Uber Technologies
Growth is attractive.
Capital allocation is the test.

01 / The central investment debate

Can Uber convert growing network activity into more earnings per share without committing away the resulting cash?

The positive case: bookings and trips are growing, established infrastructure can support additional transactions, and cross-selling mobility, delivery and advertising can improve profitability. Q2 bookings were $58.0bn, up 22% in constant currency; trips rose 18%; GAAP operating income reached $1.89bn.[S1]

The counter-case: customers and suppliers can use competing apps, autonomous fleets could demand more of the economics, and acquisition spending can consume the cash that otherwise supports buybacks. Scale is useful, but it is not a contractual right to high margins.

The acquisition is neither free upside nor already completed

The announced offer is €41.50 per Delivery Hero share, implying $14.8bn for the full equity; that headline is not the exact cash still payable by Uber. Existing holdings, derivatives, tender participation and a separate sale of 14 markets complicate the bridge. The offer period ends 5 November 2026, with completion expected in H2 2027, subject to conditions.[S6][S7]

Management targets high-single-digit percentage EPS accretion by year three. Our base case credits only 5% incremental earnings by 2030, net of financing, tax and minority interests. That is a modelling choice, not a full merger pro forma. We do not value all of Delivery Hero’s revenue as Uber-owned economics or deduct the same financing cost twice.

02 / Competitive position

Uber’s potential advantage is the combination of demand, dispatch density and a multi-product customer relationship—not exclusive ownership of autonomous-driving technology. Its 2025 filing reported 46m Uber One members, but also describes fragmented, competitive markets. Our conclusion is a meaningful distribution advantage, not an impregnable moat.[S5]

Issuer-reported quarterly figures; bookings and order-value definitions differ. Adjusted EBITDA excludes stock compensation and is not operating profit or shareholder cash. The workbook also deducts a compensation proxy, with differences in payroll treatment identified.[S1][S10][S11]

Why prefer Uber’s business mix? Mobility provides an established profit base while Delivery offers further margin potential. Q2 segment operating income was $2.22bn for Mobility and $1.06bn for Delivery, before shared corporate costs. Freight was slightly loss-making. The breadth is useful; it does not prove Uber is cheaper than every competitor.[S1]

Autonomy is a two-sided risk. A distribution marketplace may help fleets improve utilisation. Conversely, fleets with their own customer demand may bypass Uber or compress its fees. September’s Wayve launch in London was supervised; it is not evidence that large-scale driverless economics have already been demonstrated.[S9]

03 / Earnings are not distributable cash

We use normalised earnings that retain stock-based compensation, rather than treating adjusted EBITDA as owner profit. Nonetheless, Uber’s adjustments exclude some legal, restructuring and acquisition-related items that can require cash. Those costs cannot simply be assumed away.[S1]

LTM = FY2025 minus H1 2025 plus H1 2026. The final row is our proxy, not an issuer measure or a claim that this entire amount is distributable.[S3][S4]

Why trailing GAAP P/E misleads: FY2025 earnings included a $5.0bn tax-valuation release; quarterly investment revaluations also move reported profit. Our $3.30 full-year EPS estimate instead combines Q1 $0.72, Q2 $0.81, the Q3 guidance midpoint of $0.86 and an assumed Q4 $0.91. It is not company annual guidance.[S2][S3]

Cash needs a second check. H1 operating cash flow included $830m from insurance-reserve growth and a $771m deferred-tax addback. These are timing effects, not proof of fraud or worthless cash—but they weaken an assumption that recent conversion persists indefinitely. H1 repurchases of $3.53bn also competed with acquisitions and investment spending.[S4]

Yields use a $145.2bn diluted equity-value proxy. They are alternative lenses, not additive returns. The EPS scenarios already include net share retirement; adding a separate buyback yield would double count it. No dividends are assumed.

At 30 June, financial debt was $12.72bn versus $5.39bn of unrestricted cash and short-term investments. Restricted insurance assets are not surplus cash. September borrowing initially increases both debt and cash; we do not bolt the new notes onto June net debt while ignoring the proceeds or later spending.[S4][S8]

04 / Our five-year operating scenarios

The model starts with approximately $233bn of 2026 bookings and $6.77bn of normalised earnings. It forecasts core bookings and a net-income-to-bookings margin, then applies a deal contribution and dilution-adjusted share count. This avoids confusing revenue accounting changes with weaker demand.

