UBER

We rate Uber a Buy. Our central estimate of its value is $90 a share against a $71.22 price, and it returns 15.5% a year over five years on our numbers. The price assumes that autonomous fleets (robotaxis) cut Uber's US ride-hailing profit well before 2031. Our central case already charges that threat and still gets $90 (sections 4 and 5).

Terms used below. Every figure is in US dollars. "Gross Bookings" is what customers pay through the platform; "segment operating income" is Uber's own segment profit, after stock pay and depreciation. NOPAT is operating profit after tax. The cost of equity is the yearly return shareholders require, used to bring future cash back to today. Each report values a company three ways: the central case is our best estimate, the bear case a severe but plausible bad outcome, and the bull case the matching good outcome. The exit price is our estimate of the share price on 2031-12-31, taken as the median of three methods. The judged exit multiple is what we think the market will pay for each dollar of 2032 NOPAT. The long fade values the company by letting its growth and its return on new capital fall gradually toward normal levels over 15 years. The outside view applies what comparable companies trade at today. The model is model.py (output in output.txt); what the price assumes is unblinded.py (output in unblinded-output.txt); the blind record is assumptions.md.

1. How Uber makes money, and what limits it

Riders and eaters pay Uber; Uber keeps about 30% of a Mobility booking and 19% of a Delivery booking as revenue (2025, reported) and pays drivers, couriers, insurance and incentives out of it. What is left after all of that, stock pay included, was 7.4% of Mobility bookings and 3.5% of Delivery bookings in 2025, and 7.66% and 3.77% in H1 2026 (reported, segment operating income). Corporate costs take about 2% of bookings. Freight still loses money.

Three things limit it. Drivers must earn enough to show up, so Uber's take cannot rise much. Insurance is the largest cost Uber controls only in part: it self-insures through a captive, so a bad claims year lands on Uber (reserves $13.3B on 2026-06-30). And regulators and courts set the labor and liability rules city by city (accrued legal, regulatory and non-income tax claims $1.84B).

Uber is asset-light: capex was $0.34B in 2025 against $0.75B depreciation (market data cash-flow statement), and working capital gave cash ($2.2B in 2025), mostly because insurance reserves grow with bookings (10-Q Q2 2026 liquidity discussion). Reported free cash flow ($9.8B in 2025) is well above the earnings we value, because it adds back $1.8B of stock pay, which we treat as a cost.

2. Uber has beaten its own guidance in each of the last eight quarters

Management has beaten its own guide every quarter for two years (reported, against each release's outlook):

The last three quarters landed above the top of the Gross Bookings range. Our Q3 2026 forecast is the midpoint of the guide times that median beat: Gross Bookings $60.2B (guide $58.25-60.25B), Adjusted EBITDA $2.97B (guide $2.86-2.96B). Uber reports Q3 on 2026-11-04 (earnings calendar; Uber announced the call date on 2026-10-05). Q2 2026 grew Gross Bookings 24% to $58.0B (22% in constant currency, management); trailing free cash flow passed $10B (management claim, Q2 call).

3. Our forecast values Uber at $90 a share

Each segment has its own path (model result, central, $B):

How the model counts each claim on the company, once:

• Stock pay is a cost inside segment operating income. The share count is held at 2,052.7M (2,042.56M on 2026-07-31 plus options and a 4M allowance for the 2028 convertible). Unvested RSUs are not added, because the expense already pays for them.
• Insurance. Claims cost sits in segment margins. The balance sheet adds one claim: reserves less the restricted cash and investments held against them and reinsurance recoverables, after tax, $0.79B. Prior-year development was +$158M, −$78M and −$21M in 2023-25 (10-K Schedule II), so the central case adds nothing; the bear adds 10% adverse development. The float's growth and the interest on restricted assets are not counted as value: a conservative choice worth roughly $2-3 a share (our estimate, not modelled).
• Legal and regulatory charges run at 0.36% of bookings a year, the reported average of 2024 to H1 2026 (0.69% in 2024, 0.29% in 2025, 0.01% in H1 2026), plus the accrual ($1.84B on 2026-06-30, taken as $1.8B) as a claim at exit.
• Equity stakes (Didi $1.90B, Grab $2.02B, Aurora $1.76B, others $3.35B on 2026-06-30, 10-Q Note 3) at their June value less a 15% tax haircut, carried to exit at the cost of equity so they add today's value only.
• Delivery Hero is valued as a separate 75% stake: $5.8B paid in 2027, 14x its 2032 NOPAT less $4.2B of bridge items, with 60% of the $1.2B synergies (management claim) counted. If the deal fails, value is $89.
• Tax is 26% on operating profit: the recurring lines of the 2025 10-K rate reconciliation (21% federal, 2.7% state, GILTI and FDII, which are US taxes and deductions on foreign income, Brazil withholding, foreign tax credits, R&D credits, stock pay) sum to about 25.8%. Valuation-allowance releases, one-off losses and price-driven excess stock benefits are left out. H1 2026's 27.7% book rate includes deferred US tax on equity-stake gains, which this line does not cover.
• Deferred tax assets (tax benefits already earned that lower future tax bills; $10.16B on 2026-06-30, less $1.77B of accrual-type items that never turn into cash in a growing business: accruals and reserves, legal, stock pay and lease, net of the right-of-use liability, 10-K 2025 tax note) shield up to half of each year's tax.
• Operating cash. Amounts owed to drivers and merchants ($1.77B on 2026-06-30, grown with bookings to $3.2B by 2031) are held back from net cash at exit (added after the blind commit, section 8).

