On October 1, 2026, Microsoft quietly moved a date most mid-market buyers have not yet circled. Usage-based billing for new Microsoft 365 Copilot Business subscriptions purchased through CSP will now be enabled by default starting December 1, 2026, previously communicated as November 2. The default spending ceiling will be 4,000 Copilot Credits per user per month, admin-adjustable, and the configuration ships with a side effect most finance teams have never thought about: an auto-provisioned Azure subscription appearing inside their tenant on first purchase.

This is not another pricing announcement. It is a default-configuration event that transfers responsibility across three admin surfaces, creates a new category of CSP partner invoice liability, and gives both sides of the transaction roughly 50 days to decide who owns which risk. Buyers signing new Copilot Business orders in November will inherit a credit ceiling, an Azure footprint, and a consumption exposure they never negotiated. CSP partners who have not tested in the Partner Center sandbox before late November will inherit reconciliation surprises on the first December invoice.

Here is what actually changed, how the two Copilot billing meters differ, what the arithmetic looks like at mid-market scale, and the three moves both parties must make before December 1.

What Actually Changed on October 1

The date Microsoft moved: November 2 to December 1

The Partner Center October 2026 announcement confirms the revised effective date: usage-based billing will now be enabled by default for new Microsoft 365 Copilot Business subscriptions purchased through CSP, beginning December 1, 2026. The shift from November 2 compresses CSP partner readiness into a 29-day sandbox window: Partner Center sandbox environments open November 2, giving solution providers less than a month to test the new reconciliation fields before the production default-on cutover, as CRN reported.

Which SKUs are in scope, and which are not

The default-on behavior applies to new purchases of Microsoft 365 Copilot Business, Microsoft 365 Business Premium with Copilot, Microsoft 365 Business Standard with Copilot, Microsoft 365 Business Basic plus Microsoft 365 Copilot Business, and the other bundled B forms of Copilot documented on Microsoft Learn. CRN confirms that additional seats added to an existing subscription, upgrades to Copilot Business, and subscriptions transferred from a prior solution provider will also arrive with usage-based billing on by default. Existing subscriptions purchased before December 1 are unaffected unless they are modified in a qualifying way.

Microsoft 365 Copilot Business is the SMB and mid-market SKU, aimed at tenants with up to 300 seats and priced in the $18 to $32 per user per month range. It is a distinct product from the enterprise Microsoft 365 Copilot license. If your organization sits on the enterprise SKU, this specific default-on event does not apply to you today.

The 4,000-credit default nobody asked for

Microsoft set the default spending limit at 4,000 Copilot Credits per user per month. In Microsoft's own example on the Partner Center page, a customer with 100 Microsoft 365 Copilot Business users could consume up to 400,000 Copilot Credits per month under the default spending limit, with UBB charges based on actual usage. At the published rate of one cent per Copilot Credit, that ceiling implies up to roughly $40 per user per month of variable exposure stacked on top of the Copilot Business list price.

Admins can lower the ceiling, raise it, or set it to zero. But the default lands on December 1 whether anyone touches it or not.

Two Meters, One Confusion

Copilot Credits: the Microsoft 365 admin center meter

Copilot Credits are the common currency for a specific set of eligible Microsoft 365 services with usage-based billing, metered in the Microsoft 365 admin center. The Microsoft Learn overview lists the experiences enabled under this meter, including Copilot Cowork, Work IQ APIs, advanced work in SharePoint and OneDrive, Copilot Managed Runtime, and the GitHub Copilot Harness experience. This is the meter that goes default-on in CSP on December 1.

The Copilot Studio message meter is a different bill

The Copilot Studio message meter is a separate pay-as-you-go construct billed through an Azure subscription at $0.01 per message, covering Copilot Chat, SharePoint Agents, and the Microsoft Copilot Retrieval API, as documented in the Microsoft Copilot pay-as-you-go overview. It is not the Copilot Credits meter, and it is not what changes in CSP on December 1. If your team needs the deeper explainer on that meter, we covered it in the BabyBots Guide Copilot Studio Billing Just Changed.

Why the distinction matters for your December 1 decision

Most published coverage conflates these two meters. The practical consequence is that finance teams run arithmetic on the wrong rate card, and admins configure spending policies in the wrong console. The simple decision rule: if the workload runs inside Microsoft 365 Copilot Business (Cowork, Work IQ, SharePoint and OneDrive agentic tasks, GitHub Copilot Harness), it bills against Copilot Credits in the Microsoft 365 admin center. If it runs inside a Copilot Studio agent you built, it bills against the Copilot Studio message meter in Azure. Different consoles, different owners, different invoices.

