Down 54% and Still “A Buy”? Falling Knife Enthusiasts Never Learn.

Down 54% and the market is wringing its hands over whether AI will eat software. Let me translate that: a basket of investors panic-sold an entire sector on a thesis no one has actually priced with any rigor, and now the same crowd is wondering if they missed the bottom. Brilliant. Truly the stuff of compounding wealth.

Here’s what actually happened, stripped of the drama. Software stocks sold off earlier this year because investors got spooked that AI-native applications from labs like OpenAI and Anthropic would disrupt entrenched SaaS businesses with cheaper, more customizable tools. The sector has since recovered. Notice what that tells you: the panic was a sentiment event, not a cash-flow event. Nobody reran the discounted cash flows. The market just changed its mood and called it analysis.

Now, is the disruption thesis wrong? No. It’s just unfalsifiable garbage at current levels of discourse. Anyone can claim AI kills SaaS. Show me a model of which revenue lines get compressed, on what timeline, at what margin impact, and I’ll engage. What I get instead is vibes, and vibes are not a position size. If AI-native tools genuinely cannibalize SaaS subscriptions, the winners are… companies selling you the tools, and the losers are companies with 80% gross margins? Please. Both sides of that trade are crowded with people who never wrote out a single number.

And that brings us to the stock down 54%. A 54% drawdown means the price needs a 117% gain just to get back to even. That’s not a reason to buy or avoid it — it’s arithmetic, and it should immediately kill the amateur instinct that a beaten-down price equals a bargain. A stock down 54% is cheap only if the underlying earnings power survives the disruption everyone’s worried about. If the moat is real and cash flows hold, you’re buying future cash flows at a discount. If they don’t, it’s a value trap with a discount sticker, and no amount of AI narrative changes the math.

Here’s the part that should genuinely bother you: this entire dilemma is optional. You don’t need to adjudicate whether Anthropic unplugs Salesforce. You could own the whole market through a broad index fund at a few basis points and let the spectacle resolve itself without your participation. The people trading in and out of software names on AI headlines are paying — in spreads, taxes, and mistimed entries — for the privilege of guessing. Historically, most of them lose to the person who did nothing, and the person who did nothing wasn’t clever. They just had a savings rate above 40% and the patience of a thermostat.

If you must pick individual stocks, then at least demand the boring things: durable revenue, sane valuation relative to actual free cash flow, and a plausible answer to why AI strengthens rather than erases the business. Quality getting thrown out in a sector-wide selloff is a legitimate opportunity. Quality being assumed because a stock fell is how people donate money to the market.

The AI boom will produce enormous winners, enormous losers, and one guaranteed group of casualties: investors who confused a news cycle with a valuation model. The market recovers from panics. It doesn’t recover your opportunity cost.