Lyft Address

Income statement

Revenue CAGR (FY2021 → FY2025): 18.4%. On a Gross Bookings basis the FY2023 → FY2025 CAGR is 15.7%, essentially in line with the ~15% 2027 target set at the 2024 Investor Day.

Commentary — income statement trends, inflections and drivers

Revenue. The FY2024 revenue jump of 31% substantially overstates underlying growth. It reflects a step-change in accounting and marketplace-health mix — the implied take rate expanded from 31.8% to 35.9% in a single year — layered on 16% Gross Bookings growth. FY2025's 9% headline growth conversely understates underlying momentum for two reasons. First, a $168 million charge from certain legal, tax and regulatory reserve changes and settlements was recorded as a reduction to revenue; excluding it, FY2025 revenue would have been approximately $6.5 billion, or roughly 12% growth. Second, investments in driver supply — also a contra-revenue item — fell by $121.4 million as organic driver supply improved, which is a positive operating development that flatters revenue relative to a year in which supply incentives had risen. The clean read on FY2025 top-line health is the 15% Gross Bookings growth and 14% ride growth, not the 9% revenue line. In H1 2026 revenue growth reaccelerated to 15.0% (H1'26 $3,494.0 million versus H1'25 $3,038.4 million), and to 16% in Q2 alone, as Freenow's gross-basis revenue consolidated for a full period.

Cost of revenue and the insurance problem. Cost of revenue rose 11% in FY2025, outpacing revenue. The entire delta is insurance: a $337.8 million increase driven by higher ride volume compounded by higher cost per mile, plus $29.6 million of higher transaction fees, partly offset by a $35.0 million decrease in restructuring costs. Management's forward guidance is explicit and unusually useful: cost of revenue will keep rising year over year in the near term on insurance renewals, "but we expect total insurance costs will increase at a lower rate than they have historically as a result of a recently passed rideshare insurance reform bill, SB 371, which is expected to reduce our insurance rate in California." California is Lyft's single most important insurance jurisdiction, and SB 371 is arguably the most consequential regulatory development for Lyft's FY2026–FY2027 margin bridge. Confirmation of this arrived in Q2 2026, when cost of revenue fell year over year in absolute terms ($926.4 million versus $935.7 million) and, more importantly, on a per-ride basis, against 12% ride growth. This is the single most bullish datapoint in the recent financials.

Operating expenses. The FY2023 → FY2025 pattern reveals the Risher-era reallocation with unusual clarity. R&D fell 29% in FY2024 (from $555.9 million to $397.1 million) as the April 2023 restructuring took hold, then rose 14% in FY2025 to $451.4 million on rehiring ($32.5 million personnel, $17.9 million stock compensation). Sales and marketing moved in the opposite direction — up 64% in FY2024 and a further 11% in FY2025 to $875.1 million, with incentive programmes rising from $402.8 million to $445.8 million and third-party marketing partnerships up $32.9 million. The company has been systematically shifting spend from engineering to demand generation and partner economics. G&A grew 7% in FY2025 to $1,002.1 million on $45.2 million of consultant and advisory costs (largely M&A-related) and $35.5 million of higher self-retained general business liabilities, against the absence of the prior year's $29.6 million lease-termination gain.

The Q2 2026 warning sign. Sales and marketing rose 68% year over year in Q2 2026 to $320.0 million. Revenue grew 16%. Rides grew 12%. Active Riders grew 17%. The company is buying growth at an accelerating price, and net income of $50.3 million missed consensus of roughly $56 million even as revenue beat. The market reaction — a 5.5% decline on 12 August 2026 — reflects this, not the headline records.

The FY2025 net income figure is an accounting event, not an earnings event. Net income of $2,844.0 million comprises a pre-tax loss of $53.2 million plus a $2,897.3 million income tax benefit, of which $2.9 billion is the one-time release of the valuation allowance against US federal and certain state deferred tax assets. The company concluded, on the basis of "sustained profitability in the U.S. based on pre-tax book income adjusted for permanent book-to-tax differences," that realisation was more likely than not. A valuation allowance is retained against California R&D credits. Any valuation metric built on FY2025 GAAP EPS of $6.92 is meaningless. The trailing P/E of 2.29 reported by data providers is an artefact.

