What Are AI Tokens? What NetSuite’s New AI Units Reveal About Enterprise AI Pricing

When Oracle announced NetSuite AI Units, the first question most people asked was simple:

“Does this mean NetSuite is charging for AI?”

It’s an understandable reaction. It’s also the wrong question. To understand what’s actually happening, it helps to back up one step and ask a more basic question first: what are AI tokens, and why does every major AI platform need some version of them?

What Are AI Tokens?

A token is the basic unit AI models use to measure how much work a request takes. Roughly speaking, a token is a chunk of text, often a few characters or part of a word. When you send a prompt to an AI system and get a response back, both the input and the output are broken into tokens, and that count is what determines the computing cost behind the scenes.

Tokens exist because every AI interaction is generated in real time. Unlike a report or dashboard that’s already been built and simply gets displayed, a large language model has to process the request from scratch, run billions of calculations on specialized hardware, and produce a fresh response in seconds. That has a real, measurable operating cost, which is fundamentally different from how traditional software has priced access for decades: by number of users, regardless of how heavily each one used the product.

So AI providers needed a new unit of measurement. The name changes depending on who you ask.

The terminology changes. The underlying idea doesn’t: measure the resource, so the platform can stay sustainable as usage scales.

So What Are NetSuite AI Units, Specifically?

NetSuite AI Units are Oracle’s version of the same concept, applied to its ERP platform. Most NetSuite customers receive AI Units automatically through their existing user licenses, and those units are pooled at the account level rather than assigned per person.

That pooling matters more than it might seem. Think of it less like a per-seat allowance and more like a shared company budget. Some employees will rarely touch native AI features. Others, like a finance lead running forecasts daily, might use them constantly. Because the allocation is shared, the organization can direct usage where it creates the most value instead of managing individual limits.

For most businesses, the AI Units included with eligible licenses will be more than enough. Oracle has been clear that additional AI Units are meant for organizations whose AI adoption grows significantly across many users over time, not as a default purchase everyone needs to plan for on day one.

The Distinction That Actually Matters: Native AI vs. Third-Party AI

Here’s where a lot of the early confusion is coming from, and it’s worth being precise about it.

NetSuite’s native AI experiences (things like Ask Oracle) do consume AI Units. But if you’re connecting external models like ChatGPT or Claude through the NetSuite AI Connector Service, that usage does not draw from your NetSuite AI Unit pool. It runs on a separate model entirely.

That’s an architectural choice, not a footnote. It means an organization can combine Oracle’s native AI capabilities with external AI platforms based on what each use case actually needs, without forcing every AI interaction through a single consumption meter. If your team is already using the AI Connector Service to bring Claude into NetSuite workflows, this announcement doesn’t change that math at all.

Why This Signals a Bigger Shift Than Pricing

The more interesting part of this announcement isn’t the mechanism. It’s what the mechanism says about where enterprise AI is headed.

Five years ago, AI inside ERP software was mostly a demo. Something shown at a keynote, tested in an innovation lab, rarely load-bearing in daily operations. Today, Oracle is treating AI as core platform infrastructure, on the same tier as storage, compute, and API usage. That’s a meaningful jump in maturity, and it changes the questions organizations need to be asking.

Instead of “should we use AI,” the more useful questions become:

• Which AI capabilities actually deliver business value for our team?
• Which workflows should run through native ERP AI, and which are better served by external tools?
• How do we encourage adoption without losing visibility into usage?
• How do we balance productivity gains against operational cost?

Those aren’t licensing questions. They’re AI governance questions, and they’re going to keep showing up as more of the software stack gets AI-native. Getting ahead of them, rather than reacting once usage patterns are already set, is the difference between an AI rollout that scales cleanly and one that turns into a budget surprise eighteen months in.

What This Means for Your NetSuite Investment

For budgeting purposes, AI Units are becoming one more line item to understand as part of your total NetSuite investment, alongside user licenses, modules, and implementation. Right now, for the large majority of NetSuite customers, that line item costs nothing extra: it’s included, pooled, and generous enough that it won’t be a factor. The organizations who will need to think about it are the ones scaling native AI usage aggressively across many users, which is a good problem to have, and one worth planning for rather than being surprised by at renewal.

Frequently Asked Questions

Does the NetSuite AI Connector Service use AI Units? No. Connecting third-party AI models like ChatGPT or Claude through the NetSuite AI Connector Service does not consume NetSuite AI Units. Only native NetSuite AI experiences, like Ask Oracle, draw from the pool.

Do I have to pay extra for NetSuite AI Units? Most NetSuite customers receive AI Units automatically through eligible user licenses, at no additional cost. Additional units are only relevant for organizations whose AI usage grows well beyond what’s included.

Are NetSuite AI Units the same as AI tokens? They measure the same underlying idea, AI usage, but they’re not interchangeable terms. Tokens are the unit OpenAI and Anthropic use to measure model input and output. AI Units are Oracle’s account-level equivalent inside NetSuite.

Talk to Tavano Team about how NetSuite’s AI roadmap fits into your broader NetSuite investment and eCommerce strategy.