Some days the tape hands you a confession.

Some days the tape hands you a confession. Today it handed you seven.

Cerebras Systems down nearly twenty percent on the print. No margin data in front of me yet, but a gap that size on a freshly public AI infrastructure name is not noise — that is the market reading the prospectus more carefully than the roadshow crowd did, and whatever is in that filing about customer concentration or gross margin structure just got priced in the hard way. This one goes on the teardown list immediately.

Nike. Earnings per share down forty-nine point four percent year over year, and the stock is trading at almost twenty-eight times those gutted earnings. That is the number that should stop you cold. The margin compression here is not cyclical softness — it is the consequence of years of pulling back from wholesale, betting everything on direct-to-consumer, and then watching the DTC channel disappoint while the shelf space they abandoned got filled by On and Hoka. You do not rebuild distribution in a quarter.

Micron. Eighty-five point six percent revenue growth, four hundred and twelve percent earnings-per-share growth, forty-one and a half percent net margin. That is a remarkable print. The catch is the fifty times earnings multiple sitting on top of it — the Street has already priced in the next cycle, so the question is not whether Micron executed, it is whether memory pricing holds long enough to justify the valuation.

Paychex. Revenue up sixteen point four percent, net margin at twenty-five point eight, and then earnings per share somehow down five point six percent on the year. That arithmetic deserves a closer look at what is happening below the operating line — interest allocation, share count, one-time items. The business looks healthy; the per-share story does not match.

Carnival. Earnings per share up fifty point five percent, net margin at eleven and a half, and the stock barely moved. The market is either efficient here or it is bored. At twelve and a half times earnings for a business growing at that rate, somebody thinks the cycle turns before the debt does.

FedEx. Revenue up four point seven percent, earnings per share up seventeen point nine percent, and essentially no gap on the print. That earnings-per-share leverage on modest top-line growth is the buyback and cost program doing exactly what management promised. The question for a full teardown is whether the operating margin expansion is structural or one more year of expense cuts masking a volume story that never fully recovered.

Sunbelt Rentals down nearly three percent on the print with no detailed financials yet in hand. Equipment rental is a leading indicator — when that gap goes negative, you want to know if it is fleet utilization, pricing, or something in the guidance. Worth watching.

The two I am pulling apart in full: Cerebras Systems, because a nineteen-plus percent gap on a recent IPO almost always traces back to something specific in the prospectus, and I want to find it. And Nike, because forty-nine percent earnings-per-share deterioration at a brand that still commands a premium multiple is a structural argument hiding inside a cyclical headline.

That is the menu. The numbers were always there — most people just do not look. See you at the next filing.