Is AI Disrupting the SaaS Industry? A Deep Dive into Market Shifts
In a world where generative AI is increasingly capable of performing tasks autonomously, rather than merely assisting humans, the future of Software as a Service (SaaS) is under intense scrutiny. Over recent weeks, investors have shed a trillion dollars in value across various software sectors, prompting a reevaluation of the terminal value of classic SaaS models. This has led to a dramatic repricing of public software, with average multiples now around 6x EV/LTM revenue, a stark contrast to the 20x seen in 2021, representing a 70% compression. Horizontal software has been hit hardest, compressing to approximately 3x multiples.
The Historical Success of SaaS
Since 2007, investments in enterprise software have grown into a portfolio of nearly 100 companies, highlighting SaaS as a fantastic asset class. For two decades, it grew at roughly 20% year-on-year, driven by attractive economics such as low churn, high gross margins, and the ability to scale with customers. However, growth has slowed to around 12%, an eight-percentage-point decline that began even before recent AI catalysts made headlines.
The ROI case for SaaS was compelling: instead of investing in physical servers and high capex, businesses could rent services, amortise costs, and deploy modern tools rapidly. This reduced friction in adopting better software, democratised information across organisations, and enhanced collaboration, theoretically boosting productivity and strategic focus. With such clear benefits, SaaS became a no-brainer, spawning an army of companies and accounting for around $1 trillion in enterprise software spend last year alone.
The Hidden Drivers of SaaS Growth
A dirty secret of SaaS growth lies in its contractual nature. Long-term contracts with price uplifts and escalators, often ranging from 3-5% annually—and up to 10% for top companies—provided a base layer of expansion. This meant companies didn't need to find all growth from scratch each year; they could rely on contracted expansion and then seek additional growth through broader deployment, selling extra capabilities, and adding new customers.
For many categories, expansion was seat-based, but this became challenging once peak internal penetration was reached, especially with Fortune 500 companies showing less than 1% employee growth in recent years. Companies optimised go-to-market strategies, but this didn't magically expand the market. Meanwhile, shifts in interest rates led CFOs to scrutinise software budgets, moving away from growth-at-all-costs mentalities.
Generative AI's Challenge to SaaS
Generative AI is now challenging the top line of even the best SaaS companies, not by rendering SaaS obsolete, but by questioning its compounding mechanisms. While AI-native companies, including hardware winners like Nvidia, are experiencing rapid growth, enterprise budgets have not expanded meaningfully. Early ROI for generative AI in enterprises remains limited, as evidenced by public market leaders like Datadog and Servicenow, which have not seen significant topline expansion despite rolling out new products aggressively.
For instance, Datadog reports that 12% of its revenue comes from AI-native customers, growing 17% quarter-on-quarter, while Servicenow's AI product, Now Assist, has reached $600 million in annual recurring revenue. However, AI revenues remain small compared to broader revenues at many incumbents, making it difficult to see immediate top-line growth. In practice, enterprise budgets haven't grown enough to fund everything simultaneously, leading to marginal decisions: expansion seats are questioned, nice-to-have tools are consolidated, and new multi-year commitments are delayed as buyers await AI's potential.
Cyclical or Structural Change?
The key question is whether this slowdown is cyclical or represents a structural change. There is no doubt that generative AI will become pervasive in enterprises, but the pace of transformation and the unit of value that software monetises are uncertain. SaaS pricing traditionally assumed software scales with headcount, but AI introduces a credible scenario where value scales without headcount.
This doesn't guarantee the end of elite software businesses. Part of the current slowdown is likely cyclical, with enterprise buyers cautious about committing to new SaaS spend due to fears of legacy technology. However, this caution builds technical debt through deferred upgrades and ageing integrations. History shows enterprise transitions are slow—the move from mainframe to SaaS took 25 years and is ongoing, with IBM mainframes still generating billions in revenue. Elite software companies have time to adapt but must move quickly.
Early data from companies like Datadog and Servicenow suggests that incumbents who act fast can absorb AI rather than be disrupted by it. They need to aggressively pursue AI opportunities, understand if their solutions will remain relevant in an agentic world, and land AI-native customers, as the Fortune 500 landscape may shift in five years.
Where AI Will Disrupt vs. Enhance
AI will genuinely disrupt categories where good SaaS alternatives never fully emerged. Software development is a prime example, with AI coding assistants reshaping code writing and review processes. Customer service is another area, where AI agents can handle interactions that previously required both software and headcount. In these greenfield categories, AI isn't competing with entrenched SaaS but filling gaps that SaaS never closed.
The winners in this new era will be companies that tackle hard problems their customers cannot solve independently and tie that capability to measurable value. SaaS is not dead, but the game has changed. The companies that recognise and adapt to this reality fastest will define the next decade of enterprise software.
