IBM’s second-quarter results showed a sharp decline in mainframe-related business, highlighting changes in the SaaS revenue model in the AI era. As debate over the end of SaaS spread through overseas markets in early 2026, the sell-off also reached Japanese stocks, with SaaS names falling sharply on the same day.
In IBM’s second-quarter results, hybrid cloud (Red Hat) remained solid, up 11%, while IBM Z fell 42%. Consulting remained broadly flat, illustrating how infrastructure demand in the AI era is driving growth while legacy assets are weighing on earnings. The earnings briefing did not indicate that AI agents had directly caused cancellations; instead, it presented the view that the trough of the mainframe product cycle had been a factor.
In February 2026, the phenomenon dubbed “SaaSpocalypse,” referring to the sharp decline in SaaS stocks, was reported, and Japanese names including ラクス, Sansan, 弁護士ドットコム and freee also fell that day. The declines were attributed to a broad sell-off in SaaS stocks overseas, with sector-wide movements taking precedence over the fundamentals of individual names.
The view that the traditional seat-based SaaS pricing model could become a drag on revenue as AI agents gain wider adoption has spread. At the same time, the advance of in-house development and productivity gains from AI agents are expected to alter demand in the SaaS market, leading to the view that structural change is unavoidable. Nomura Securities’ view is that in-house development remains limited, while the replacement of employees by AI agents is a realistic concern.
From Chapter 3 onward, the performance of major US SaaS companies is examined through figures for ServiceNow, Salesforce and Adobe, organizing the fact that revenue continued to grow even as their share prices fell sharply. ServiceNow continued to post growth in RPO and cRPO, which indicate a high degree of contract duration. Salesforce announced the introduction of the new AWU metric and raised its growth outlook. Adobe also maintained double-digit growth, but its share price reaction was muted. Together, these cases suggest that the declines were driven not by an “earnings collapse” but by changes in valuation multiples.