Balance sheet: notes and sources

FY2021–FY2023 balance sheet detail was not obtained in this research pass and is marked as unavailable rather than estimated.

Commentary — balance sheet

The FY2025 balance sheet expanded 66% to $9.03 billion, but $2.91 billion of that — 81% of the increase — is the newly recognised deferred tax asset. Strip it out and total assets grew roughly 13%, driven by the acquisitions (goodwill up $188 million to $440 million; intangibles up $136 million to $179 million) and by restricted assets supporting the insurance book.

Three features deserve emphasis.

First, working capital is structurally and deeply negative — minus $1.61 billion at year-end 2025 and minus $2.02 billion at mid-2026, with a current ratio of 0.65 and 0.59 respectively. In an ordinary industrial this would signal distress. Here it is a design feature: Lyft collects rider fares upfront, holds insurance reserves that are long-dated liabilities against short-dated cash inflows, and carries no driver receivables ("we maintain no accounts receivable from drivers"). The negative working capital is a float, and it is the reason operating cash flow ($1.17 billion in FY2025) so dramatically exceeds any measure of accounting profit. Roughly $829.0 million of FY2025 operating cash flow came from working capital movements, principally a $479.0 million increase in insurance reserves and a $385.6 million increase in accrued liabilities.

Second, that float is a leveraged bet on reserve adequacy. Insurance reserves grew 28% in FY2025 and a further 5.8% in H1 2026, to $2.31 billion. The company states plainly that it has "in the past experienced adverse development where we have needed to increase historical reserves attributable to liabilities in prior periods." If reserve development turns adverse at scale, both the income statement and the cash-flow statement reverse simultaneously. This is the principal analytical risk in the financials and it is not visible in any headline metric.

Third, the capital structure is clean but convertible-heavy. Total debt of roughly $1.0 billion consists of $460 million of 2029 notes and $500 million of 2030 notes, both convertible, both hedged with capped calls ($47.9 million purchased in 2024, $42.0 million in 2025). The $420 million revolving credit facility was entirely undrawn at year-end 2025, with $61.3 million of letters of credit reducing availability to $420 million less that amount, leaving the company with a net cash position of roughly $835 million excluding restricted assets. There is no near-term maturity wall: the earliest maturity is March 2029.

Cash flow

Trailing twelve months to 30 June 2026: operating cash flow $1,195.1 million; free cash flow $1,112.3 million — the fourth consecutive quarter above $1 billion on a TTM basis.

Commentary — cash flow

The cash-flow inflection is the most genuinely impressive element of the Risher turnaround. Lyft moved from minus $248 million of free cash flow in FY2023 to plus $1,116 million in FY2025 — a $1.36 billion swing in two years — while capital expenditure fell by two-thirds, from $150 million to $53 million, as the micromobility and Flexdrive fleets were rationalised. Free cash flow conversion of roughly 211% of Adjusted EBITDA in FY2025 vastly exceeds the ">90% annually" target set in 2024, but investors should discount this heavily: the excess is insurance-reserve float and accrual build, not distributable earnings. A more conservative view treats Adjusted EBITDA of $529 million, less roughly $322 million of stock compensation, as the sustainable owner's-earnings base — which is a considerably less flattering picture.

Capital return has begun in earnest: $50 million of buybacks in FY2024, $500 million in FY2025, $400 million in H1 2026 alone, with a fresh $1.0 billion authorisation added in February 2026 on top of $250 million remaining from the original programme. Shares outstanding fell 4.94% year on year to 378.5 million, giving a buyback yield of roughly 4.9%. No dividend has been paid or declared, and none is signalled.

