AutoNation Retires Its $2,000 Used-Car Target — and Admits the Service Bay Is Leaking

A seventh straight quarter of EPS growth was bought with buybacks, while the CEO quietly shelves the gross-profit-per-unit number he spent a year promising and flags a new problem: customers deferring repairs

AN · Earnings Call · 2026-10-30

Seven straight quarters, and the tape still says no

AutoNation delivered its seventh consecutive quarter of year-over-year adjusted EPS growth — $5.11 against $5.01. Read the walk on slide five and the story is less flattering. After-sales added $0.38 of EPS, AutoNation Finance added $0.27, and share repurchases added $0.64. Vehicle volumes subtracted $0.52, vehicle profitability subtracted $0.28, and the catch-all line subtracted $0.39, which CFO Tom Szlosek broke into less SG&A leverage (~$0.13), wholesale vehicle profitability (~$0.14) and higher interest expense (~$0.08). The buyback was the largest single contributor to a 2% EPS gain. That is the structural picture, not a one-quarter artifact. Revenue has been flat-to-down for two years — $6.9B, down 1% year over year in the latest filed quarter — and gross profit of $1.2B, down 3%, still 8% below the 2022Q3 peak. The equity has noticed: the stock peaked at $229.98 on 29 July and sits 21.9% below that high, down 6.5% over 90 days. Price to net income of 7.9x, down 33% year over year, is what a market says when it doesn't believe the EPS is operational.

The service bay springs a leak

After-sales is the whole thesis: record gross profit of $617 million, ninth consecutive record quarter, roughly half of total company gross profit, customer-pay traffic up 5% and warranty traffic up 13%. It is also where the first genuinely new negative signal of this call appeared. “We are seeing higher levels of deferred work, and we are seeing less total hours per RO” — Mike Manley, Chief Executive Officer · 2026-10-30 That is a new entry in this company's keyword set: deferred work and hours per RO rank at the very top of the Q3 theme list, where a quarter ago the language was about penetration and conquest. Manley was careful to call it "not lost work" and to blame general economic pressure rather than competitive labor-rate pressure — but he then layered on a structural admission: the COVID-era hole in the 3-to-10-year vehicle park peaks in 2027. Traffic up, ticket content down, and the most profitable slice of the car park shrinking into next year is a hard combination. Margin told the same story quietly. Gross margin of 17.8%, down half a point; after-sales gross margin 47.8% versus 48.7%, entirely mix from wholesale parts, which grew 16% at a lower rate. Manley also conceded wholesale is now absorbing high fuel costs on milk runs, a fuel-cost theme that dominated the broader market's keyword list this quarter and echoes through airlines, truckers and railroads. AN is bearing it to protect the customer base — low margin, strategic purpose.

The target that now 'lives in the past'

Here is the contrast that matters. In May, Manley was explicit: “our internal view is that we should be moving towards $2,000 a unit” — Michael Manley, Chief Executive Officer · 2026-05-01. In July he was already reframing: “sometimes this maniacal focus purely on used volume actually misses the point” — Michael Manley, Chief Executive Officer · 2026-07-31. On this call, the number became a relic.

I don't think that I'm going to be that core I don't want to be constrained by a fixed used car margin because I think that, frankly, lives in the past and should live in the past.
Mike Manley, Chief Executive Officer · 2026-10-30

He replaced it with a dynamic return on capital model that incorporates inventory velocity, replenishment supply and customer lifetime value — and explicitly said “you'll probably hear me talk less about GPU” — Mike Manley, Chief Executive Officer · 2026-10-30. Reported used GPU was $1,524 versus $1,489, with sequential improvement in August and September after a weak July. The strategic logic is coherent; the rhetorical retreat from a public number, four quarters after setting it, is the tell. Consensus will now have to underwrite a model the company hasn't yet quantified.

FinCo: the annuity, and the reason CFS looks worse

The one unambiguous build is captive portfolio AutoNation Finance: $2.9 billion, up from $2.0 billion, quarterly operating profit $15 million versus $2 million, 24% return on equity, 91% non-recourse funded, four ABS deals in 18 months totalling $2.4 billion plus a new one next week and a new syndicated pre-ABS warehouse closed in October. Management guides to ~$120 million of FinCo cash flow in 2026 and at least $150 million in 2027. That build deliberately degrades the headline. Customer financial services gross profit fell to $349 million from $375 million, and per-unit to $2,698 from $2,775 — with 3.6 points of product terminations and a point from loans retained by FinCo, partly offset by incremental platform fees. Economics are migrating from upfront commissions to recurring finance income. Szlosek noted SG&A excludes FinCo entirely, so the reported SG&A ratio is penalised relative to a holistic operating margin. Which brings the quarter's real miss. “our cost performance in Q3, particularly in July and August, could have been better” — Tom Szlosek, Chief Financial Officer · 2026-10-30 — an unusual admission from a management team that spent February reaffirming a 66-67% SG&A-to-gross target. September initiatives in field labour, local discretionary spend, corporate advertising and commercial sales are supposed to carry into 2027. On funding, Szlosek was measured: “there are occasional exceptions, but I think generally speaking, the team has a really good process for pricing that gives consideration to the changes in the funding costs” — Tom Szlosek, Chief Financial Officer · 2026-10-30. With funding cost the dominant worry across every bank and lender reporting this week, that is a promise, not a result.

What AutoNation stopped saying

Two absences stand out. First, tariffs. Tariff refund was the single hottest keyword in the entire market last quarter, and a striking number of companies reporting in the last five days — ACCO, ACU, AOS, BOOM, ORLY, SMP, VFC, VAIAS — booked refunds. AutoNation's only tariff mention is as a prior-year comp headwind. The exposure has resolved into arithmetic. Second, the language of the shopping funnel. Manley repeatedly invoked agentic search — that real-price transparency now lives on third-party sites and that agentic shopping “will keep that pressure on there” — Mike Manley, Chief Executive Officer · 2026-10-30 in Q4. A franchise dealer talking about AI-mediated price discovery, while simultaneously arguing its dense physical network plus a digital layer is the moat, is the most forward-looking thing said on the call. It is also an admission that front-end gross is structurally contested. So the trade is this: a cheap multiple, a $1 billion incremental authorisation taking total buyback capacity to ~$1.4 billion, $169M repurchased in the latest quarter, up 408%, and a genuine annuity being built inside FinCo — against a shrinking prime service park, a quiet retreat from a public margin target, and a cost base that slipped in July. The used inventory correction is self-correcting. The 2027 vehicle-park hole is not.