The AI-will-eat-everything story hit software stocks hard. Late last year and into early this year, investors grew convinced that large language models would let companies rip out expensive enterprise software and rebuild it cheaply overnight. Some software names fell 50% to 80% on that fear alone.
The earnings haven't backed it up. Margins have held. Growth rates haven't cracked. Free cash flow keeps coming. That gap between the panic and the actual numbers is usually where the opportunity sits.
The sell-off went too far, per Whitney Tilson of Stansberry Research, who isn't calling the AI threat fake so much as selectively applied. The market lumped together software that AI can genuinely replace with software running the operational guts of companies, and it rarely bothered to tell the two apart.
Microsoft (NASDAQ: MSFT) is a case in point. Its tools are so deeply wired into daily corporate life that most people never see the internal systems running underneath, and its recent earnings pop showed what happens when a deeply embedded software name proves it isn't going anywhere. Microsoft isn't one of the three names getting the spotlight here, but it frames why the other three matter.
The Earnings Case the Sell-Off Missed
Start with ServiceNow (NYSE: NOW). The company's software runs IT service management, risk management and operations for large enterprises, and its second-quarter 2026 results, reported in late July, beat guidance and came with raised full-year revenue projections.
The stock had fallen by more than half from its highs during the broader sell-off. Even after rebounding, shares trade at a forward earnings multiple in the mid-20s, well below the roughly 60x multiple the stock has historically commanded.
That's a rare setup: a company still growing at a double-digit clip and throwing off strong free cash flow, priced closer to a market multiple than a premium one. The logic, per Tilson, is simple. The cost of running ServiceNow is a rounding error next to workforce costs, and enterprises don't rip out mission-critical software to save a rounding error.
A Life Sciences Moat the Market Is Underpricing
Next is Veeva Systems (NYSE: VEEV), a name most retail investors haven't heard of because it's built specifically for pharmaceutical and biotech clinical trials. That narrowness is the moat. Regulators are accustomed to seeing data presented in Veeva's format, and most major pharma and biotech companies run on it.
Shares dropped sharply during the sell-off after historically trading near 60 times forward earnings, and have since recovered to put the forward multiple in the low-20s. Veeva carries roughly $7 billion in cash against essentially no debt, and management has leaned into buybacks now that shares look cheap relative to history. Growth has cooled from an earlier 20%-a-year pace into the mid-teens, but for a business this dominant in a regulated niche, that still supports a premium multiple over time.
The Value Play With the Widest Range of Outcomes
Then there's Adobe (NASDAQ: ADBE), the most beaten-down of the three and the one facing the most credible AI competition. Content-creation tools are evolving fast, and AI-native rivals are chipping away at the edges.
At a forward earnings multiple near 10, roughly a third of its historical range, the market is pricing in a real growth slowdown. Yet the most recent quarter still showed double-digit revenue growth, in line with its historical trend, and Adobe is folding AI tools into its existing products rather than standing still. Analyst opinion is split, with several downgrades this year alongside other firms initiating coverage with double-digit upside targets, which captures just how unresolved the debate is. If sentiment simply stabilizes, a re-rating back toward 20 times earnings would mean a double from here. If the AI-competition thesis proves out instead, this is the name most likely to keep drifting lower.
What the Analyst Downgrades Are Really Signaling
Analysts turned bearish on several of these names near the bottom, cutting price targets and ratings right as the underlying numbers held up. A consensus reversal at the low is often a more contrarian signal than a warning. One prominent AI-focused hedge fund unwound its entire public portfolio in late July after getting caught leveraged long AI infrastructure names and short software stocks, including Adobe, a reminder of just how one-sided the AI trade had become before software started to recover.
Tilson's software call isn't his only contrarian bet right now. He's making a similar case on the commodity side, arguing the market has moved on from nuclear energy too early and is underpricing geothermal as the next major beneficiary of surging AI-driven electricity demand. Investors who want the full rundown on his next commodity supercycle pick can find it here.
The 2000 dot-com crash offers a useful parallel. Traditional retailers like Home Depot (NYSE: HD) got hammered on fears that e-commerce would gut them, and while Amazon (NASDAQ: AMZN) eventually thrived, so did the retailers investors had left for dead.
The risk across all three names is real. AI adoption could still erode software pricing power over a multi-year horizon in ways that haven't shown up yet. The upside is that quality software businesses trading at compressed multiples, with stable margins and intact growth, have historically been a durable place to put capital. Stay focused on the earnings, not the AI headlines, because that's what actually moves these stocks from here.
The article "Buy the Dip or Run: 3 Software Stocks Down 50% Face Their Moment of Truth" first appeared on MarketBeat.
