At a glance: Three S&P 500 names report today and all three come after the close — Oracle, Adobe and Copart. Oracle's 11.9% implied move is the largest single-name event of the week and the one print the wider AI capital-spending complex will trade off. Cooper Companies is graded below: a 2.7% EPS beat, a revenue miss, and an 18.5% fall after the bell.

The tape ahead

Oracle is the marquee print of the week and the straddle says so: 11.9% implied into tomorrow's expiry against a 1.6% average realized move over the last eight prints, the widest implied-to-realized gap on the board by a distance, and on the thinnest beat rate of today's three names — 62%, against 100% for Adobe and 75% for Copart. Consensus of $1.74 in EPS (down 17.6% vs prev.Q.) on $19.13B of revenue (down 0.3% vs prev.Q.) reads as a sequential step-down that is almost entirely the fiscal calendar: the May quarter is Oracle's seasonal high-water mark on license and quarter-end closings, and the revenue line is effectively unchanged. Against the $1.47 of a year ago, the growth is intact. Nobody is trading this on the EPS line. The question is conversion — the remaining-performance-obligation backlog has been inflated by a handful of enormous multi-year AI-infrastructure contracts, and the market wants evidence that signed capacity is becoming recognized revenue on schedule, at a gross margin that survives a mix shift toward rented compute. Behind that sit the three things that have made this an AI-credit story as much as an equity one: a capital-spending program far larger than the operating cash flow funding it, free cash flow that has gone deeply negative, and a debt stack raised specifically to put up data centers for a customer list concentrated in a small number of AI labs. Trailing EPS growth (+35.1% 1-year, +24.0% 3-year, +5.2% 5-year, +11.0% 10-year) shows how recent the inflection is: the one-year rate is the build-out already landing in earnings, the five-year rate the low-single-digit software compounding it displaced. The read-across is wide — every name levered to AI data-center capital spending trades on the backlog line tonight.

Adobe reports alongside it, with consensus at $6.09 in EPS (up 2.1% vs prev.Q.) on $6.70B of revenue (up 1.3% vs prev.Q.) — steady sequential progress against a year-ago $5.31 that shows the earnings base still compounding even where the multiple has not. The set-up is awkward in a specific way: Adobe has beaten in all eight of its last prints and guides conservatively enough that clearing the headline carries almost no information, which leaves the AI disclosure and the fourth-quarter guide to do the work. What the market wants is evidence that the free users pulled in by the generative tools are converting into paying seats at a price that does not cannibalize the core creative subscription, and that net-new subscription growth is organic rather than acquired — a recent deal has made that harder to read from the outside. Competition is the other half of the argument, with design-native and model-native rivals growing off smaller bases and increasingly sitting where the creative workflow begins. Governance adds its own noise: a chief-executive handover is already scheduled for December and the head of the creative business leaves this month. Options price 8.2% against a 1.4% average realized move, rich by any historical measure and a fair reflection of how binary the monetization question has become. Trailing growth (+35.1% 1-year, +18.3% 3-year, +9.0% 5-year, +29.7% 10-year) is not the issue here; the terminal value is.

Copart closes the slate and is the quietest of the three by design. Consensus of $0.38 in EPS (down 11.0% vs prev.Q.) on $1.14B of revenue (down 7.5% vs prev.Q.) is partly seasonal — the January-to-April quarter carries winter catastrophe volume the summer quarter does not — but consensus also sits below the $0.41 earned a year ago, so this is a down quarter on both comparisons, and seasonality alone does not account for that. The pressure point is US insurance unit volume: softer collision claim frequency, drivers carrying thinner coverage in response to higher premiums, and lost assignment share have pulled the core unit count down for several quarters running. Against it, total-loss frequency keeps grinding higher as vehicle complexity and repair costs push more damaged cars past the economic repair threshold, and management has been pointing at non-insurance volume — dealer consignment and whole-car — plus international as the growth it controls. Cost is the other watch item, with operating expense per unit rising on long-haul transport and yard expansion, so margin can compress even where the mix improves. A 7.2% implied move against a 2.1% realized average is a real premium on a 75% beat rate, and trailing growth (+13.6% 1-year, +12.0% 3-year, +16.7% 5-year, +22.5% 10-year) describes a compounding record that is slowing rather than breaking. Whether the insurance-volume drag is cyclical or structural is exactly what tonight is for.

Scorecard — reported since we last wrote

We wrote yesterday that Cooper's straddle was pricing the strategic review and the guide rather than the quarter, and the tape settled that argument emphatically. EPS of $1.15 beat the consensus we carried by 2.7%, but revenue of $1.07B missed by 2.9% — the two surprises disagreeing — and reported EPS was still 5.0% below the prior quarter. The stock then fell 18.5% after the close, a reaction more than twice the implied move the options market had set for the event and entirely disconnected from a beat that close to consensus. When a print moves a mid-cap that far on a headline that near the number, the information was in everything around it — the outlook, and what the board had to say about its own review — and not in the quarter at all.

Informational only — not investment advice. Figures are consensus/estimates and option-implied values from public sources and may be revised.