Which AI software catalysts are people watching next?

A primary catalyst reshaping the software landscape is a significant capital rotation driven by shifting enterprise priorities . CIOs are deprioritizing spending on traditional software, which has fallen on their to-do lists, in favor of AI tokens and infrastructure from foundation model providers like Anthropic and OpenAI [1, 2]. This reallocation is motivated by a perception of faster ROI from AI investments [1, 4]. The scale of this shift is substantial; in 2025, OpenAI and Anthropic alone were projected to add new run-rate revenue equivalent to **almost half** of that generated by the entire public software industry [15, 16, 21]. This trend has created major headwinds for legacy SaaS companies, whose long-term value is being questioned by investors [14, 18], and is fueling the vast majority of net new revenue growth in the software sector [28, 29].

Beyond the immediate spending shift, a fundamental architectural evolution is underway, centered on the rise of a "dynamic agent layer" [6, 12]. This represents a move away from passive "systems of record" toward AI agents that sit closer to the user, interpreting intent and executing tasks autonomously [6, 8]. This paradigm shift is not merely about making human workers more efficient but about replacing certain functions entirely, with businesses showing a preference for buying autonomous agents over hiring and training people . This new layer threatens the data gravity of established platforms by offering a **10x better user experience** and will likely consist of purpose-built agents for specific industries, which most enterprises will prefer to buy rather than build [6, 8]. This agent-centric future is also expected to activate CPUs for inference to a degree that may surprise observers focused solely on GPUs .

This disruption is bifurcating the software market, creating a complex environment for identifying future winners [7, 17]. While the emergence of AI-native challengers presents a direct threat to incumbent-dominated categories , the narrative of a "SaaSpocalypse" is not universally accepted . Some analysts believe it will be **at least a year** before investors can clearly identify which software companies will successfully navigate the transition . There is evidence that specialized incumbents can thrive by integrating AI to demonstrate clear value and re-accelerate growth, as seen in Figma's rebound . The central question for established players is whether AI will serve as a catalyst for renewed growth or a terminal threat , while some leaders maintain that AI will ultimately make the entire software industry larger, despite creating big losers among companies that fail to adapt .

The current market is characterized by unprecedented growth dynamics and an evolving value stack, representing what some see as the most significant platform shift since the internet . A new wave of AI software companies is achieving growth from zero to $100 million in revenue within one to two years, a velocity rarely seen in previous software eras [10, 11]. While some analysts believe value is accruing across all layers of the stack—from chips to applications—making zero-sum thinking a flawed strategy in this expansionary phase [9, 20], others posit that long-term value and defensibility will concentrate at the semiconductor and infrastructure layers, with the model layer becoming commoditized . This tension highlights the ongoing debate about the ultimate structure and profit pools of the AI-powered software industry.