¹ Normalised core net income divided by bookings, not operating margin or revenue margin. ² Cumulative reduction in model diluted shares, after employee dilution. Deal earnings in 2031 are −8%, +5% and +10% of core earnings in bear/base/bull, respectively, after incremental financing and other ownership costs. No probabilities are assigned.

Base is not a low bar. Core bookings nearly double; net profitability improves; financing and integration are manageable; and repurchases resume. The bull case requires stronger growth and a premium valuation. The bear case represents margin erosion despite activity growth; its approximately 39% capital loss is not a worst-case floor.

05 / The base-case path

Core booking growth fades from 16% in 2027 to 12% in 2031. We assume no net share retirement in 2027, followed by 1%, 1.5%, 2% and 2%. The workbook tests whether that retirement programme fits economic cash generation after illustrative strategic-investment, acquisition-cash and debt-reduction reservations.

This is a feasibility check—not proof that the takeover is fully funded. Tender participation, target debt, derivative settlement and integration spending remain uncertain. Unallocated model cash is not added as a second terminal asset.

Valuation cross-check: the market is not obviously giving away the core business

A fixed-share equity cash-flow model starts at $7.0bn, below the $8.18bn historical FCF-less-SBC proxy. It assumes 14% cash growth for five years, 8% for the next five, then 3% perpetually. No separate buyback benefit is added.

This is an existing-business cross-check, not a complete post-acquisition DCF: takeover consideration and incremental acquired cash flows are not consolidated. Existing net debt is not deducted again from equity cash flows; no extra investment-portfolio value is added. Around half of the central value comes from the terminal period.

Nearer-term illustration: our 2027 EPS of $3.92 at 20–24× implies roughly $78–$94. At 22×, the figure is $86, approximately 21.8% above the reference price. This is a valuation sensitivity, not a prediction about when the shares will trade there.

The accessible consensus snapshot was last updated 10 August: 2026 EPS $3.31 and 2027 EPS $4.40. Our 2027 assumption is lower. We do not present that older secondary snapshot as current September consensus.[S13]

06 / Back-solving the 20% hurdle

The base value of $173.56 is just below that hurdle. With the starting P/E unchanged, its return falls to 19.0%; the remaining improvement to 19.6% comes from a modest multiple increase.

Exact base-case entry: $173.56 ÷ 1.20⁵ = $69.75. That is arithmetic, not a margin of safety. An 8% discount to that threshold gives approximately $64, which supports our preferred $63–$65 region while the thesis remains intact.

The investment is sensitive to valuation discipline: at an 18× terminal multiple, the same base EPS produces approximately $142, only 14.9% annually. Earnings can grow substantially while the stock still misses your objective.

For an Australian investor, these USD scenarios are not AUD return forecasts. Australian-dollar strength would reduce the translated return; weakness would increase it. Taxes, fees and exchange-rate movements are excluded.

07 / What would confirm or break the thesis?

Monitoring thresholds are our judgments, not management guidance. An isolated weak quarter is not automatically thesis failure. The underlying filing also identifies legal, worker-classification, insurance and regulatory exposures that can alter costs or business structure.[S5]

Next evidence: Q3 performance against EPS guidance of $0.84–$0.88; the 5 November tender deadline; financing details and the eventual post-deal earnings framework. A September restructuring was reported to involve roughly 3,300 jobs; we assign no automatic margin windfall, since savings may be reinvested and cash charges are uncertain.[S1][S7][S14]

08 / Conclusion and limitations

Uber remains a credible five-year compounding candidate. It is not a demonstrated 20%-return bargain at any price.

At $70.82, our central model delivers about 19.6% annually, but requires real growth, improved profitability and disciplined allocation through a major acquisition. $63–$65 is a more attractive purchase region in this framework; a falling price is not a buying signal when the underlying thesis has weakened.

The valuation cross-check is less emphatic than the earnings scenario: the core DCF’s central value is near today’s reference price. That disagreement is informative. It argues for a buffer and continuing diligence, rather than calling the stock grossly mispriced or treating the optimistic scenario as the expected outcome.

Scope: public-source investment research and an editable scenario model, not personalised portfolio advice, an audited forecast, a full merger purchase-price-allocation model or an assurance of benchmark outperformance. No position size or leverage recommendation is implied. Five-year CAGR is a compounded endpoint return; it does not mean a 20% gain every calendar year. The HTML is a dated snapshot; changes to the workbook do not automatically update this file.