Autonomous vehicles (AVs): which case charges what. The threat is in the central case: Mobility growth fades from 17% to 8%, and the Mobility margin stops at 8.1% from 2028 instead of rising, because Waymo, Tesla and Zoox take part of the dense US metros. The bear charges more: Mobility growth to 4% and margin to 6.3%, with a 12x multiple. The option is in the bull only: Uber becomes the booking layer for many AV fleets, Mobility margin rises to 9.2% and growth stays above 10%. AV cash ($8.0B over 2027-31, our share of the $10B commitment) leaves every case and returns as stakes worth 20%, 50% or 100% of cost.

Value, three ways (central, at 2031-12-31): the judged exit multiple, 18x 2032 NOPAT, gives $158. The long fade (growth 7.7% falling to 3.5% over 15 years, return on new capital 40% falling to 20%, discounted at the 10.40% cost of equity because Uber holds net cash from 2028) gives $136. The outside view, the peer median forward P/E (price divided by next year's expected earnings) of 16.9x applied to 2032 NOPAT plus net cash at face, gives $151. The median of the three, $151, is the exit price. At a 10.40% cost of equity (5.087% 10-year Treasury yield, beta 1.265, and a 4.20% equity risk premium, the extra return investors require for holding stocks, from Damodaran, each dated in data/market-inputs.json) it is worth $90 today.

The 18x is judged from mature marketplaces' forward P/E (BKNG 13.4x, EXPE 11.8x, EBAY 18.3x, LYFT 16.9x, ABNB 27.9x; market data, fiscal 2027, fetched 2026-10-15), a turn above their median because Uber still grows about 8% at exit. The value-driver formula gives 19.3x at 6% growth and a 40% return on new capital; it is a cross-check only. The multiple implies 5.6% growth while the model's path grows 7.7% in 2033; the fade method uses the path and lands at 14.8x. The median of the three is now the outside view. Uber's own history gives no guide: GAAP operating profit only since 2023.

The fade's 40% starting return on new capital is below what the model's own forecast implies: the forecast has no net investment besides AV and Delivery Hero (capex runs below depreciation and working capital is a source of cash, 2023-25), so NOPAT nearly doubles on about $8B of AV spend that is valued separately.

Case-weighted fair value $89 (25/50/25, our weights); weighted five-year return 15.2% a year.

4. The $71.22 price assumes robotaxis hurt Uber early

Holding everything else at our central case, today's $71.22 needs one of: an exit at 12.4x 2032 NOPAT against our 18x; Gross Bookings growth 4.9 points a year lower (12% in 2027 to 3% in 2032); or a Mobility margin of 6.2% from 2028 against 7.66% today. Moving Mobility alone 68% of the way to our bear AV path also gets there. In plain terms the price assumes that autonomous fleets cut Uber's Mobility profit in the US well before 2031.

The price has fallen from a weekly close of $98.51 (week of 2025-09-15) to $71.27 (week of 2026-10-12, still forming), 28%, while Gross Bookings grew 21-25% a year in each of the last four reported quarters (market data weekly closes; releases). It closed at $72.46 the week before the Delivery Hero offer and $65.94 the week after (weeks of 2026-07-13 and 2026-07-20); we did not test what else moved it that week. Capfolio's Alpha model agrees the stock is cheap (valuation B, rank 74 of 709) but ranks it Neutral overall on weak momentum and expectations; Alpha leans on price momentum, and a gap from it is expected.

