Oracle has a problem. Its cloud applications business — the division selling human resources, finance, and supply chain software to enterprises — is growing, but not fast enough. Customers aren’t rushing to upgrade. Some are even eyeing competitors. And the company’s answer to this creeping stagnation is as old as the software industry itself: give something valuable away for free.
The something, in this case, is artificial intelligence.
According to The Information, Oracle has been embedding AI-powered features directly into its Fusion Cloud applications suite at no additional charge — a strategy designed to make its existing software stickier and more attractive at a moment when enterprise software vendors across the board are struggling to translate AI hype into real subscription revenue growth. The approach marks a sharp departure from competitors like Salesforce, which has tried to charge premiums for its AI add-ons, and it reflects a calculated bet by Larry Ellison’s company that the real money isn’t in AI features themselves but in the infrastructure underneath them.
The SaaS Growth Wall
Oracle’s cloud applications revenue has been decelerating. The company reported its fiscal Q3 2025 results in March, and while total cloud revenue hit $6.2 billion — up 23% year over year — the applications portion of that figure grew at a more modest pace compared to the infrastructure side. Cloud application revenue came in at $2.2 billion, up roughly 9%. Not terrible. But not the kind of number that excites Wall Street when the AI narrative demands acceleration.
The broader context matters. Enterprise software spending has cooled considerably since the post-pandemic boom. CFOs are scrutinizing renewals. Migration projects from on-premises systems to the cloud, which drove years of growth for Oracle, SAP, and others, are maturing. Many large enterprises that were going to move have already moved — or have decided not to. The easy wins are behind them.
Oracle isn’t alone in feeling this. Salesforce has seen its growth rate compress. Workday has faced similar headwinds. SAP, which is in the middle of pushing customers from its legacy ECC systems to S/4HANA Cloud, has managed to maintain momentum largely because of that forced migration cycle, but even it has acknowledged that AI needs to become a growth driver soon.
So Oracle made a choice. Rather than gate AI capabilities behind a premium tier — the approach Salesforce initially took with its Einstein GPT and later Agentforce products — Oracle decided to bake them in. Free. As part of the existing subscription.
The logic is straightforward: if customers are hesitant to spend more, don’t ask them to. Instead, make the product they’re already paying for dramatically better, reduce churn, and win competitive deals against rivals who are charging extra for similar capabilities.
It’s a land-and-expand play, but with a twist. The expansion Oracle is banking on isn’t in application seats. It’s in infrastructure consumption.
Here’s how it works. When Oracle adds AI agents and generative AI features into Fusion Cloud apps — things like automated invoice processing, intelligent recruiting assistants, AI-generated financial narratives — those features run on Oracle Cloud Infrastructure. Every AI inference, every model call, every data processing job consumes OCI compute resources. The more customers use AI-enhanced applications, the more OCI revenue Oracle generates. The applications become a funnel for infrastructure spend.
This is the same strategic logic that has made Oracle’s infrastructure business the fastest-growing part of the company. OCI revenue surged 49% year over year in the most recent quarter, driven heavily by demand for GPU compute from AI training workloads. Oracle has signed massive contracts with companies like OpenAI, xAI, and Cohere to provide cloud infrastructure for model training. Adding its own application customers to that demand pool only strengthens the flywheel.
The Competitive Calculus
Oracle’s free-AI strategy puts direct pressure on Salesforce, which has spent the last 18 months trying to build a premium AI revenue stream. Salesforce CEO Marc Benioff has championed Agentforce — the company’s AI agent platform — as the next major growth vector. But monetizing it has proven tricky. Salesforce initially priced AI features as add-ons, then adjusted its approach after customer pushback, bundling some capabilities into higher-tier plans while keeping others as paid extras.
The contrast is stark. Oracle says: AI is included. Salesforce says: AI costs more. For enterprise buyers evaluating their options, that pricing difference can tip decisions, especially when budgets are tight.
But the comparison isn’t perfectly clean. Salesforce’s AI ambitions extend well beyond embedded features — Agentforce is designed to be a platform for building autonomous AI agents that can handle complex multi-step business processes. Oracle’s current AI offerings in Fusion Cloud are more narrowly scoped, focused on enhancing existing workflows rather than creating entirely new ones. The depth of capability differs even if the marketing language sounds similar.
