Software stocks are falling out of favor after years of market leadership, as investors fear the group may be one of the first major victims of artificial intelligence advancements. Big software companies Intuit , ServiceNow , GoDaddy , AppLovin and Adobe are the S & P 500’s worst performers so far this year, in that order, as of Friday’s close. As AI adoption has accelerated, including experimentation with AI agents , the stocks appear to reflect growing concern that automation could fundamentally reshape enterprise software businesses and weigh on the companies’ pricing power. The latest AI agents such as Anthropic’s Claude Cowork are built for basic productivity tasks such as file management that overlap with what major software companies offer. Many, including Adobe and Salesforce , use “per-seat” subscription models that charge companies a recurring fee for each individual user, or seat. It’s possible AI agents will be able to replace these human seats, undermining software’s predictable revenue streams, which are a big draw for investors in these stocks. “The market’s pretty smart and starts to realize where things are going,” said Melius Research’s head of technology research Ben Reitzes. “Many of the SaaS companies are AI- and agent-last and seat-first, and it’s an ‘Innovator’s Dilemma.'” Reitzes, who first declared that “AI is eating software” in April 2024, was referring a concept popularized by Harvard professor Clayton Christensen, that explains why disruptive technology supplants previously successful incumbents. “The pricing models and the way the whole market is served could change under AI. And it’s one of those things where it’s just very hard to fight it,” Reitzes told CNBC. Software companies have been benefiting from significant price increases they made during the Covid pandemic and the Covid-era hiring sprees that inflated seat growth, Reitzes said. “Those two things alone are very troublesome, where they rely on price increases and fake seats to grow,” he said. Cost of coding expected to fall Some analysts argue that AI will not entirely replace software because it lacks deeper context. Reitzes, on the other hand, believes that hurdle may be overstated given that AI agents can draw context directly from customer data, rather than enterprise systems. “Many of the obstacles to deploying enterprise software are going to be knocked down because the cost of coding is going to get much closer to zero, and so you’re going to be able to try a lot of solutions that obviate these expensive applications, as long as you can get access to the data layer. And over time, AI will get better at understanding context,” Reitzes said. “The key is if agents can get access to the data and accuracy.” To be sure, many analysts believe this plunge in software stocks could be short-lived as leading Software-as-a-Service companies further integrate AI into their offerings. Software vendors have started to blend usage-based models with per-seat pricing in an effort to monetize AI. Separating winners and losers Both Arjun Bhatia, an analyst at William Blair, and Gil Luria, head of technology research at D.A. Davidson, both believe the sell-off could be overdone. Fears of AI model disruption and cyclical outflows do “not justify the broad-based indiscriminate selling we are seeing across the sector,” Bhatia wrote in a Thursday note to clients. What could improve sentiment in the group is if software companies scale AI monetization and see a general reacceleration in growth rates, he said. Luria said the sell-off has opened up a good entry point for investors. Software companies remain scalable, recurring, high incremental margin businesses, he said. “AI hasn’t actually disrupted any of these businesses, but the conventional wisdom is that it will disrupt all of them. The better the AI models get the more investors are convinced that we will no longer need enterprise software and that it will all be replaced by vibe coding,” Luria told CNBC. He expects the process of separating the winners from losers is underway this year, and demonstrating revenue acceleration will be vital. “We point out that those winds are fickle and can switch directions quickly — stocks like CRM, ADBE, PATH and ESTC have all gone through a full phase of being considered a huge AI winner and then a huge AI loser. Which is why we now find ourselves with several software companies trading at previously unheard of multiples on cash flow,” Luria wrote in a Jan. 14 note, in which he reiterated his buy ratings on Snowflake , DataDog and Box . Reitzes, too, said he likes Snowflake and MongoDB as software names less vulnerable to AI disruption. The companies fall into the category of “data connectors,” he said, meaning they make data usable and secure across systems rather than work on end-user applications.
