The death of software has been greatly exaggerated.
“If AI is going to kill businesses, the signs are already showing,” DA Davidson analyst Gil Luria said in a research note, observing that the industry is now in the third year of the AI shift.
The biggest impact on the software industry so far has been “customers who are scared of the narrative and won't commit,” he added. The latter may be a factor that is starting to change as companies realize that neither they nor their software vendors have been overrun by AI.
As this “scared” money returns to the market, DA Davidson, Piper Sandler and Truist Securities analysts have identified stocks that are likely to lead the recovery. All three companies have in common that they provide the infrastructure on which AI is built.
DA Davidson's top picks focus on specialized growth and infrastructure resiliency. Luria's 2026 stock pick is Commvault (CVLT), with a strong upside of over 50% and a $220 price target on the back of sustained momentum and a rebound in profitability.
Other notable stocks include supply chain and retail software company Manhattan Associates (MANH). The business is a “subscription acceleration story” with an ROIC of over 100% and a $250 price target. Next up is marketing platform Zeta Global (ZETA), which benefits from “traditional marketing technology replacement” and has a $29 target.
Rounding out the list are Box (BOX), which is gaining momentum through its “Enterprise Advanced” upgrade with a $45 price target, and Datadog (DDOG), billed as a “complete observability platform” for complex AI-driven environments with a $225 price target.
Alongside Luria's list, Piper Sandler analyst James Fish is eyeing Gen Z winners and infrastructure initiatives. Fish highlights that Rubrik (RBRK) has a $75 price target upon completing its transition to SaaS. Nutanix (NTNX) is gaining share from VMware, so it costs $50. And AXON has a recurring model price target of $563 in public safety and drone integration.
Meanwhile, Terry Tillman, an analyst at Trust Securities, writes that software skepticism often focuses on “seat-and-license pricing,” or the idea that if AI makes humans more efficient, companies will buy fewer software licenses. But he argues that the industry is simply evolving, with the rise of agent AI (autonomous bots that perform tasks 24/7) driving a shift towards consumption-based pricing.
“As workflows move from human-initiated tasks to autonomous agents running at scale, pay-per-use becomes the most logical way to capture value,” Tillman wrote, citing examples such as computing, data processing, and transactions. Unlike humans, AI agents never sleep, so billable events can occur around the clock.
