Flash Intel: What the Dealer Earnings Actually Say, the Toyota Halt, and the July SAAR Math

Five public dealer groups reported Q2 this week. Four beat estimates. One missed badly. Toyota halted three factories in Japan after an earthquake and does not know when they restart. July new-vehicle pace is the strongest of 2026. Here is what the numbers actually say, not what the press releases say.

The public dealer earnings: what drove the results, what held them back

Lithia: the used car operation saved the quarter

The headline was a 13% earnings beat. What actually happened: Lithia's new vehicle business declined. Same-store revenue fell 1.6% and new vehicle units sold dropped 2.2%. That is not a strong business. What covered it was the used vehicle operation turning a $339 per unit improvement in gross profit in a single quarter, and the finance arm, Driveway Finance, posting record originations with income up 70% year over year. The company is becoming more of a finance and used vehicle business that happens to also sell new cars. That is not a criticism — it is the model working. SG&A tightened 290 basis points in one quarter, which is the cost discipline story analysts rewarded. Aftersales up 3.1% as well. The store that is winning right now is the one where service and used vehicle are carrying the month. Lithia proved it at scale.

What to watch: Pinewood AI rollout in North America. Management flagged potential store-level disruption. That is a real operational risk worth tracking if you follow how tech rollouts typically go at multi-brand dealer groups.

Penske: international exposure and trucks are the story no one is writing about

Revenue up 11%, earnings beat by 6.7%, 23rd consecutive dividend increase. The headline looks clean. Look under it: six-month adjusted EPS is down 13% year over year. The quarter beat because same-store retail automotive revenue rose 6%, but free cash flow was negative $72 million versus positive $119 million a year ago. That is a significant reversal. The bright spots are commercial trucks, which are recovering, and the international footprint, which added $47 million in revenue from currency tailwinds. Service and parts gross margin improved 80 basis points, continuing the fixed ops story. New vehicle GPU of $4,782 is the highest in the group this week, which says something about how Penske manages allocation and mix. The 23rd consecutive dividend increase signals confidence. The cash flow number says watch closely.

Carvana: a 20% gross margin in an industry where 16% is considered strong

The EPS beat was small. The margin story is not. Carvana is running a 20.13% gross margin TTM in a business where franchise dealers consider 15-17% a good quarter. That gap exists because Carvana reconditioning cost per unit, logistics model, and lack of physical showroom overhead are structurally different from a franchise dealer. They are not better at selling cars. They are better at the cost side of selling cars. The question for Q2 is what happened to acquisition cost per unit in a quarter where 29.6% of trade-ins carried negative equity. When the full call drops, that is the number to look for. If their acquisition cost rose, the margin advantage compresses. If they held it, the model is more durable than the industry wants to believe.

Sonic: a beat with nothing behind it to pull apart

EPS $1.82 versus $1.76 estimate, 3.4% beat. Gross margin 15.9%. No material surprises in either direction. Sonic operates an eclectic brand mix and its results tend to track the market without standing out in either direction. That is not a knock. In a softening market, not standing out negatively is its own form of execution.

Group 1: the one that actually tells you something about the market

EPS $9.61 versus $10.92 estimate. A 12% miss. That is not noise. Group 1 has one of the larger international footprints among the U.S. publics, including a significant UK operation. Currency, regulatory environment, and consumer demand in the UK are all working against them right now. The domestic business is fine. The consolidated number is not. What makes this worth paying attention to: Group 1 is the cleanest read on whether international dealer operations are a strategic advantage or a drag right now. The answer this quarter is drag. Full call details will tell us whether this is fixable in Q3 or structural.

AutoNation reports Friday. The full earnings deep dive publishes after that.

Toyota, Nissan, Honda: what an earthquake in Japan means for your lot

A 7.1 magnitude earthquake struck Kumamoto, Japan on July 28. Toyota halted three plants in Fukuoka, Miyata, Kanda, and Kokura, through at least July 31. Lexus output is suspended. Toyota tried to restart Wednesday morning, stopped again by afternoon. Their public statement: the situation is "changing daily." Nissan is evaluating its Kyushu assembly operations. Honda extended suspension of its Kumamoto factory. Daihatsu halted its Oita plant.

Kyushu is also Japan's semiconductor hub. Chipmakers halted facilities simultaneously to assess structural damage. The overlap matters because modern vehicles, including every hybrid and EV, are chip-intensive. Parts maker damage assessments are still coming in. Aftershocks are continuing. There is no reliable timeline for full restart.

Why this matters for a dealer in the U.S. right now: the models with Kyushu production exposure are some of the fastest-moving units on dealer lots. RAV4, Camry, Corolla, and Lexus RX all have manufacturing in that region. At 54, 61, 79, and 108 average days on lot respectively, these are not emergency situations today. But if this halt extends two to three weeks into August, allocation on those models could tighten by late August or September. The dealers who notice that early have a pricing conversation the others have not had yet.

The July SAAR number: what J.D. Power is seeing vs. what our model projects

J.D. Power and GlobalData are tracking July new-vehicle sales at a 16.9M annual pace, the strongest month of 2026. Our own SAAR model has Q3 projected at 15.74M as a quarterly average. That is a 1.16M gap, and it is intentional, not an error.

J.D. Power reports a monthly annualized rate, which is what is happening in showrooms right now. Our model projects a quarterly average under sustained macro pressure: low consumer sentiment, elevated loan rates, rising delinquency. A strong July does not mean the quarter averages strong. If July is being pulled forward by buyers rushing ahead of the August 19 Canada tariff, August and September pay that back. The quarterly average is what matters for planning.

We are tracking the gap and taking it seriously. This week we reduced the weight our model places on consumer sentiment, since the market is clearly transacting above what sentiment alone would predict. We also activated a pull-forward assumption in the model for Q3 only, which lifted our projection modestly. We are also watching for Q2 delinquency data from the Federal Reserve, which is still six months stale in our inputs and is a key variable. When that posts, expected in August, we update immediately. Until then: J.D. Power is telling you what July looks like. We are telling you what the quarter might average if the headwinds are real. Both are worth knowing.

Flash Intel is a quick read on what happened and why. Full deep dive on the public dealer groups publishes after AutoNation reports Friday. Sources: company press releases, Finnhub, Reuters, J.D. Power/GlobalData via CBT News. July 30, 2026. dgactual.com