’s existing stake inside the $14.0B of stakes at $5.0B.⁶ If Uber pays a fair price, the deal adds nothing to value per share; the table of checks above prices a 20% overpayment and a blocked deal.

The offer is EUR 41.50 a share, EUR 13.0B for all the shares, $14.5B at the Q4 2026 rate and $14.8B at Uber’s rate of 1.14.²⁴⁴⁷⁵⁷ Uber puts its all-in price at $13.7B because it bought its stake below the offer price: about $9.4B is still to pay and about $4.3B is already paid. The $9.4B assumes the swaps need no further cash. If they settle for cash at the offer price, the cash rises by up to $1.5B, $0.74 a share.C (calculated)²⁴²⁵⁶

Delivery Hero’s own accounts show a different company from the deck. In the first half of 2026 it earned Adjusted EBITDA of EUR 427M on GMV of EUR 25.7B, but EBITDA after management adjustments was EUR 79M, stock pay was EUR 157M and the net result was a loss of EUR 359M. Free cash flow before extraordinary items was EUR 348M, helped by working capital that it expects to reverse in the second half.⁴⁷⁴⁸

On the whole company the price is 17.3 times 2025 Adjusted EBITDA of $1.1B. That figure is for the 50 markets Uber keeps while the enterprise value includes the 14 markets sold to SSW; netting the $1.6B sale, the multiple is 15.8 times. The company’s 8 times 2027 includes over $1.2B of synergies that Uber has yet to earn.C (calculated)²⁵ I cannot bridge Delivery Hero’s own 2025 figure, above EUR 900M for the whole group, to Uber’s $1.1B, so I treat the multiple as Uber’s claim rather than a fact I have tested.⁴⁹ A 20% overpayment costs $1.86 a share, and each $1B of value gained or lost is $0.49. The deal moves my value by a few dollars; robotaxis and margins move it by tens.

The financing is the part that touches the cases. Uber is buying the SSW promissory notes for about EUR 1.4B in cash and has committed a shareholder loan for any Delivery Hero debt that falls due on the change of control, convertible bonds among it.⁴⁷ The EUR 4.5B of notes sold on 15 September are not in the 30 June balance sheet, and they add cash and debt together.²⁷ Uber’s own Foodpanda Taiwan purchase was blocked in January 2025 and cost $236M; if regulators stop this deal Uber owes EUR 700M, $0.78B at the Q4 2026 rate and $0.38 a share.⁵²³⁵⁷ The initial acceptance period for Delivery Hero shares runs to 5 November 2026, with two more weeks after it.⁴⁸

The margin of safety

• Price$65.74Close, Q4 2026 [55]
• Base value$86My estimate E (estimate)
• Margin of safety below$6030% under my base value C (calculated)

My rule for a discount to base value is 25% for a business with net cash and no customer above a tenth of revenue, 30% when one cycle or two customers drive results, and 35% when fixed charges such as leases sit ahead of the shares. Uber has no large customer, and it is net cash only if the stakes are counted. I use 30% because two outcomes drive results, whether robotaxi operators keep the rider and what driver rules and accident claims cost, and because about EUR 14B of acquisition financing sits ahead of the shares.⁶²³

The discount covers errors in my base inputs. It does not cover the bear case: at the margin-of-safety price the bear value is 50% lower, and at today’s price 55% lower.C (calculated)⁵⁵⁶

The price is 9% above that line, so the stock would have to fall about 8% to reach it.C (calculated) The lowest close of the past year was $65.27 on 28 October 2026, 0.7% below the price; the high close was $99.72 on 3 November 2025, and the price is 34% below that.C (calculated)³⁵⁵

The price, and how it got here

Uber fell 34% from its $99.72 close on 3 November 2025 to $65.74 in Q4 2026, while the S&P 500 rose 13%.C (calculated)^3535556 The six company events in the table compound to a 10% fall, about a quarter of the 34% fall measured in log terms.C (calculated) I could not confirm what drove the rest, including the two largest falls on days with no event I found, 20 November and 10 December 2025, or the 7% fall from 16 July to 23 July 2026. Six days earlier, when I first priced this page, the close was $66.70.¹

Uber confirmed on 5 October that its Q3 call is on 4 November 2026 at 8:00 a.m. Eastern.⁵⁴ Uber guided Gross Bookings of $58.25B to $60.25B and non-GAAP earnings per share of $0.84 to $0.88.⁷

Where this could be in 5 and 10 years

These are my model’s values at later dates, in dollars per share of today’s 2,042.6M, in each year’s money. The dates are October 2031 and October 2036. Each is the present value at that date, at 10.6%, of the owner cash after it, the unused tax assets and the value after year 10, plus $2.01 a share of other balance sheet items held flat.