Ratios

Ratio commentary. Almost every profitability ratio for FY2025 and the TTM period is corrupted by the deferred-tax release. ROE of 141–153% is an artefact of a one-time non-cash tax benefit sitting on a still-modest equity base; ROIC of minus 3.9% and ROCE of minus 2.8% — which strip the tax item and measure operating returns on capital — are the honest numbers, and they are negative. Lyft does not yet earn its cost of capital on an operating basis. With a WACC around 12%, the operating return gap is roughly 16 percentage points. Interest coverage and net-debt-to-EBITDA are not meaningful because the company runs a net cash position and reports negative GAAP operating income. The cash conversion cycle is not meaningful: there is no inventory and no driver receivable. The Altman Z-Score of −0.03 should be read with caution — it is designed for capital-intensive manufacturers and treats Lyft's structural negative working capital as distress rather than as float.

Management guidance and consensus

Management guides Q3 2026 to Gross Bookings of $5.50–5.67 billion (+15–19%) and Adjusted EBITDA of $183–203 million (3.3–3.6% of Gross Bookings), with sequential margin expansion and accelerating rides growth across all business lines in the second half. Full-year 2026 is expected to exceed one billion rides. The 2027 framework — approximately $25 billion of Gross Bookings, an approximately 4% Adjusted EBITDA margin, and more than $1 billion of free cash flow — was reaffirmed on 10 February 2026 with no change to guidance.

Consensus for FY2026: revenue of $7.35 billion (+16.4%), adjusted EPS of $1.48 (+18.2%) and free cash flow of $1.13 billion, from 38 contributing analysts. Three-year forecasts imply 12.33% revenue growth and 19.64% EPS growth.

The required bridge to the 2027 targets: a 16.2% Gross Bookings CAGR from the FY2025 base (against 15.0% delivered in FY2025 and 19–23% in the last three quarters, acquisition-assisted); and margin expansion from 2.9% to approximately 4.0%, roughly 110 basis points over two years, against 30 basis points delivered in FY2025 and 30 basis points year over year in Q2 2026. Both are achievable at the current run rate. Neither has slack.

Bull case

1. The insurance cost curve has visibly bent, and the market has not priced it. Cost of revenue fell year over year in Q2 2026 in absolute dollars and per ride, against 12% ride growth — the first such inflection in the review period. Management attributes this to safety feature innovation, proactive policy reform, and specifically California SB 371. Insurance is the largest single line in Lyft's cost structure and the historical reason its margin ceiling looked so low. If cost of revenue grows meaningfully slower than Gross Bookings through FY2027, the 4% margin target is comfortably met and the operating loss closes. This is the highest-conviction bull argument because it rests on a delivered result rather than a promise.

2. The valuation embeds terminal decline that the operating data contradicts. At 5.2× EV/free cash flow, 0.31× EV/Gross Bookings and 10.9× EV/FY2025 Adjusted EBITDA, the equity is priced as a melting ice cube. The operating data shows eleven consecutive quarters of double-digit ride growth, seven of double-digit rider growth, record bookings, record riders, and four consecutive quarters of trailing free cash flow above $1 billion. With 23.7% of shares sold short and $1.25 billion of buyback authorisation against a $6.28 billion market capitalisation, the setup is mechanically asymmetric: management could retire a fifth of the share count while the short base is squeezed on any AV catalyst — precisely what occurred on 17 September 2025, when the stock rose as much as 16% in a session.

3. Lyft owns the scarce physical layer of the AV transition, not the commoditised one. Every AV developer will need depots, maintenance, cleaning, charging, repositioning and 24/7 service-level management in every city. Waymo and its peers do not want to build that. Flexdrive already has it, has proven it — the June 2026 Nashville handover met every service-level agreement — and is expanding it with an 80,000 square foot purpose-built facility opening in October 2026. In a world where autonomy software commoditises across Waymo, Baidu, Tensor, Mobileye and NVIDIA-based entrants, the fleet-operations layer is the durable rent, and Lyft has a genuine head start on it.

Bear case

1. Growth is being purchased, and the price is rising sharply. Sales and marketing rose 68% year over year in Q2 2026 to $320.0 million, against revenue growth of 16% and rides growth of 12%. Net income of $50.3 million missed a ~$56 million consensus. Incentive programmes rose from $402.8 million to $445.8 million in FY2025 and third-party marketing partnerships by a further $32.9 million. If a 68% marketing increase buys 12% ride growth, the marginal customer acquisition cost is deteriorating badly, and the FY2027 margin target requires that spend to normalise at exactly the moment competitive intensity is peaking around AV launches. Note also that consensus models flat free cash flow in FY2026 despite 16% revenue growth.

