October 31, 2026 • AI Costs & Budgeting • By Jeff Utecht

AI Cost for School Districts: Paying for Seats vs. Paying for Use

District CFO and superintendent reviewing AI budget charts together at a conference table in a district office

Every AI budget conversation I have with district leaders eventually lands on the same question: what are we actually paying for? Most of the time the honest answer is “seats.” A license for every staff member, renewed every year, whether that person opens the tool daily or never logs in at all.

But AI is metered work — text read and text written — and the providers themselves sell it that way. So when we think about AI cost for school districts, I want leaders to see the two pricing models side by side, with real published prices and simple arithmetic you can check yourself.

Here’s that math, using prices I looked up today, October 31, 2026.

Key takeaways

  • Per-seat AI licenses charge for every user every year, regardless of how much each person actually uses the tool.
  • AI providers sell the same underlying models by the token, and a typical teacher request costs about a penny or two at current published API rates.
  • A district-owned private server is a one-time purchase with no per-user fees; open-source models on it carry no token fee, and commercial models can be connected by API on a pay-for-use basis.
  • Token prices change too, so the real advantage is owning the platform and being free to switch models when prices or quality shift.

The real AI cost for school districts: per-seat licenses

Let’s start with a published price. Google’s AI Pro for Education add-on is listed at $20 per user per month on an annual plan, or $24 per user per month on a flexible plan. Google also notes it is only available to users 18 and over — so it covers your staff, not your students.

For a district licensing 500 staff at the annual rate:

500 staff × $20 per month × 12 months = $120,000 per year — and the same again next year.

Other tools follow the same pattern. MagicSchool’s Plus plan is $12.99 per month for an individual teacher, or $99.96 billed yearly; its school and district pricing is by quote only, so I won’t guess at it.

To be fair: those subscriptions buy real things. Polished teacher features, AI built directly into the apps people already use, support teams and curriculum tie-ins. But be clear-eyed: you’re paying the same for your heaviest user and your staff member who logged in once in September.

Tokens explained: paying for use instead of seats

When AI providers sell to developers, they don’t sell seats. They sell tokens. A token is a small chunk of text, roughly three-quarters of an English word. Providers charge per million tokens the model reads (input) and per million tokens it writes (output).

A typical teacher request might be a few thousand tokens. For the math below I’ll assume a generous 3,000 input tokens and 1,000 output tokens per request. That’s my working assumption, not a measured figure; your district’s real usage will vary.

Now the published prices:

Now picture a heavy user: a teacher making 10 requests every school day, 20 school days a month. That’s 200 requests × $0.016 = about $3.20 a month on Claude Sonnet 5.5 — compared with a $20 seat.

Scale it to a district. A labeled hypothetical: 500 staff, 300 of whom use AI regularly at 150 requests a month, over a 10-month school year.

300 users × 150 requests × 10 months = 450,000 requests
450,000 × $0.016 = about $7,200 per year on Claude Sonnet 5.5

Even if all 500 staff were heavy users at 200 requests a month for all 12 months, that’s 1,200,000 requests × $0.016, or about $19,200 — still a fraction of $120,000.

Where the private server fits in the math

Here’s how the PrivatEDU AI server works on cost. It’s a one-time $20,000 purchase that includes pre-configured server hardware, 12 months of remote support, and a year of virtual professional development from LeadershipAIplus. There are no per-user licensing fees, and one server supports roughly 200 concurrent users.

Two important details, stated plainly:

  1. Open-source models running on the server have no per-token fee. Models like Llama and Mistral run locally on district hardware.
  2. Commercial API tokens are not included in the $20,000. If the district connects Claude, OpenAI or Gemini, it sets up its own API account with each provider and pays that provider directly for the tokens it uses. Those specific requests go to that provider under its API terms, and the district decides when and for whom that happens.

So in my hypothetical, year one looks like $20,000 for the server plus about $7,200 in API tokens if every request went to Claude Sonnet 5.5 — roughly $27,200, against $120,000 in seats. Route a good share of the work to local models and the token bill shrinks further.

And notice the Gemini price above: Flash doubles on January 1. Token prices move, just like seat prices. The difference is that when you own the platform, you can switch the model behind a teacher’s assistant in an afternoon. You aren’t locked in. That’s the same lesson schools are learning right now as Gems become Skills and Custom GPTs are retired: when someone else owns the platform, they decide what changes.

What district leaders can do this month

  1. Pull your real login data. For every AI tool you license, ask how many seats were active in the last 30 days. Divide total cost by active users, not purchased seats.
  2. Estimate your token footprint. Pick a few typical staff tasks, count roughly how much text goes in and out, and multiply by current published API prices. Show your assumptions.
  3. Separate staff and student needs. Many per-seat products are adult-only. Decide what student access should look like and what guardrails it needs.
  4. Ask every vendor the churn question. “What happens to the assistants our teachers built if you change or retire this feature?”
  5. Give your CFO a five-year view. Recurring seats versus one-time ownership plus metered use, with a buffer for price changes.

Why a private server is the smarter long-run investment

I’ve partnered with schools since 2009, starting with 1:1 laptop rollouts, and the lesson I keep relearning is that owning and supporting the platform matters more than the device or the app on top of it. A private server is not one more chatbot. It’s the district’s own AI platform: you control the data, the models, the features and the cost. I recommend running it on-site, where prompts and logs stay on district-owned hardware inside your network — the strongest position for control and data security. That’s the heart of why districts choose private AI.

It’s also why I’m proud to be a program partner of LeadershipAIplus: educational leaders supporting educational leaders, not a vendor chasing seats. Districts from Kansas to Washington to Berlin are running a private server through the program, and the professional learning comes with it, including courses on using and administering the server. If you want a deeper planning view, our guide to private AI servers for K12 walks through data boundaries, access and governance.

FAQ

How much does AI cost for a school district?

It depends on the model. Google AI Pro for Education lists at $20 per user per month annually, so 500 staff is $120,000 a year. Pay-per-token API use for the same staff can cost a fraction of that, depending on usage.

Are API token costs included in the PrivatEDU AI server price?

No. The $20,000 one-time price covers the server hardware, 12 months of remote support and a year of virtual PD. Open-source models on the server have no token fee; commercial models connected by API are billed by the provider directly to the district.

What is a token in AI pricing?

A token is a small chunk of text, about three-quarters of an English word. A typical classroom request of a few thousand tokens costs a cent or two at current published API rates.

The bottom line

Seats charge you for people. Tokens charge you for work. Ownership lets you choose. When AI frees up time for teachers, as I wrote in AI Gave Teachers Time Back. Who Gets to Keep It?, the budget should follow the value, not the headcount.

So here’s my question for you: if you divided last year’s AI spend by the people who actually used it, would you sign that contract again?

Ready to run the numbers for your district? Schedule a Systems Review.

Sources

One thought on “AI Cost for School Districts: Paying for Seats vs. Paying for Use”

  1. Pingback: Private AI Server for Schools: October Updates

Leave a Reply

Your email address will not be published. Required fields are marked *