FEATURED

The Cloud SaaS Monetization Reality Check

Something broke in enterprise software earlier this year, and the market still hasn’t fully sorted out what it means.

The IGV – the iShares Expanded Tech-Software Sector ETF – is down roughly 30% from its September 2025 peak. That’s approximately $2 trillion in market cap, gone. Software forward price-to-sales ratios compressed from around 9x to 6x, levels not seen since the mid-2010s. For the first time in the modern era, software stocks are trading at a discount to the broader S&P 500. That last part is worth sitting with. For two decades, investors paid a premium for software because the economics were genuinely superior – 70-80%+ gross margins, recurring revenue, near-zero marginal delivery cost. That premium is gone. The question now is whether it’s coming back.

What triggered it isn’t complicated, even if the implications are.

AI agents – autonomous software capable of running multi-step enterprise workflows without human involvement – are threatening the per-seat licensing model that built the SaaS industry. The math is simple and brutal: if an AI agent handles 40% of the tasks a 10,000-seat CRM platform was bought to support, the enterprise doesn’t need 10,000 seats anymore. Goldman Sachs has started calling the emerging alternative “Results-as-a-Service” – charging for outcomes rather than headcount. That framing is useful, but it also means the revenue model for most horizontal SaaS vendors is structurally uncertain in a way it never has been before. CIO surveys from early 2026 show roughly 40% of IT budgets being redirected away from traditional SaaS subscriptions toward agentic platforms and LLM infrastructure. That is not a marginal shift.

Slight tangent, but it matters: Salesforce served as the opening signal here. A rare revenue miss in late 2025, followed by weak guidance, triggered what some called the “Salesforce Contagion” – a rolling reassessment of whether even deeply entrenched CRM and ERP platforms are truly immune. They’re probably not. That doesn’t mean they’re worthless. It means the old multiple doesn’t apply.

Oracle is a different conversation entirely.

Q3 fiscal 2026 results, reported March 10, came in well above expectations across every major line. Total revenue hit $17.2 billion, up 22% year-over-year – the first quarter in over 15 years where both organic total revenue and non-GAAP EPS grew at 20% or more simultaneously. Cloud Infrastructure (IaaS) revenue came in at $4.9 billion, up 84% – accelerating from the 68% growth posted in Q2. Total cloud revenue (IaaS plus SaaS) reached $8.9 billion, up 44%. Those numbers beat analyst consensus of $16.9 billion on the top line and $8.84 billion in cloud. Oracle’s multi-cloud database revenue grew 531% year-over-year. AI infrastructure revenue inside OCI grew 243%. These are not estimates or forward projections – they are Q3 actuals.

The backlog is what makes this structurally different from an AI branding story.

Remaining Performance Obligations came in at $553 billion at the end of Q3, up 325% year-over-year. That figure grew from $523 billion in Q2 – which was itself up 438% from a year earlier, driven by new commitments from Meta, NVIDIA, and others. These are contracted, multi-year revenue obligations. They are not announcements or letters of intent. Oracle also confirmed plans to raise $45-$50 billion in capital expenditure this fiscal year to expand cloud infrastructure capacity, with over 10 gigawatts of computing power coming online over the next three years. To put that in context, management’s long-range OCI revenue trajectory targets $32 billion in fiscal 2027, scaling to $73 billion, then $114 billion, and reaching $144 billion annually by fiscal 2030.

What’s interesting is how different Oracle’s risk profile looks compared to the SaaS vendors getting sold. Enterprise database migrations – the kind running through Oracle’s sovereign cloud contracts with governments and large-scale enterprises – are not discretionary spending. They don’t get cut when a CIO starts rationalizing AI budgets. That stickiness is the structural argument. Oracle isn’t competing with AI agents. It’s providing the infrastructure those agents run on.

Here’s where I’m at on the risks, because they’re real.

Oracle reported $13.18 billion in negative free cash flow over the past 12 months. Renting out Nvidia GPU capacity generates lower margins than selling software licenses. GAAP EPS for Q3 came in at $1.27, compared to non-GAAP EPS of $1.79 – a gap that reflects the capital intensity of the build-out. Oracle acknowledged that demand for AI capacity currently exceeds supply, which is a good problem to have until it becomes a delivery problem. The company stated it delivered over 400 megawatts of capacity in Q3 and accelerated rack-to-revenue times by 60% – but executing a $144 billion five-year revenue target while managing that much capital deployment is an operational challenge of a different order. Investors should watch RPO conversion rates carefully as capex scales.

The broader SaaS selloff, meanwhile, has not been uniform. The deepest damage has landed on horizontal platforms – the tools that organize and surface information across business functions but don’t own the underlying data or the infrastructure layer. The platforms with genuine data moats, mission-critical workflow ownership, or consumption-based pricing models have held up relatively better. That distinction is the actual trade here, not a blanket view on software as a sector.

For active traders, the framework worth tracking: Oracle’s Q4 FY2026 guidance calls for revenue growth of 18-20% in constant currency, with non-GAAP EPS of $1.96 to $2.00. The stock surged as much as 10% in after-hours trading following the Q3 print, though it had been down approximately 23% year-to-date heading into that report versus the S&P 500’s near-flat performance over the same period. The key level to watch is whether IaaS acceleration continues into Q4 – if the 84% growth rate holds or expands, the backlog conversion story becomes significantly more compelling. If it starts decelerating, the capex commitment becomes a heavier topic.

The part people skip in all of this: the SaaS rout and Oracle’s outperformance are not separate events. They’re the same event. Capital is moving away from seat-based software toward infrastructure that doesn’t care how many humans are in the workflow. Oracle sits on the right side of that shift. Whether it can execute at the scale its backlog implies is the only question that matters from here.

For informational and educational purposes only. Not investment advice. Trading involves risk, including loss of principal.