NextFin News - The era of "seat-based" software dominance is facing its most severe structural test as artificial intelligence begins to decouple productivity from headcount. In the first quarter of 2026, the iShares Expanded Tech-Software Sector ETF (IGV) has plunged nearly 22%, reflecting a deepening skepticism among investors that traditional enterprise giants can maintain their margins in a world where AI agents, rather than human employees, perform the bulk of administrative tasks.
The market's anxiety is visible in the valuations of the industry’s standard-bearers. Since the start of the year, Microsoft has seen its stock drop more than 21%, while Salesforce and Workday have retreated 26% and 36% respectively. This sell-off, colloquially dubbed the "SaaSpocalypse" by some Silicon Valley insiders, was accelerated in February when Anthropic released its Claude Cowork agent, a tool capable of automating complex business workflows that previously required multiple software subscriptions and human oversight.
Rishi Jaluria, a senior software analyst at RBC Capital Markets, has emerged as a prominent voice of caution regarding this transition. Jaluria, who has historically maintained a balanced view of the sector but recently shifted toward a "show-me" stance, argues that the market is currently struggling to price a transformation that may take five to ten years to fully materialize. According to Jaluria, while these companies are still reporting stable headline numbers, the lack of accelerating growth suggests that AI monetization has yet to offset the potential erosion of the traditional per-user licensing model.
Jaluria’s cautious outlook is not yet a universal consensus. At a recent UBS conference, Rajesh Jha, a high-ranking executive at Microsoft, countered the narrative of terminal decline. Jha argued that AI is not eroding the "seat" count but rather evolving it; in his view, AI agents may eventually be treated as digital employees, effectively becoming "new seats" that require their own licenses. This perspective suggests that the total addressable market for software could expand as businesses deploy thousands of autonomous agents to handle tasks that were previously too expensive or complex for human staff.
The tension between these two views is forcing a radical shift in how software is sold. For decades, the industry relied on charging per user per month. However, market intelligence firm IDC now predicts that pure seat-based pricing will be largely obsolete by 2028. In its place, a "consumption-based" or "outcome-based" model is emerging. Salesforce is already pivoting with its "Agentforce" platform, which allows customers to build and deploy AI agents, while Microsoft has introduced a high-end E7 enterprise tier priced at $99 per seat to capture the premium value of its AI capabilities.
Despite the aggressive product rollouts, the threat of "vibe coding"—where non-technical employees use AI to build custom internal tools—remains a wildcard. If a mid-sized firm can use an LLM to build a bespoke CRM or HR portal in a weekend, the value proposition of a $150-per-month Salesforce license diminishes. One Microsoft salesperson described a recent meeting where a CTO bluntly asked why they should continue paying for a platform they could now "just build" themselves using internal AI resources.
However, the "build-it-yourself" movement faces significant hurdles in security and integration. A Salesforce employee recently noted that while building a simple app is easy, ensuring that app complies with global data privacy laws and integrates with a dozen other legacy systems is a different matter entirely. For many large enterprises, the cost of a subscription is as much about risk mitigation and "plumbing" as it is about the software's features. Asana CEO Dan Rogers has echoed this sentiment, suggesting that as companies add more AI tools, the need for a central "coordination layer" will actually grow exponentially rather than disappear.
The immediate future for Big Software depends on whether these companies can transition their pricing models fast enough to catch the falling knife of seat-count contraction. While the "SaaSpocalypse" narrative has successfully reset valuations, the actual displacement of enterprise software remains more of a long-term threat than a current reality. For now, the industry is caught in a painful middle ground: the old world of predictable per-user revenue is fading, but the new world of agent-driven growth has yet to prove it can deliver the same legendary margins.
Explore more exclusive insights at nextfin.ai.