5. We expect Uber to keep growing; the market prices a robotaxi squeeze

The market prices Uber as a business whose best market is about to be taken by robotaxis. We price it as a network that keeps growing double digits for three more years, then slows to 8%, with its Mobility margin capped at about today's level by AV competition. Our central case already charges the threat; it does not need the AV option to work. The two views are settled by the US Mobility numbers: trips and margin in the cities where Waymo and Tesla run at scale. Uber said US trips and bookings accelerated in Q2, with strong growth in Los Angeles and San Francisco (management claim, Q2 call). If that holds through 2027, the market's view is wrong.

6. The bear case ($35) is a real possibility

The bear ($35) is a real possibility, not a tail. Robotaxis may cost less than half of a human-driven ride (an analyst claim in a 2026-09-29 headline, not checked); if Waymo and Tesla run their own apps in the top US metros, Uber loses the most profitable rides first, and its 8% Mobility margin depends on those rides. Uber is also spending heavily: $14.6B for Delivery Hero (a lower-margin delivery business, close in H2 2027), $2.3B for ezCater, $10B for AVs, with debt rising to fund them. Insurance and personal-injury litigation can surprise (a $40M wrongful-death verdict on 2026-09-18, headline). The Uber executive leading the robotaxi expansion is leaving next month, November 2026 (Dow Jones and Reuters headlines, 2026-10-13). Our 18x exit multiple assumes the market will again pay a growth multiple; it pays 12x on our numbers today.

7. What would confirm or break the call

Catalysts:

• 2026-11-04: Q3 2026 results. Our forecast: Gross Bookings $60.2B, Adjusted EBITDA $2.97B; the Q4 guide against our seasonality.
• First half 2027: Delivery Hero offer document and acceptance level (our central assumes 75%).
• Second half 2027: Delivery Hero close and regulatory decisions.

Falsifiers:

• Q3 2026 (2026-11-04): Mobility segment operating margin at or below 7.4% of bookings (no gain on Q3 2025's 7.42%) cuts central to about $83; at or above 7.9% with a Q4 bookings guide of 17% or more in constant currency keeps $90.
• Mobility margin below 7.0% in any later quarter: toward the bear (7.0% from 2028 is $79).
• US Mobility trips falling year on year in any quarter: the AV threat is arriving faster than the central case.
• Gross Bookings below the low end of the guide in any quarter (never in eight quarters): the guidance record no longer holds.
• Adverse prior-year reserve development above $0.3B in a year: the insurance claim is larger than charged.
• Delivery Hero acceptance below 50% (deal fails): $89, little change.

8. Two rounds of review moved the value from $96 to $90

Round 1: our own checks of the draft

The analyst ran the report checklist (math reruns, figures sourced, the bridge from enterprise value to equity, coherent cases) on the draft. Multiples, dollar values and which method set the exit price in this table are as of the draft (central $95; the draft's model output, commit 21586fded, printed a central fade of 16.4x, a bear fade of $64 and a bear multiple of $62, with the fade as the median). After round 2 below, the bear's multiple ($59) and fade ($57) sit within $2, the central fade gives 14.8x, and the median exit price is the multiple in the bear and the outside view in the central case.

Round 2: two outside reviews (2026-10-15)

An auditor checked the model's inputs and an advocate argued the opposite side. Each effect is measured alone against the $95.34 central before this revision (rerun of model.py on a scratch copy). The accepted changes together give $89.99, not the sum of their parts: with the fade discounted at the cost of equity, the median moves from the fade to the outside view, which the discount rate does not touch.

9. Sources

• SEC EDGAR (CIK 1543151), read 2026-10-15: 10-K 2025 (filed 2026-02-13); 10-Qs Q3 2025, Q1 2026, Q2 2026 (filed 2026-08-05); releases Q2 2024 to Q2 2026 (8-K ex. 99.1); 8-Ks 2026-01-12, 2026-07-16 (Delivery Hero), 2026-08-07, 2026-09-15. Excerpts in data/filing-excerpts.txt.
• Q2 2026 call, Quartr via market data (notes).
• Market data statements, segmentation and peer estimates, fetched 2026-10-15 (data/).
• Market data: US 10-year, EUR/USD and betas (2026-10-15 20:02 UTC); weekly closes, headlines, forecasts and earnings calendar (read after the blind commit).
• Damodaran implied ERP, 2026-10-01.
• Market data: quote (2026-10-15 19:59:59 UTC) and analyst ratings (updated 2026-09-10).
• Alpha: alpha rank UBER --raw --json --allow-stale-model-cache, as of 2026-10-14.