UpTrajectory Review
Oracle's recent financial results highlight a significant surge in cloud infrastructure revenue, driven by the increasing demand for AI capabilities among businesses. The company's commitment to invest over $70 billion next year reflects a strategic pivot to meet this demand, particularly as customers seek to leverage AI for productivity and competitive advantage. This shift includes a pilot program for outcome-based AI billing, which aims to align pricing with the value delivered to businesses.
For small business owners, Oracle's move towards outcome-based AI billing could signal a transformative approach to budgeting for AI services. This model may alleviate concerns about unpredictable costs associated with traditional token-based pricing, allowing operators to better plan their investments in AI technology. However, it's essential to remain cautious; as analysts suggest, while the interface may appear friendlier, the underlying token economics are still in play. Operators should closely monitor how this pricing model evolves and consider its implications for their own budgeting strategies.
“Oracle’s move towards outcome-based AI billing should not be misread as the death of token economics.” — CIO Magazine
Takeaway: Watch for how outcome-based AI billing can simplify budgeting for AI investments in your business.
From the original item — CIO Magazine:
Cloud revenues are booming, as are infrastructure investment costs at Oracle, this week’s release of the company’s fourth quarter financial results revealed.
During a conference call with analysts, newly appointed CFO Hilary Maxson said that cloud infrastructure revenue grew 93% year-over-year, which she said reflected demand for AI workloads and database services. The company, therefore, plans to invest upward of $70 billion in capital expenditures next year, driven by committed customer demand.
Those customers, said CEO Mike Sicilia, “are now focused on how to leverage AI in their own businesses. They want AI to increase productivity, enhance customer service, and create real competitive advantages. So they want to do it quickly, and within their existing budget envelope.”
They are ready to implement enterprise-ready agentic offerings that help run their business, he added, noting, “Over the past year, we have delivered more than 1,000 AI agents across our application suites. These agenda-based offerings can reason, decide, and execute work across processes.”
All of that activity has resulted in implementation of a pilot program that Sicilia said will “align pricing directly to the value derived.” Launched this past quarter, he described it as a “limited rollout” involving 33 organizations.
Sanchit Vir Gogia, chief analyst at Greyhound Research, said, “Oracle’s move towards outcome-based AI billing should not be misread as the death of token economics. Tokens are not disappearing; they are being hidden behind a friendlier commercial interface. The meter is still running somewhere.”
However, he added, the logic is sound, “because token pricing is a dreadful language for enterprise budgeting. Nobody wants to build a strategic AI roadmap around a taxi meter with a PhD. Independent research shows agentic workloads consuming up to a thousand times more tokens than simple tasks, with identical runs varying as much as thirtyfold.”
Oracle’s advantage, said Gogia, “is proximity to the system of record, which lets it define outcomes more credibly than vendors who see only prompts and outputs. The danger begins precisely there; outcome pricing sounds cleaner until the vendor becomes both the supplier and the referee.”
This is an industry-wide repackaging, and the destination is hybrid pricing, he observed. “Tokens will remain in the plumbing even after they vanish from the invoice. The right question is not whether token complexity has gone, but where it has been buried.”
Scott Bickley, advisory fellow at Info-Tech Research Group, added that this type of model, “is where I think ultimately CIOs will want to go, because it’s much easier to define what you are paying for and the business value.”
Right now, he said, “everyone is struggling with variable consumption licensing models, and tokens are essentially a black box, very hard to decipher and figure out what you’re actually buying. They’re also moving targets. You have different underlying mechanics for how the models consume tokens.”
There are, said Bickley, a lot of variables that go into how many tokens are going to be consumed for a particular query or prompt or action. Vendors at the SaaS layer “have been trying to abstract tokens into their own version of AI work units, or AI credits, or AI actions, or Agentforce actions, whatever you want to call them.”
But that abstraction, said Bickley, “still requires you to understand the underlying mechanics of how consumption is being used to deliver an action or an outcome. If you can totally abstract that pain away and give me an outcome for a price, then you’re well ahead of the majority of the pack.”
Gogia added, “the most important thing CIOs need to know about Oracle’s latest earnings is that the company has crossed a strategic threshold. This is no longer a software company talking about AI. It is becoming an industrial-scale AI infrastructure company, with the debt and the execution risk that such a shift implies.”
He also advised CIOs to separate construction momentum from execution certainty, “because conflating the two would be naïve. The real bottleneck in AI is no longer GPUs. It is power, permits, and politics.”
In addition, the angle they should not miss is that these earnings are not only about AI infrastructure. “They are equally about Oracle’s attempt to make itself harder to avoid in a multicloud enterprise world, and the database numbers carry that story,” he noted.