The Arithmetic at Mid-Market Scale

100, 500, and 2,000 seats under the default ceiling

Microsoft's own 100-user example of up to 400,000 credits per month is the anchor. Extended to realistic mid-market tiers at the one-cent-per-credit rate, the default ceiling implies the following maximum monthly variable overlays, all before any admin intervention:

Default-ceiling monthly exposure by seat count

• 100 seats: up to 400,000 credits per month, approximately $4,000 per month in variable Copilot Credits consumption at the default ceiling.
• 500 seats: up to 2,000,000 credits per month, approximately $20,000 per month in variable Copilot Credits consumption at the default ceiling.
• 2,000 seats: up to 8,000,000 credits per month, approximately $80,000 per month in variable Copilot Credits consumption at the default ceiling.

None of these numbers reflect actual usage. They are the headroom the default policy authorizes. The gap between what we are entitled to spend and what we expect to spend is the exposure a CFO needs to see before December 1.

What 4,000 credits actually buys a user

Microsoft frames the 4,000-credit allowance operationally. According to the Partner Center documentation, that monthly budget would allow a single user four weekly Cowork customer meeting preparation and follow-up sessions, four weekly inbox triaging sessions, four weekly commit and pipeline risk reviews, two RFP responses with compliance matrix completions, and one generated branded customer pitch deck. CRN's reporting breaks the task economics into tiers: light tasks consume up to roughly 300 credits, medium tasks up to about 700 credits, and heavy tasks over 700 credits each. The ceiling is generous for an average knowledge worker and dangerous for a power user turned loose on long-running agentic workflows.

The budgets alert, they do not enforce problem

Microsoft's spending policy documentation describes two behaviors most finance teams will not expect. First, prepaid credits draw first; capacity packs and pre-purchase plans are consumed before pay-as-you-go kicks in. Second, running tasks that exceed the per-user cap can complete without being billed in that cycle, and budget alerts repeat weekly rather than acting as hard enforcement. On the Azure side of the house, the same pattern holds: budgets notify, they do not stop consumption. If there is no per-user cap paired with the tenant limit, a single user can draw against the shared pool until the shared pool is gone.

The December 1 switch is not a pricing announcement. It is Microsoft handing every mid-market buyer a 4,000-credit-per-user default ceiling, an auto-provisioned Azure subscription, and a CSP invoice obligation — and giving both sides 50 days to decide who owns which risk.

The Exposure Nobody Is Mapping: Buyer vs. CSP Partner

Every published summary of the December 1 switch treats it as a one-sided event: either a buyer story or a partner story. In practice it is a two-sided risk event with a third silent actor. Mapping who owns what is the first move that produces clarity.

What the buyer owns

The buyer owns the spending policy configuration, the per-user cap decision, the governance of the auto-provisioned Azure footprint, end-user behavior inside Cowork and Work IQ, and the budget authority to approve or deny credit-increase requests. The buyer also owns the internal communication to employees about what a 4,000-credit ceiling means and what the escalation path is when a user hits it.

What the CSP partner owns

The CSP transacting partner owns the invoice. According to CRN, consumptive charges flow through the CSP's invoices, and the transacting partner that receives the bill is ultimately responsible for all charges under the CSP transacting account, including unexpected consumption. The partner also owns reconciliation-file readiness in the Partner Center sandbox, the new promotion identifier and discount source field handling, and the dispute escalation path when subscription mappings go wrong. For indirect resellers, the distributor sits in the middle of that chain.

What Microsoft owns

Microsoft owns the default configuration itself: the 4,000-credit ceiling, the auto-provisioning behavior, the geographic rollout sequence, the catalog API fields partners read to detect UBB eligibility, and the behavior of running tasks when a per-user cap is exceeded. Microsoft does not own the consequence of either party failing to act inside the 50-day window.

The Auto-Provisioned Azure Subscription Is a Governance Event

Three teams who need to review it before December 1

When UBB activates on a new Copilot Business purchase for a customer without an existing Azure footprint, Microsoft creates an Azure plan, an Azure subscription, a resource group, and a default usage-based spending policy inside the tenant automatically. That is a billing artifact, but it is also a security, FinOps, and compliance artifact. Three functions should touch it before December 1:

• FinOps: own the tagging, cost center attribution, and chargeback mapping for the new subscription before the first invoice cycle, otherwise December consumption lands in an unallocated bucket.
• Security: review the auto-created resource group against baseline tenant controls, confirm Purview and DSPM coverage extends to the new Azure footprint, and verify identity and access boundaries match the rest of the Azure estate.
• Compliance: confirm the data residency region of the auto-provisioned subscription matches policy, particularly for regulated workloads and multi-national tenants.

Why three out of four CSP buyers will get a brand-new Azure footprint

Microsoft Learn notes that roughly three out of four Copilot Business customers sold through CSP have no Azure footprint today. For that majority, the December 1 default-on event is not an incremental configuration change: it is the first Azure subscription their organization has ever had. The security and FinOps processes that enterprise Azure customers take for granted do not exist yet in these tenants. The subscription appears, and nobody is watching for it.

