ADBE, ORCL, CPRT After Hours: Three Drawdowns, Three Different Trades

Three software-and-services names reported after Thursday's close, and all three walked into the print carrying deep 2026 wounds. Adobe is down roughly 29% this year, Oracle and Copart each about 21%. Same headline, three different stories. A price drop without a comparison set is just a number — so let's sort the genuinely de-rated from the aptly de-rated, and assign each one a job in a portfolio.

Adobe: cheap relative to the quality still on the books

Adobe beat. Fiscal Q3 delivered non-GAAP earnings of $6.13 a share, ahead of expectations, on record revenue of $6.76 billion, up 13% from a year earlier. Yet the stock enters the reaction already down 11% in five sessions and nearly 29% for the year.

That gap — solid results, punishing tape — is what makes Adobe the interesting one. Strip the price action out and you have one of the strongest profitability stacks in software: an 89% gross margin, a 40% free-cash-flow margin, roughly 62% return on equity, and a balance sheet with essentially no net debt. Now run valuation against a peer: Adobe trades near 14 times forward earnings and about 4 times sales, versus Microsoft at roughly 27 times earnings and above 11 times sales. A high-quality compounder sitting at roughly half the multiple of a large software peer is, by most factor frameworks, the definition of cheap quality.

The catch is growth. This is a low-double-digit grower now — around 12-13% — not the 20%-plus franchise it once was, and the market is arguing the creative franchise faces AI disruption. Momentum agrees with the pessimists for now: price sits under both its 50- and 200-day averages with a relative-strength index near 40, and AInvest's aggregate signal labels the name Hold rather than Buy. Nothing about the quarter restarted the growth engine.

The portfolio read: Adobe belongs in the patient, quality-value leg of a barbell — the sleeve that holds high-margin cash generators at sane multiples while the market fights over momentum. The trigger that changes the case is not the price; it's whether roughly 12-13% growth stabilizes as the new floor. If it does, today's discount is the opportunity. If the next few quarters show the sub-20%-growth era sliding lower, the cheapness is a value trap wearing quality's clothes.

Oracle: real growth, paid for with the balance sheet

Oracle is the other end of the barbell and the most volatile of the three. Revenue rose 30% to $19.3 billion, cloud infrastructure revenue jumped 121% to $7.4 billion, and earnings of $1.92 a share cleared the ~$1.67 consensus. Remaining performance obligations — the contracted work not yet billed — surged to a record $664 billion after more than $30 billion in new AI-cloud bookings, and Oracle raised its full-year revenue outlook. The stock jumped about 7% after hours. Growth here is not a story; it is a disclosed number.

The question is what it costs to get it. This is growth funded by debt and capex, and the factor stack shows the friction: trailing capex of roughly $76 billion a year against about $47 billion of operating cash flow, net debt near $88 billion, and debt at roughly 1.9 times equity. Return on invested capital sits near 13% — respectable, but thin once you finance the buildout with $236 billion of total debt. You can see the strain in the valuation curve: Oracle trades at about 23 times trailing earnings but roughly 35 times forward earnings. A forward multiple higher than the trailing one is the market saying it expects near-term earnings to be diluted by depreciation and interest — the exact cash-burn fear Reuters says the strong quarter tempered, rather than retired. AInvest's aggregate signal still labels Oracle Buy.

The portfolio read: this is the aggressive growth sleeve, sized with discipline. It is not a hedge against anything — it is the bet itself, and it only works if the record backlog converts into profit rather than staying trapped as depreciation. Oracle's job is to deliver the upside of the AI buildout; the discipline is to hold it at a size where a quarterly miss on that conversion is survivable, not ruinous.

Copart: a fortress you'd be overpaying for

Copart is the cautionary case, and it is the one where the aggregate and the factor math diverge. The auto-salvage auctioneer reported fiscal Q4 earnings of $0.35 a share on revenue of about $1.15 billion — shy of the roughly $0.39-0.41 and $1.21 billion analysts had braced for. The deeper problem is that revenue grew just 0.4% from a year earlier and fell for the quarter in a row. This is a business that has stopped growing.

The quality is beyond dispute: an 87% gross margin, a near-7.9x current ratio, and no net debt at all — a fortress balance sheet that throws off steady cash. And AInvest's aggregate signal labels Copart Buy. But that is the tension worth naming. At roughly 18 times trailing earnings and about 6 times sales, the market is paying a premium multiple for a business whose growth has stalled and that just gave back its own guidance. Copart's announcement that it will acquire ACV Auctions in an all-cash deal, expected to close by year-end, is the one concrete catalyst that could put units back on the platform and rekindle growth — until it does, you are paying for quality without paying for growth.

The portfolio read: Copart's natural home is the stability sleeve, but the discipline here is not to buy the sticker. On the factor math, the right move is to wait for a cheaper entry or for the acquisition to demonstrate it can restore unit growth. Quality is not a license to overpay; it is a reason to be patient with the entry price.

What the three say together

Put the three in one portfolio and they assemble themselves into a barbell. Adobe is the cash-generating quality leg at a de-rated valuation, patient in temperament — slow, safe, and finally cheap enough to matter. Oracle is the AI-growth leg that carries the upside and the risk, funded on borrowed money and debt-funded capex, to be sized small and held with discipline. Copart is the reminder that steady quality still has a price, and that a static Buy from an aggregate screen — however well-intentioned — is not a license to pay it at 18 times no growth.

Same drawdowns, different trades. One is a value opportunity conditional on growth holding; one is a growth bet conditional on the buildout converting; one is a quality name simply waiting for a better number. Which one is "in trouble" depends entirely on which job you bought it for — and that is the part no single drawdown percentage can tell you.