Lithia: A $2,000 Used-Car Floor, an Agentic Cost Plan, and a 37% Drawdown
The quarter beat — EPS +9% on a $10 billion top line — but the tape stopped paying for the operational turn three months ago, and the balance sheet is where the argument now lives.
LAD · Earnings Call · 2026-10-21
A beat the stock has already refused
On July 29, the day Lithia held its second-quarter call, the stock printed a high of $427.48. Ninety trading days later it walks into this third-quarter print down 37% from that peak and 12.8% over three months. Then management delivered adjusted diluted EPS of $10.34, up 9% year over year, on revenue just shy of $10 billion. The gap between the operating story and the tape is the entire dossier.
Because operationally, this was the cleanest quarter Lithia has strung together in years. Total vehicle GPU of $3,958 fell only $63 year over year despite same-store revenue slipping 1.4% against a nearly 8% prior-year compare. Aftersales gross margin hit 59.5%, up 100 basis points, growing profit in an eighth consecutive quarter. North American same-store SG&A fell to 65.7% of gross profit, a level the CEO has been chasing publicly for over a year — on the July call he called 66.2% “the first time that Lithia & Driveway returns to the #1 position in prominence as the lowest SG&A in North America” — Bryan DeBoer, President and CEO · 2026-07-29, and then improved on it again.
The used-car machine: from thesis to underwriting
Used GPU reached $1,993, up 13% and $234 per unit, with the core product bucket — the three-to-eight-year-olds — carrying the load, GPU up roughly $500 per car. Yet used units were down about 3%. That is the trade management has been evangelizing for six quarters: give up a little volume to stop leaving money on the lot.
The new word for it is the micro funnel, and it is genuinely fresh language for this company — a reframing of the problem from pricing to psychology:
A lot of our micro funnels still haven't figured out that they don't need to sell the car to the first person that walks in because their funnel is small, and they flinch and sell it to the first customer.
The mechanism is unchanged from the April call, when Bryan DeBoer said Lithia was selling at “approximately 95% of what the 1 price used car retailers are selling the same Carrefour” — Bryan DeBoer, President and CEO · 2026-04-29. What is new is that the gap is now closeable in practice, not theory: a $1,500-per-car prize on a $30,000 average car, extended across ~350,000 units, is the ~$500 million management keeps pointing at. And unlike prior quarters, they will now underwrite it forward — asked whether ~$2,000 is a floor into 2027, DeBoer answered, “looking into 2027, that is absolutely accurate” — Bryan DeBoer, Chairman and CEO · 2026-10-21, with the caveat that used GPU typically softens seasonally in Q4.
That caveat matters. Lithia's fourth quarter is structurally its worst — snow belt, and a UK calendar with no major registration month. day supply is roughly 60 days and cleaner than last year, which is the only cushion on offer.
Agentics: the vocabulary changed, and the promise finally has a number
In the third-quarter keyword set, "AI solutions" was this company's number-one theme. This quarter it is simply gone from the top 30, replaced by Agentics and a wonderfully specific artifact of it — windshield wipers.
The UK is the live experiment. On the call, DeBoer credited AI-generated appointment setting with driving 10%-11% additional pre-sales per booking — fluid changes, wipers, the upsell that normally happens only after a technician inspection, if at all. He also noted the UK has been running the tooling for “about six to nine months, and we're just now starting to see some of those nice gains” — Bryan DeBoer, Chairman and CEO · 2026-10-21. That is a far more concrete AI datapoint than most industrials offer.
The quantification: of the roughly 1,000 basis points of SG&A reduction the company wants, half comes from agentics and half from volume leverage and fixing underperforming stores. But the sequencing is unforgiving — “the drive to a sub 60% SG&A will trail the rollout of the Pinewood stores probably by 12 to 24 months” — Bryan DeBoer, Chairman and CEO · 2026-10-21. Pinewood completes in North America by end-2028. So the target is a 2029-2030 event, not a 2027 one. Meanwhile the acquisition of Pinewood by Ridgeview Partners closed this month, generating a valuation gain while management insists strategic alignment is unchanged — a neat financial engineering trick, but it does not accelerate the cost bridge.
Doc fees, the FTC, and a competitor-clearing event
The single most prominent new keyword this quarter is Doc fees, and it is a rare case of a regulatory shift that helps the incumbent. DeBoer's read: Lithia was already disclosing doc fees in every state and any repricing happened in Q2, so “it should allow them to be more competitive in the long run” — Bryan DeBoer, Chairman and CEO · 2026-10-21 against less transparent independents. He attributes part of peers' new-GPU pressure to higher-doc-fee-state exposure.
Note what else this section is doing: it plants the actual goodwill argument, which is really a valuation plea. Public dealers trade at 12-13x earnings or 7.5x for Lithia; private transactions clear at 3-6x normalized EBITDA on goodwill. Management's whole capital allocation story now rests on the claim that the public multiple is wrong.
The balance sheet is where the argument gets hard
Driveway Finance is the best-performing asset in the company: $937 million of originations, 4.7% net interest margin, penetration at 18.3% toward a 20%+ target, average origination FICO of 752, front-end LTV of 94.6%, and an allowance for loan losses down to 2.8% from 3.1% as the book matures. Finance operations income of $32 million rose nearly 70%.
But DFC is also a balance-sheet consumer. Effective net cash sits at -$9.2 billion, down 85% year over year, and liabilities to assets at 75.5%, up 4.5 points year over year. Net leverage is 3.1x excluding floorplan and DFC securitizations, down from 3.2x, against a 5.75x covenant and a stated <3x target. The last filed quarter showed $262 million of net income, up just 1%.
So when an analyst asked directly about de-levering into a hiking cycle, the answer was revealing: roughly $1.5 billion of annual free cash flow, a preference for staying at 3x, and buybacks over M&A because North American store prices are too high. Lithia repurchased $85 million of stock at an average of $345 — above where the shares sit today. The prior call framed capital allocation as “1/3 on buybacks, 1/3 on M&A” — Bryan DeBoer, President and CEO · 2026-07-29; this quarter it was all buyback plus the $250 million Gossett Hyundai/Kia deal in Memphis.
At 8.9x trailing net income, the bull case writes itself. The bear case is that a 37% drawdown into a 9% EPS beat is the market saying it does not believe the SG&A bridge is a 2027 story. It isn't — it's a 2029 one, gated on a DMS rollout, a used-car pricing culture change, and a credit book that is growing faster than the cash flow funding it.