Copilot's December 1 Billing and Licensing Changes: What Changes and What Doesn't

Two separate changes land on the same date

On 1 December 2026, two unrelated Microsoft changes take effect for Microsoft 365 Copilot. They are being widely reported as one thing, and that conflation is causing the wrong people to panic and the right people to miss their deadline.

They are:

1. Usage-based billing becomes the default for new Microsoft 365 Copilot Business licences purchased through the Cloud Solution Provider (CSP) channel, with a preset spending limit of 4,000 Copilot Credits per user per month.
2. The base-subscription prerequisite disappears. New Microsoft 365 Copilot and Microsoft 365 Copilot Business offers can be bought without a qualifying Microsoft 365 base subscription, and the existing offers that do require one reach end of sale in CSP.

The first is a billing-configuration change with a narrow blast radius and a genuine governance problem inside it. The second is a catalogue change that will break licence automation at organisations that never touch Copilot Business at all.

Neither changes a single Copilot capability. Microsoft is explicit on that point: the purchasing change "affects how customers purchase Copilot, not the underlying Copilot capabilities."

Change 1: usage-based billing on by default

From 1 December 2026, pay-as-you-go becomes the default billing configuration for new Microsoft 365 Copilot Business licences bought in CSP — standalone offers and bundles alike. Microsoft's stated reason is to remove setup friction: new licences arrive with the Azure subscription wiring that usage-based billing needs, so customers can reach metered experiences without a multi-step billing configuration.

The metered experiences this unlocks are Copilot Cowork, the Work IQ APIs, and GitHub Copilot Harness.

Note the original date. This was first communicated as 2 November 2026 and then moved to 1 December. If your change calendar still says November, it is wrong, and anyone who built a readiness plan around the earlier date has four more weeks than they think.

The 4,000-credit default is a ceiling, not a budget

The default spending limit is 4,000 Copilot Credits per user per month, and administrators can change it.

Microsoft's own worked example is the number to put in front of your finance partner: a customer with 100 Copilot Business users could consume up to 400,000 Copilot Credits per month under the default limit, with charges based on actual usage.

Read that carefully, because the two halves pull in opposite directions and both matter:

• Charges follow real consumption. Nobody is billed 4,000 credits for a user who consumes nothing. The limit is not a commitment and not a subscription fee.
• The default limit is nevertheless an authorisation. It is the amount of consumption your tenant is pre-approved to incur per user per month, set by a vendor default rather than by anyone in your organisation.

A ceiling that no one in your finance function chose, multiplied by seat count, is a governance gap even when actual spend lands nowhere near it. The question to answer before 1 December is not "will we hit 400,000 credits?" It is "who in this organisation decided 4,000 was the right number, and who is allowed to change it?"

If the honest answer is "nobody decided, it arrived that way," that is the finding. Treat it the way you would treat any other inherited default with a cost attached.

Not every market, not on day one

The rollout begins with eligible customers in supported markets and is initially not available in Australia, Belgium, Brazil, France, Germany, India, Italy, Korea, the Netherlands, Poland, or Spain.

For a multinational, that matters more than it first appears. If your Copilot Business estate spans, say, the UK, Germany, and the Netherlands, two of those three are outside the initial rollout. You will be operating two different default billing configurations inside one tenant estate, on a timeline Microsoft has not yet published. Plan for the split rather than for a single cut-over, and ask your partner for the market list in writing when availability expands.

What it does not touch

Existing subscriptions are not switched over. The change applies to new purchases of Copilot Business licences.

And Microsoft has not announced default-on pay-as-you-go for the larger Microsoft 365 Copilot licence. If you are an enterprise running Microsoft 365 Copilot rather than Copilot Business, your administrators still have to enable usage-based billing deliberately. The 1 December default does not reach you.

That distinction is where most of the current commentary goes wrong, and it is worth saying plainly inside your own organisation before somebody escalates a billing change that does not apply to your licences.

Change 2: the prerequisite goes away, and the old offers stop selling

This is the one that will generate tickets.

From 1 December 2026, new Microsoft 365 Copilot and Microsoft 365 Copilot Business offers can be purchased without a qualifying Microsoft 365 base subscription. The new offers are available across all channels including CSP, for commercial, education, government, and nonprofit customers.

The new offers:

• have their own product and SKU IDs
• are priced the same as the current offers
• have IDs that Microsoft said would be shared ahead of the price list preview on 1 November 2026

From 1 December 2026, these new offers replace the add-on offers for new Copilot and Copilot Business purchases.

The rename already happened, and it is the real trap

From 16 October 2026, the current offers that require a qualifying base subscription carry "(add-on)" in their names — for example, "Microsoft 365 Copilot (add-on)" and "Microsoft 365 Copilot Business (add-on)."

Microsoft's guidance on this is unusually direct, and it is an engineering instruction rather than a commercial one: update scripts, catalogue mappings, and automations that rely on offer names so they key off product and SKU IDs instead.

