SaaSpocalypse Explained: Why enterprises are suddenly worried about the future of SaaS

The idea that software-as-a-service (SaaS) may be facing an existential crisis—popularly labelled the “SaaSpocalypse”—is no longer confined to enterprise IT discussions. It has begun to surface in public markets, executive commentary and investor notes, raising concerns about how artificial intelligence (AI) could reshape the software industry’s profit pool.

That unease intensified after Anthropic unveiled its Claude “coworker” agent, now armed with task-automating plugins that allow it to execute workflows autonomously. The shift marks a subtle but profound change—from AI as a feature embedded within software, to AI as a direct competitor to software vendors and IT services firms themselves. The anxiety is not about declining demand for software. Instead, it reflects a deeper reassessment of how companies buy, deploy and extract value from digital tools. What is being questioned is the economic logic that powered the SaaS boom for more than a decade: recurring per-user licences, expanding application portfolios and interface-heavy workflows?

The market reaction was swift, with global tech titans and Indian IT majors seeing a sharp decline in their market shares last week, with  Jeffrey Favuzza of Jefferies describing the event as a “SaaSpocalypse” in comments to Bloomberg, calling it nothing less than “an apocalypse for software-as-a-service stocks.”

While market movements are often driven by sentiment, they reflect a broader concern: whether AI could compress the pricing power that has long underpinned SaaS valuations.

What’s driving the anxiety?

At the heart of this reassessment is generative AI and the rapid rise of AI agents. Together, they are challenging a long-held assumption of enterprise IT—that every business problem requires a dedicated SaaS application with its own interface, workflow and licensing model.

For years, SaaS thrived by offering specialised tools for narrowly defined functions such as customer relationship management, expense processing, HR queries, IT ticketing and analytics dashboards. Today, large language models and AI agents can increasingly perform many of these tasks directly, often without users needing to log into multiple software products.

Instead of navigating dashboards and menus, employees can now interact with enterprise systems via natural-language prompts. An AI agent can pull data from multiple sources, apply business rules and execute actions across systems—replicating workflows that were previously embedded inside SaaS applications. This shift is forcing enterprises to confront an uncomfortable question: are they paying for software, or for outcomes?

Why SaaS costs are under sharper scrutiny

Cost pressures are sharpening this debate. CIOs and CFOs point to three persistent challenges. First, licence sprawl, with dozens or even hundreds of SaaS tools accumulated over the years, often through decentralised departmental purchases. Second, uneven utilisation, particularly in per-user subscription models, where actual usage rarely matches licence counts. Third, rising investments in AI infrastructure and models, which are now competing directly with SaaS renewal budgets.

In an environment of tighter IT spending, SaaS subscriptions—once treated as non-negotiable—are being scrutinised line by line. Renewals are no longer automatic. In many cases, they are being postponed, downsized or renegotiated, especially where AI-led alternatives appear viable.

In a publicly shared post, Zoho CEO Sridhar Vembu said SaaS had leaned too heavily on sales-led growth and ever-expanding enterprise budgets, describing AI as “the pin that is popping this inflated balloon.” His critique reflects a growing view that the SaaS model, rather than individual products, is being stress-tested.

Which parts of SaaS are actually at risk?

Is SaaS truly under threat? The answer is more nuanced than the term “SaaSpocalypse” suggests. SaaS as a category is not disappearing, but its growth model is under pressure.

Core systems of record—enterprise resource planning, core finance, payroll and cloud infrastructure—remain deeply entrenched. Replacing them is risky, expensive and rarely justified. These platforms continue to anchor enterprise operations and data integrity.

The real pressure is on workflow-heavy and interface-driven applications—tools that primarily exist to move data between systems, trigger approvals or surface dashboards. These are precisely the areas where AI agents are proving most effective.

A redistribution, not a replacement

Much of the current anxiety stems from framing AI as a replacement for enterprise software rather than a reallocation of responsibility. Ritesh Kapadia, field chief technology officer at iLink Digital, argues that every major technology transition tends to follow this pattern.

“Every major shift in enterprise technology is often framed as an ending, but it is usually a redistribution of responsibility. The current ‘SaaS apocalypse’ narrative fits that pattern,” Kapadia says. While AI agents are increasingly taking on coordination, orchestration and decision support, accountability for reliability, security and scale, he notes, still firmly sits with enterprise software.

This distinction is critical. AI does not reduce the importance of strong software foundations—it heightens it. “AI only works when the foundations are strong. Data integrity, cloud reliability, security, compliance, and disciplined integration remain non-negotiable and become even more critical as AI takes on greater operational roles,” Kapadia adds.

In this emerging model, SaaS does not disappear. It becomes less visible at the surface, but more important as the stable backbone that allows AI-driven workflows to operate reliably across the organisation.

How large enterprises are responding

Most large organisations are not ripping out SaaS overnight. Instead, they are adopting a phased approach. This includes consolidating vendors, rationalising overlapping tools, replacing niche SaaS products with internal AI-led platforms, and shifting procurement conversations from licence counts to measurable business outcomes.

The shift has also revived interest in low-licence or “zero-licence” models, where enterprises aim to minimise per-user subscriptions while building automation layers on top of existing systems. The emphasis is moving away from feature abundance towards flexibility, control and orchestration.

What this means for Indian IT services firms

For Indian IT services providers, the SaaSpocalypse narrative presents both risk and opportunity. Slower growth in SaaS spending could weigh on implementation, integration and partner revenues. At the same time, enterprises need extensive support to redesign workflows, integrate AI agents, govern data and ensure security and compliance.

Much of the work is shifting from configuring SaaS platforms to re-architecting business processes around AI—an area that plays to the strengths of large services firms with deep domain expertise, strong engineering talent and long-standing enterprise relationships.

Pressure on SaaS vendors to evolve

For SaaS vendors, incremental AI features will not be enough. They will need to embed AI deeply into their products, rethink pricing models and demonstrate that their software delivers outcomes that AI agents alone cannot.

Interface-heavy products risk being marginalised. Those that evolve into intelligent, adaptive platforms may emerge stronger.

The bottom line

The SaaSpocalypse is not about the death of SaaS, but a reset in enterprise software economics. SaaS is no longer the default choice. It must now compete with AI-driven alternatives that promise similar outcomes at lower marginal cost. For enterprises, the shift marks a return to first principles: paying for results, not interfaces.