Oracle CEO says token-fee model will help companies manage AI costs. What that actually means for ORCL stock.

Oracle (ORCL) is changing the way companies pay for artificial intelligence, and the company’s CEO says the new approach will put customers back in control.

During Oracle’s Q4 2026 earnings call on Wednesday, CEO Mike Sicilia outlined the transition to token-based and outcome-based pricing for AI capabilities across the software suite. The goal, he said, is simple: to help customers spend their money smarter on artificial intelligence without being shocked.

“All of this helps customers manage costs and align spending with the value created,” Sicilia said on the conference call.

How Oracle’s new AI pricing actually works

Oracle has introduced two new pricing tracks in addition to traditional software subscriptions.

The first is the token bundle model. Customers can purchase bundles of tokens that work across Oracle’s application suite, essentially allowing them to pre-purchase AI capacity in easy and predictable packages. Think of it like buying a data plan. Consumers know what they are getting and how much it will cost.

The second is outcome-based pricing. Customers pay based on what the AI ​​actually delivers, rather than a flat fee. Oracle gave several examples on the conference call.

• The interviewing agent sets a fee per screened candidate, or
• Hospitality Upsell Tools. Prices are set as part of the transactions that are generated.

Cicilia said Oracle has been testing outcomes-based models for some time in sectors such as construction, health care and hospitality. But the company is now rolling out these structures more broadly across its Fusion applications and industry-specific products.

“We’re simplifying the way customers use agent capabilities and pay,” he said. “Our new agency pricing is aligned with customer value.”

In the quarter ended May 31, 33 customers pre-purchased token bundles to unlock access to advanced AI inference and models. Aon Service Corporation (AON) and Liberty Energy (LBRT) were also among the companies mentioned on the conference call.

Sicilia was clear that many of Oracle’s AI capabilities in core applications will continue to be included at no additional charge. Token bundles are more specifically for customers who want access to larger and more capable inference models.

Why this matters: broader industry issues

Oracle’s shift in pricing comes at a time when the cost of enterprise software is rising unexpectedly.

New AI monetization strategies and pay-as-you-go pricing are the biggest drivers of software-as-a-service (SaaS) cost increases, surpassing even application sprawl, according to a January report from software management company Zylo. CFO Dive. Organizations are experiencing unbudgeted charges as vendors pile AI add-ons onto existing subscription agreements.

It’s exactly this volatility that Oracle is trying to avoid. The company believes that by packaging AI capabilities into predictable bundles or directly tying costs to measurable outcomes, customers can more easily justify and plan their AI budgets.

Sicilia pointed to Oracle’s full-stack position as a key advantage here. Because Oracle runs the infrastructure, database, and applications, we can actually measure what AI is doing for our customers.

Record quarter behind price news

The price announcement was made following a positive financial report. Oracle had total revenue of $19.2 billion in its fiscal fourth quarter, an increase of 21% year-over-year (YOY). Cloud revenue increased 47% to $9.9 billion, driven by 93% growth in cloud infrastructure, reflecting the surge in demand for AI workloads.

For the full year, Oracle’s revenue exceeded $67 billion for the first time. Remaining performance obligations, or future contract revenue, increased 363% year over year to $638 billion.

New Chief Financial Officer Hilary Maxson, who joined Oracle two months ago, acknowledged that gross margins will face pressure in the near term as the company ramps up data center capacity. But she was optimistic about its trajectory.

“We expect infrastructure margins to improve rapidly as our data centers reach full contracted revenue levels,” he said on the conference call.

For fiscal year 2027, Oracle expects total revenue to increase approximately 34% at constant currency, excluding one-time investment gains recorded in fiscal year 2026, and non-GAAP earnings per share to increase 18% at constant currency to $8.05.

What it means for companies using Oracle

The pricing changes give enterprise customers more flexibility, but they also signal the direction in which enterprise software is headed more broadly.

The days of flat-rate, per-seat licenses for AI tools appear to be coming to an end. Vendors across the industry are considering how to charge for features that don’t scale neatly to employee headcount.

Oracle’s approach, transparent token bundling and outcome-linked pricing, is one of the more structured answers to that challenge so far.

For companies already deeply involved in the Oracle ecosystem, the message from Wednesday’s conference call is that more AI capabilities will be built into existing subscriptions. But for those who want to go further, there is now a clear, if novel, path to doing so without signing a blank check.

Is ORCL stock undervalued?

ORCL stock is down 46.7% from its all-time high, and adjusted earnings per share are expected to expand from $7.63 in fiscal 2026 (ending May) to $23.33 in fiscal 2031.

If the tech stock’s forward P/E ratio is set at 22.92x, similar to its current multiple, the stock could nearly triple over the next four years.

Of the 43 analysts covering Oracle stock, 33 have recommended a “strong buy,” one has recommended a “moderate buy,” eight have recommended a “hold,” and one has recommended a “strong sell.” ORCL’s average price target is $257.30, implying an upside of 39.7% from there.

On the date of publication, Aditya Raghunath did not have any positions (directly or indirectly) in any securities mentioned in this article. All information and data in this article is for informational purposes only. For more information, please see the Barchart Disclosure Policy here.