If you have a procurement workflow, a licence-reconciliation job, a CMDB sync, a chargeback report, or a provisioning script that string-matches on "Microsoft 365 Copilot", it is already matching a different string than it did in mid-October. Depending on how the match is written, that is either a silent no-match or a silent wrong-match, and both fail quietly in a reconciliation job that nobody reads until quarter end.

This is the item to check this week. It does not wait for 1 December, and it affects organisations that will never buy a Copilot Business licence.

End of sale is not end of life

The "(add-on)" offers reach end of sale in CSP on 1 December 2026.

Microsoft is explicit that end of sale is not end of life: existing customers can continue to renew and adjust seat quantities. For existing subscriptions, no action is required. Product and SKU IDs, pricing, renewal terms, auto-renew settings, and service all remain unchanged. There is no migration and no licence reassignment.

So if your renewal lands in March, nothing happens to it on 1 December. What changes is which offer a new purchase is written against — which means a mid-term seat addition and a net-new purchase may now sit on different SKUs in the same tenant. Worth knowing before your asset register flags it as a discrepancy.

Two adjacent changes worth folding into the same review

AI Builder capacity add-ons retire from sale on 1 November 2026. Customers with active subscriptions can keep using existing AI Builder credits through the end of the current subscription term, and Microsoft's guidance is to plan on Copilot Credits for anyone who wants to continue using AI Builder features afterwards. Critically, partners must turn off auto-renew before each customer's next renewal date to stop a retired add-on renewing.

Office LTSC 2021, Project LTSC 2021, and Visio LTSC 2021 went out of support on 13 October 2026. The Copilot-relevant detail: Microsoft notes that Copilot is supported only on the cloud-backed apps included in Microsoft 365 suites, and that on-premises versions of Office are not eligible. The Azure Information Protection labelling client for perpetual Office has also been retired. If any part of your estate is still on perpetual Office and in scope for Copilot, those two facts decide the sequencing for you.

What to do before 1 December

A short, concrete list. None of it requires a project.

1. Fix the name-matching first. Grep your automation for literal Copilot offer names. Replace with product and SKU IDs. This is already live as of 16 October and is the only item on this list that is actively breaking things today.
2. Decide your spending limit deliberately. Do not inherit 4,000 credits per user per month by default. Pick a number, write down who picked it, and record who is authorised to change it. A default you did not choose is not a decision.
3. Separate Copilot Business from Microsoft 365 Copilot in your own comms. Most of the alarm circulating is about a default that only applies to new Copilot Business purchases in CSP. Say so internally before somebody acts on the wrong briefing.
4. Map your markets. If you operate in any of the eleven markets outside the initial rollout, expect two billing configurations across your estate and get the expansion timeline from your partner in writing.
5. Put Copilot Credits consumption on an existing report. Usage-based billing introduces a variable line item. It needs an owner and a monthly review in whatever forum already reviews cloud spend — not a new one.
6. Check your AI Builder auto-renew. If you have AI Builder capacity add-ons, confirm auto-renew is off before the next renewal date and decide whether Copilot Credits cover the same scenarios.
7. Confirm the eligibility of your Office estate. On-premises Office is not eligible for Copilot. If LTSC 2021 is still in production anywhere in scope, that sequencing question comes before any licensing decision.

The pattern underneath all of this

The licensing-prerequisite change and the usage-based default point the same way: Copilot is moving from a per-seat add-on with a predictable monthly cost toward a platform with a consumption component attached to it.

A per-seat add-on is a procurement decision made once a year. A consumption component is an operational responsibility — it needs a limit, an owner, a reporting line, and a review cadence, the same way Azure consumption does. Organisations that already run cloud spend that way will absorb this without noticing. Organisations treating Copilot purely as a licence line will meet consumption governance for the first time, under a default somebody else set.

The 4,000-credit ceiling is not the risk. Not having decided on it is.

If you are working out which controls belong in the decision, our Copilot governance work covers spending limits and consumption ownership alongside the data-governance controls, and the delivery framework sets out where a licensing decision sits relative to readiness and deployment. For the licence-tier question underneath all of this, the E3 and E5 governance framework is the better starting point, and consumption planning goes deeper on modelling query cost. Where the spending limit should be reviewed is a question for your governance board's charter.

Sources

Every date and figure above comes from Microsoft's published announcements, not from interpretation:

• Partner Center announcements, October 2026 — "Update: Usage-based billing is default on for new Microsoft 365 Copilot Business licenses starting December 1, 2026" (announced 1 October 2026, updated 14 October 2026); "Copilot offers and name updates" (14 October 2026); "AI Builder capacity add-ons retiring" (22 October 2026); "Office LTSC 2021 end of support" (15 October 2026)
• Understand usage-based billing and cost management for Copilot Credits
• Managing AI experiences enabled by usage-based billing
• Microsoft 365 Copilot licensing

Dates and market availability are subject to change by Microsoft. Verify against the Partner Center announcement before acting on a specific date.