What are people saying about AI's threat to SaaS pricing?

A consensus is forming that AI poses a significant threat to traditional SaaS pricing, particularly the per-seat, per-month model [3, 9]. This model is seen as a strategic vulnerability because AI agents and automation are expected to reduce the number of human employees required for many tasks, thereby directly cannibalizing the revenue base of companies priced on human user counts [23, 26]. The business model for customer support software like Zendesk, for instance, is described by one expert as **becoming an extinct business model** due to AI's ability to handle the majority of queries . This pressure extends to any SaaS product where pricing is tied directly to user productivity that can be automated, such as Salesforce [1, 2]. While some analysts believe all traditional SaaS companies are under threat [8, 11], others argue the "SaaSpocalypse" is an oversimplification, with the impact varying based on the company's position .

In response to these pressures, SaaS companies are exploring more resilient pricing structures. Per-employee models for systems of record like Workday are considered more defensible because they are not directly tied to a specific task or outcome that AI can automate away [1, 10]. Many see a necessary shift toward consumption or usage-based pricing, which better aligns with the high and variable Cost of Goods Sold (COGS) associated with running AI models [13, 15, 29]. However, this approach faces a significant hurdle, as **pure consumption-based billing is often met with enterprise resistance** due to the customer's need for budget predictability [1, 3, 19]. This has led to a growing belief that the future lies in blended or hybrid models that combine a stable per-seat fee with a consumption-based component to account for AI costs and value delivery [16, 20, 28]. The challenge remains that the value created by AI does not fit neatly into any single traditional pricing structure .

Beyond pricing models, AI threatens the fundamental economics and defensibility of the SaaS industry. A primary concern is that AI will dramatically lower software creation and data switching costs, eroding the protective moats of large incumbents like ERPs and ServiceNow [2, 6, 27]. This commoditization could be accelerated by AI agents absorbing the user interaction layer, further reducing customer lock-in . This trend, combined with high COGS from inference and lower barriers to entry, puts pressure on historically high SaaS margins [17, 22]. The potential fallout is severe valuation compression, with some analysts predicting price-to-sales multiples could fall **from 5-10x to 1-2x** as the utility-like nature of these businesses is exposed . The threat is considered so significant by some that AI models are seen as capable of rewriting an incumbent's entire business within three to four years .

However, the narrative of an imminent SaaS collapse is contested. Anish Akaya of Andreessen Horowitz calls the "SaaSpocalypse" narrative incorrect, citing evidence that **75% of public SaaS companies** have successfully raised prices since the release of ChatGPT, with many increasing prices by 25% or more [7, 24, 21]. This perspective suggests that while AI will dramatically reduce switching costs, it may not wholesale replace existing tools but rather create new categories . A key point of disagreement exists around systems of record; while some view them as resilient [1, 10], Klarna's CEO Sebastian Siemiatkowski argues they are existentially threatened by AI's ability to commoditize their function [2, 4]. Similarly, Atlassian's CEO views AI as a positive strategic development that will increase platform value and stickiness, a direct contrast to the more bearish outlooks causing investors to exit the sector [1, 12].