The software sector entered 2026 expecting a recovery after a bruising 2025, but a sudden burst of AI innovation has reignited investor fears that the industry’s old business model is under threat. Instead of a turnaround, the group is experiencing its worst start in years as the market wrestles with the prospect that traditional software may be replaced by AI agents that can complete complex tasks faster and cheaper than human workers.
The latest selloff was triggered by Anthropic’s Jan. 12 release of Claude Cowork, an AI tool capable of generating spreadsheets from screenshots and drafting reports from scattered notes. Although the product is still a “research preview,” its capabilities strike directly at the heart of software companies’ core offerings, amplifying fears that their platforms could be made redundant. The reaction was swift: Intuit, the owner of TurboTax, plunged 16% last week in its worst decline since 2022, while Adobe and Salesforce dropped more than 11% each.
The broader impact is stark. A basket of software-as-a-service stocks tracked by Morgan Stanley has fallen 15% in 2026 after an 11% decline in 2025, marking the weakest start to a year since 2022, according to data compiled by Bloomberg. For investors, the speed of change is unsettling. Bryan Wong, portfolio manager at Osterweis Capital Management, described the market’s uncertainty as the fastest he has seen in his career, noting that rapid innovation makes it increasingly difficult to forecast growth.
Anthropic’s new tool is precisely the kind of threat that has kept investors wary of software stocks. It demonstrates how quickly AI can be built and deployed, bypassing years of product development and undermining the advantages of incumbents. Jordan Klein, a tech specialist at Mizuho Securities, says this is why many institutional investors remain bearish: they see no catalysts that could justify higher valuations for software companies and believe that the sector may remain structurally disadvantaged as AI services evolve.
That pessimism has widened the gap between software and other tech subsectors. While the Nasdaq 100 flirts with record highs, long-established software firms like ServiceNow are trading at multi-year lows. The market’s concern is not just competition, but the slow pace at which established companies have been able to monetize AI. Salesforce’s Agentforce has been promoted as a major innovation, but has yet to materially shift revenue growth. Adobe has embedded generative AI into its creative tools, but its most recent earnings report omitted updates on key AI adoption metrics, raising questions about traction.
Despite their advantages in distribution and data, incumbents must show accelerating growth to restore investor confidence—and the outlook does not support that. Earnings growth for software and services companies in the S&P 500 is projected to slow to 14% in 2026 from about 19% in 2025, according to data compiled by Bloomberg Intelligence. In contrast, chipmakers are enjoying clearer growth visibility, driven by massive commitments from tech giants like Microsoft, Amazon, Alphabet, and Meta to build AI infrastructure. Semiconductor profits are expected to rise nearly 45% in 2025 and accelerate to 59% in 2026, according to Bloomberg Intelligence.
The market’s divergence reflects a broader shift in investor sentiment. “The reason chipmakers are outperforming is that their fundamentals are getting a lot better and there’s more certainty about their growth,” said Jonathan Cofsky, portfolio manager at Janus Henderson Investors. “At the same time, there’s a lot less certainty about how AI will change the software ecosystem.”
The valuation story is equally dramatic. The Morgan Stanley software basket now trades at 18 times projected earnings—its cheapest level on record—far below the decade-long average of more than 55 times. For years, software’s high multiples were justified by subscription models and recurring revenue streams that investors assumed could be extrapolated indefinitely. But as Wong explained, the rise of AI agents capable of working around the clock and completing complex projects in hours threatens to undermine that long-held assumption.
Still, not everyone believes the sector is doomed. Some Wall Street firms are beginning to argue that software could finally stabilize in 2026. Barclays expects a break in the sector as customer spending remains steady and valuations become increasingly attractive. Goldman Sachs suggests that AI adoption could expand software’s total addressable market, turning the technology from a threat into a tailwind. D.A. Davidson contends that the market’s narrative-driven fear has overwhelmed fundamentals, making this year a potential entry point for selective investors.
“We’re not in a position where we can say the turn is here,” said Chris Maxey, chief market strategist at Wealthspire, “since existential fears about AI will be here for a while, but the sector does look more interesting.” The message is cautious: software stocks are not yet a “screaming buy,” but for some investors, the worst of the selloff may be nearing its end. Whether that optimism is justified, however, will depend on whether incumbents can prove that AI will enrich their platforms rather than replace them.