2. Lyft lost the autonomy race in April 2021 and cannot buy its way back. Selling Level 5 for $550 million looked prudent in a cash crisis; it now looks like the most expensive decision in the company's history. Lyft has assembled seven AV relationships, and not one is exclusive or proprietary. NVIDIA's GTC 2026 announcements made the asymmetry explicit: Uber received a named 28-market robotaxi programme launching in Los Angeles and San Francisco in H1 2027, while Lyft was listed alongside Bolt, Grab and TIER IV as "also scaling robotaxi development." Waymo, Lyft's flagship partner, has a deeper, earlier and larger relationship with Uber across Phoenix, Austin and Atlanta. When AV supply is the scarce input, the platform with 10× the demand gets served first and at better terms. Lyft's own 10-K concedes these partnerships "may lead to actual or perceived conflicts of interest or misalignment of strategic objectives."

3. The financials disguise a business that still does not earn a profit, resting on a $2.3 billion reserve estimate. FY2025 GAAP net income of $2,844 million comprises a pre-tax loss of $53 million and a $2,897 million tax benefit. The operating loss widened from $118.9 million to $188.4 million. Return on invested capital is minus 3.9% against a cost of capital near 12.1%. Free cash flow of $1.1 billion is generated overwhelmingly by float: $829 million of FY2025 operating cash flow came from working capital, principally a $479 million build in insurance reserves. Those reserves — now $2.31 billion, 25% of total assets — are actuarial estimates in a class of business where Lyft has admitted prior adverse development, and where medical, repair and litigation-funding inflation continue to run hot. A single reserve strengthening would reverse the earnings and the cash flow simultaneously, and could also call into question the $2.9 billion deferred tax asset predicated on "sustained profitability."

Catalysts and monitorables — next twelve months

Analyst verdict

Lyft is two businesses wearing one ticker, and the market is pricing the wrong one.

The first is a mature North American rideshare marketplace that has been genuinely, methodically repaired. Free cash flow swung $1.36 billion in two years; capital expenditure fell 65%; eleven consecutive quarters of double-digit ride growth; and — most importantly — cost of revenue fell year over year per ride in Q2 2026 for the first time, which is the beginning of the answer to the insurance question that has capped this equity's multiple for five years. Management has hit its numbers repeatedly and reaffirmed the 2027 framework without flinching. The valuation, at 5.2× EV/free cash flow, does not credit any of it.

The second is a company making a large, late, non-proprietary bet on autonomy. Here the honest assessment is less kind. Selling Level 5 in 2021 was a decision Lyft cannot undo, and the consequence is on full display: seven AV partnerships, none exclusive, and an NVIDIA relationship in which Uber received a 28-market programme and Lyft received a mention. Flexdrive is a real and underrated asset — the June 2026 Nashville handover proved it — but fleet operations is a services margin, not a platform margin, and it does not confer the demand-side control that ultimately decides who captures AV economics.

Between those two readings sit the numbers that should govern position sizing. The operating loss widened in FY2025. ROIC is minus 3.9% against a 12.1% cost of capital. The cash flow is float, and the float is a $2.3 billion actuarial estimate on a book with acknowledged prior adverse development.

Verdict: a credible, cheap, well-managed operational turnaround attached to a strategically disadvantaged position in the transition that will determine the industry's terminal structure. The next twelve months — the Nashville depot, Waymo supply sharing, the insurance renewal and the sales-and-marketing line — will decide which reading proves correct. The asymmetry, with 23.7% short interest and $1.25 billion of buyback capacity against a $6.28 billion market capitalisation, is real. So is the risk.

---

End of dossier. Figures for FY2023–FY2025 and H1 2026 are drawn from Lyft's Form 10-K for the year ended 31 December 2025, the Form 10-Q for the quarter ended 30 June 2026, the associated earnings releases and prepared remarks, and the DEF 14A filed 10 April 2026. Market and consensus data are as of 11–12 August 2026. Items marked nd were not verifiable during this research pass and have not been estimated.