Cash is the owner cash after stock pay, tax and AV spending, added up over ten years, with the $10.2B tax asset. The return a year starts from $65.74, counts price growth and that cash, and assumes the price meets my value in October 2036.E (estimate)

For the base to be right at five years, Gross Bookings have to reach $393B in the year to June 2031 and Non-GAAP Operating Income $17.2B, 4.4% of bookings. At ten years it needs $528B and $27.2B, 5.1%. By then the whole value is the value after year 10.E (estimate)

Growth, the Mobility margin and the discount rate separate the cases. Gross Bookings grow 5.6%, 9.4% or 11.1% a year. The Mobility margin ends at 5.0%, 8.0% or 9.5%. The $10B of AV spending and the 10.6% rate are the same in all three; at 9.6% today’s base is $100, at 11.6% $75. In the bear, Uber keeps its riders but its take shrinks, so value barely rises, from $30 to $32.E (estimate)

The likeliest way my base is wrong is the Mobility margin. It is 7.5% now and I take it to 8.0%. If competition and robotaxi partners squeeze it to 5.0%, today’s base falls by $23.E (estimate)

The same dollar elsewhere

My base returns 14.3% a year to 2036 and my bear loses 1.7%.E (estimate) The closest comparable on this site is MercadoLibre, a marketplace with its own delivery network: its base returns 12.9% a year, with its price 11% below base against 24% here. The business and the price are two questions. The business is strong, with one open question, the rider in robotaxis. The price is below my base and does not cover my bear.

My $86 also needs three things together: bookings growth near 18% and then 15%, partners staying on Uber, and a Mobility margin of 8.0%. At 70%, 60% and 65%, my guesses, that is 27%.E (estimate)

These are estimates of value, not forecasts of the share price. The market price can sit above or below value for years.

VIIThe case for the price

What most people miss

Waymo files its California trips with the state utilities commission. Its main deployment program carried about 190,000 trips a week in April and May 2025, 285,000 in Q4 2025 and 325,000 in Q2 2026.C (calculated)⁶⁵ So the count is up about 70% in a year. But growth averaged above 20% a quarter in 2025 and about 7% a quarter in 2026, and June came in below May. The filing does not say why. Riders in California use Waymo’s own app, so this is a test of going around Uber. Across all its cities, Alphabet reported more than 500,000 rides a week in April.⁵⁰

Two top executives paid more than today’s price in September. Dara Khosrowshahi paid $10.0M for 141,000 shares on the 10th, at about $71. Andrew Macdonald, the chief operating officer, paid $5.3M for 70,000 on the 4th, at about $76.C (calculated)⁶³⁶⁴

The bull case

Bill Ackman’s Pershing Square Inc. showed 34.3M shares worth $2.48B in its 13F for 30 June 2026, its largest line at 12.7% of the filing. The adviser’s 13F for 31 March showed 30.0M.⁶⁰⁶¹ In Fortune on 5 August he said “Uber stock is extremely cheap for a business of this quality”. His reason is habit. Riders, he says, want one trusted app and will not scroll through Tesla’s and Waymo’s.⁶² Khosrowshahi’s version, on the 5 August call, is about supply: many AV systems, as in AI, each launching city by city, so filling the cars is valuable.²⁹

Two analysts add numbers, as reported in the press. Nick Jones of BNP Paribas (29 September) sees robotaxi costs falling from near $1 a mile to 16 to 42 cents, and forecasts $106.⁶⁶ Mark Mahaney of Evercore ISI (15 October) has $150, above my bull of $138, on Uber becoming the network that runs autonomous fleets. 51 analysts averaged $100.86.⁶⁷

Brad Gerstner is the check. Altimeter’s 13F fell from 13.3M shares at 30 September 2024 to 5.6M at year-end.⁶⁹⁷⁰ As reported, he told CNBC that November he was moving money into Tesla, though the stock was inexpensive, and later called self-driving “a risk fundamentally to the Uber business model”.⁷³⁷⁴ Altimeter rebuilt to 8.0M by March 2026 and had 7.4M at 30 June.⁷¹⁷² I found nothing from him on why.

