By Carrie Osman, CEO and Founder of Cruxy.
The cracks in enterprise software’s most sacred pricing model are widening, and private equity is starting to feel the ground shift beneath its feet.
For years, seat-based licensing was the bedrock of SaaS economics. Sign up enough users, hit breakeven, and watch new revenue flow straight to the bottom line. It was the model that made software companies command the multiples they once did. But AI is quietly dismantling that logic and the funds that piled into software in the late 2010s are now scrambling to understand what they actually own.
The warning signs are there. When news broke about Claude’s COBOL modernisation capabilities, IBM’s share price tumbled – a stark reminder that even entrenched enterprise vendors are not immune. If a company of IBM’s scale and legacy can face that kind of repricing, the question for PE-backed SaaS firms with far shallower moats is obvious: what happens to them?
The answer has two parts and both are uncomfortable.
The first is the direct disruption threat. Many software businesses owe their defensibility not just to their product, but to the proprietary data sitting beneath it and the workflows locked around it. The real question investors need to ask is: how realistic is it that a customer could ‘vibe-code’ their own alternative? If the answer is “increasingly plausible,” that is a moat problem. And yet too few investors are interrogating products at this level before committing capital.
The second threat is slower-burning but arguably more structural: the decline of seat-based pricing. As agentic AI automates workflows, a single-seat licence can now do the work that once required five. This opens up a new world of pricing logic to underpin SaaS valuations. Meanwhile, in a token-dependent AI product, every new customer generates a proportional increase in API costs. The cost curve no longer flattens. Gross margins that once made SaaS magical start to look a lot more ordinary.
This is already reshaping how some companies operate. Datarails has just become the master of its own destiny by launching a new product which invites collaboration with third-party AI systems. Elsewhere, ServiceNow and others are moving to hybrid pricing models – part seat, part consumption – in recognition that the old model won’t hold. The shift is not yet existential, but it will steadily erode growth and margins across the asset class.
For PE funds, the exposure is real and, in many cases, poorly understood. Portfolios built on the promise of SaaS scalability are now facing a structural repricing – not just of individual assets, but of the whole model. Private equity fund Apax’s decision to walk away from its Pinewood Technologies acquisition reflects a broader anxiety: that the competitive moats underpinning software valuations are harder to assess than they appeared, and that the market has not yet adequately priced this risk.
Funds need to go back to basics on product due diligence – interrogating not just what software does, but what data it holds, how those APIs are priced (many remain dramatically under-monetised), and whether the workflow is defensible enough to survive AI-enabled disruption. Some are already doing this: Hg’s Catalyst initiative is one example of a fund proactively helping portfolio companies stress-test their positioning.
The seat-based model isn’t in existential danger yet. But for enterprise SaaS businesses that fail to make the transition to consumption or value-based pricing, AI won’t just threaten their growth. It will price them out of the market entirely.