SAP presents another competitive angle. The German software giant has been integrating its Joule AI assistant across its cloud applications and, like Oracle, has largely included AI features within existing subscriptions rather than charging separately. SAP’s advantage is the sheer breadth of its installed base — it runs the back-office operations of a majority of the world’s largest companies. But SAP’s cloud transition is still incomplete, and many of its customers remain on older on-premises versions that can’t access AI features at all. That creates an opening Oracle is eager to exploit.
Microsoft, meanwhile, operates on a different plane entirely. Its Copilot AI features for Dynamics 365 — Microsoft’s ERP and CRM platform — come with additional per-user monthly fees, ranging from $20 to $30 depending on the application. Microsoft can get away with this partly because of its dominant position in enterprise productivity software and partly because Copilot is deeply integrated with the Microsoft 365 tools that virtually every knowledge worker already uses. But even Microsoft has faced slower-than-expected Copilot adoption, suggesting that enterprises aren’t eager to pay AI premiums regardless of the vendor.
Oracle’s bet is that the premium model is a dead end for application-layer AI. That the real monetization happens below the surface.
There are risks. Giving away AI features compresses application margins in the short term. If OCI consumption from application workloads doesn’t scale fast enough, Oracle could find itself subsidizing AI capabilities without a clear payoff. And competitors could match the strategy — Salesforce could decide to bundle AI into base subscriptions, SAP could accelerate its own inclusion efforts — neutralizing Oracle’s differentiation.
There’s also the question of whether Oracle’s AI features are actually good enough to matter. Enterprise buyers are increasingly sophisticated about AI. They’ve seen the demos. They’ve heard the pitches. What they want is measurable productivity improvement, not just a chatbot bolted onto an existing interface. Oracle has been investing heavily in AI R&D, and its applications team has been working on domain-specific models trained on enterprise data patterns, but the proof will be in adoption metrics and customer retention numbers over the coming quarters.
Larry Ellison has been vocal about Oracle’s AI ambitions on recent earnings calls, framing the company as uniquely positioned because it controls both the application layer and the infrastructure layer. “We have a complete stack,” he told analysts in March. That vertical integration — database, middleware, applications, and cloud infrastructure all under one roof — is Oracle’s structural advantage. It allows the company to optimize AI workloads in ways that a pure-play application vendor like Salesforce or a pure-play infrastructure provider like AWS cannot.
Whether that architectural advantage translates into market share gains remains an open question. Oracle’s cloud applications business still trails Salesforce in CRM and competes fiercely with SAP and Workday in ERP and HCM. Free AI features might help Oracle hold existing customers and win some competitive deals at the margin. But the enterprise software market is notoriously sticky — companies don’t switch ERP systems because of a chatbot, no matter how good it is.
What Wall Street Is Watching
Investors have largely rewarded Oracle’s infrastructure story. The stock is up significantly over the past year, driven by OCI’s growth and the company’s massive remaining performance obligations — essentially, contracted future revenue — which topped $130 billion in the most recent quarter. The applications business, by contrast, has been a secondary narrative. Solid but unspectacular.
The free-AI strategy could change that calculus if it accelerates application revenue growth or, more importantly, if it drives measurable increases in OCI consumption from application workloads. Oracle hasn’t broken out how much OCI revenue comes from its own applications versus external customers, and analysts have been pressing for more granularity. That disclosure — or lack thereof — will be a key thing to watch in upcoming earnings reports.
The broader industry implication is significant. If Oracle’s approach works — if giving away AI features at the application layer while monetizing at the infrastructure layer proves to be the winning model — it could force a strategic rethink across the enterprise software industry. Salesforce, which doesn’t own its own hyperscale cloud infrastructure, would be particularly exposed. The company runs on AWS and Google Cloud, meaning every AI inference it processes generates revenue for a competitor, not for itself. That structural disadvantage becomes more acute in a world where AI features are expected to be free.
And that may be the most consequential aspect of Oracle’s move. Not the immediate competitive impact, but the signal it sends about where value accrues in an AI-driven enterprise software market. The application layer is becoming a commodity — or at least, the AI enhancements to it are. The infrastructure layer is where the margin lives.
Oracle has been saying this for years. Now it’s putting real money behind the thesis. The rest of the industry is watching to see if the bet pays off — or if free AI turns out to be worth exactly what customers pay for it.
WebProNews is an iEntry Publication