Against reality

The price needs less than my base case. Any one of three things pays for it. Bookings could grow 6.0% a year for ten years. Uber could deliver 23% of my margin gains. Or operators could take 29% of Mobility bookings.E (estimate) Today Mobility bookings are growing 22%, Mobility earns 7.6%, and AVs are under 0.5% of trips.⁷²⁹

If bookings grow about 18% to June 2027 and 15% the year after, Non-GAAP Operating Income reaches 3.9% of bookings by mid-2028, and partners stay on the platform, $86 is possible. Today’s 3.7% is on that path. Mobility above 15% through 2027 and Non-GAAP Operating Income at 4.5% of bookings by Q4 2027 would point past $86, toward $138.E (estimate)

My $86 gives robotaxis no credit and no damage beyond the $10B I deduct. Uber can keep every rider and still lose if suppliers take the profit. The Phoenix pilot with Waymo has ended, and press reports say Waymo plans its own app in Austin and Atlanta from January 2028.⁶⁸⁴⁰ The CFO says there will be an effect on profit and has not sized it.²⁹ No filing yet shows what Uber earns on a robotaxi trip.

What they have to hit

Where I come down

The slowdown in Waymo’s California count moved me most, because it comes from a regulator’s filing. Ackman’s habit argument moved me less: Yipit puts Waymo at 15% of rider spending in San Francisco, on trips inside its zones, so riders there do open another app.⁵¹ I would raise my values if Mobility margin passed 8.0% by 2028 with robotaxis a visible share of trips.

VIIIRisks, and what would change my view

The risk Uber names in its own filing that I weight most:

If we fail to offer autonomous vehicle technologies on our platform at competitive scale

The filing goes on to say its financial performance and prospects would be adversely affected.⁵

1. Robotaxi operators. Mobility earns 69% of segment profit.C (calculated) Alphabet said in April 2026 that Waymo had passed 500,000 fully autonomous rides a week, and Waymo raised $16B at a $126B value in February.C (calculated)⁵⁰³⁶³⁸ Tesla reports nearly 2 million cumulative unsupervised miles.³⁹ My bear case prices a loss of margin and growth in Mobility, and the check in Valuation prices a loss of bookings. Press reports of 24 July 2026, citing an Uber spokesperson, say Waymo plans its own app in Austin and Atlanta from January 2028, with the current contract running to May 2028; I found no filing or release from either company, and Uber’s CEO said on 5 August that Uber expects to operate there next year.⁴⁰⁴¹²⁹
2. Driver rules. Classification is contested in courts and legislatures worldwide. California’s attorney general has a suit pending for the years before Proposition 22. New Zealand’s Supreme Court ruled four drivers employees in November 2025. Mexico passed a reclassification law in December 2024. The EU Platform Work Directive creates a rebuttable presumption of employment, and member states must transpose it by 2 December 2026.R (reported)⁵⁶⁴²⁴³ Accrued legal and regulatory liabilities were $1.8B at 30 June 2026, and Uber added back $3.6B of legal and regulatory charges from 2018 to 2025.⁶⁵ Each 0.2 points of bookings in extra cash cost lowers my base by $3.94.E (estimate) A ruling that reclassified drivers in a large market would cost more than any figure I can model.
3. Insurance. The cost accrued for accidents was 2.5% of bookings in 2025, against Non-GAAP Operating Income of 3.3%. The reserves are $13.3B and have held, and California’s rule change helps now.C (calculated)⁵⁶⁴⁴
4. Delivery Hero. Regulators may stop it, it adds debt, and the CFO puts the platform migration in 2029. Delivery Hero lost EUR 359M in the first half, and its full-year cash flow guide of slightly above EUR 250M against EUR 348M in the first half implies a negative second half.²³²⁹⁴⁷
5. Earnings quality. GAAP net income is dominated by tax releases and stake marks, and the tax assets that shield cash taxes run out as profit grows.⁵⁶
6. Competition. DoorDash is larger in U.S. delivery. Didi Food and Meituan are raising supply costs in Brazil. Lyft’s bookings grew 23% in Q2 2026, with Freenow in the base.^35293452
7. Consumer-protection litigation. The FTC sued in April 2025 over Uber One enrollment and cancellation. A press report of 13 April 2026 says the court let most deception claims proceed; I did not read the docket and cannot confirm the later status.⁴⁵

What would move me

The first test comes on 4 November 2026. Uber confirmed on 5 October that its Q3 call is that morning at 8:00 a.m. Eastern; Nasdaq lists the same date for results.⁵⁴² Uber guided Gross Bookings of $58.25B to $60.25B, 18% to 22% growth in constant currency, Adjusted EBITDA of $2.86B to $2.96B and non-GAAP earnings per share of $0.84 to $0.88.⁷ The Delivery Hero acceptance period ends the next day, and the EU deadline follows on 2 December.