The 50-Day Playbook

Phase 1 (October 12 to November 1): Contract and posture

Use the first three weeks to make the decisions nobody has to make under time pressure. Procurement reviews MSA language for consumption liability coverage: does the current master agreement with the CSP cover variable Copilot Credits charges, or does it assume a fixed per-seat subscription? The CIO sets the 90-day Cowork posture, choosing among off, limited pilot inside one business unit, or broad availability with a lowered ceiling. The CFO decides the opening credit ceiling the organization will configure on day one, independent of what Microsoft defaults to.

Phase 2 (November 2 to November 30): Sandbox and pre-configuration

The Partner Center sandbox opens November 2. During this window, the CSP partner tests reconciliation against the new promotion identifier and discount source fields and confirms their billing system reads the catalog API property that signals UBB eligibility. On the buyer side, the Microsoft 365 admin creates the default tenant spending policy before any new purchase can trigger auto-provisioning, sets a per-user cap underneath the tenant limit, and configures alert thresholds. FinOps, security, and compliance pre-approve the auto-provisioned Azure subscription template so that when it appears on first purchase, no one is surprised.

Phase 3 (December 1 to December 31): Go-live monitoring

The default is live. The buyer runs weekly reviews of credit-increase requests, pulls cost reports every two hours during the first week to catch runaway consumption patterns, and audits the first invoice against expected usage. The CSP partner audits reconciliation output against the sandbox baseline, resolves any subscription mapping errors inside the Microsoft-defined escalation path, and documents dispute patterns to feed back into Q1 contract reviews.

Three Questions to Put to Your CSP Partner in Writing Before November 30

These are not conversations to have on a quarterly check-in call. Put them in writing, dated, with a requested response date before November 30. The written record is what protects both parties if the December invoice contains surprises.

• Will new orders you process for us arrive with UBB default-on, and at what per-user credit ceiling? The answer should name the specific SKUs affected and confirm whether your partner intends to apply any pre-configured ceiling lower than 4,000 credits before the subscription provisions.
• Has your billing and reconciliation system been tested in the November 2 Partner Center sandbox against the new promotion identifier and discount source fields? The answer should include a specific sandbox test date and whether reconciliation output matched expected format.
• What is the escalation path and SLA for incorrect subscription mappings, auto-provisioned Azure artifacts that land in the wrong region, and disputes over unexpected consumption? The answer should name the escalation owner, the response window, and whether disputed amounts are held or billed pending resolution.

The Multi-National Nuance: 11 Countries Are Not on the Guest List

For multi-national mid-market buyers, the December 1 effective date is not universal. According to CRN, 11 countries will not initially support usage-based billing: Australia, Belgium, Brazil, France, Germany, India, Italy, Korea, the Netherlands, Poland, and Spain. If your organization is headquartered in or operates significantly across any of these markets, the rollout is sequenced, not simultaneous. That changes two things. First, a uniform global spending policy is premature; policy design should assume differential effective dates for the next several quarters. Second, procurement should ask the CSP partner for the Microsoft-published rollout sequence for the carve-out countries, so that country-level readiness is scheduled rather than reactive.

A Contrarian Position: Why Zero Is Sometimes the Right Starting Ceiling

The industry consensus will be to accept the 4,000-credit default or lower it modestly. The BabyBots position is different. For a specific subset of mid-market buyers in the first 90 days, zero is the right opening ceiling. Not forever. Not as a permanent posture. As a controlled start.

Setting the spending limit to zero disables usage-based billing, which also disables access to Cowork, Work IQ APIs, and GitHub Copilot Harness experiences until the limit is raised. That tradeoff is acceptable, and arguably correct, when any of the following conditions apply:

• The organization has no documented data labeling and sensitivity baseline, meaning agentic workflows that read SharePoint and OneDrive content will inherit oversharing risk the organization has not yet remediated.
• The organization has no existing Azure footprint, no FinOps function, and no security team capable of reviewing the auto-provisioned subscription inside the 50-day window.
• The organization has not yet decided which user cohort should have access to Cowork at all, and a broad default-on experience will create change-management work that outruns the governance team.

In those cases, open at zero, raise the ceiling deliberately in January after a controlled Cowork pilot inside one business unit, and let the January invoice reflect scoped consumption rather than default consumption. The point is not that pay-as-you-go is wrong. The point is that default is not a synonym for recommended.

What You Now Know

If you only read the Microsoft announcement, you know the date, the ceiling, and the included experiences. If you only read the trade press, you know the partner-side framing and the executive quotes. What you did not know this morning, and now do:

• The December 1 default-on event transfers responsibility across three admin surfaces (Partner Center, Microsoft 365 admin center, and Azure), not one.
• The 4,000-credit ceiling implies up to roughly $4,000, $20,000, and $80,000 per month of variable consumption headroom at 100, 500, and 2,000 seats respectively, before any admin intervention.
• Copilot Credits and the Copilot Studio message meter are different bills on different consoles with different owners, and the December 1 event affects only the first.
• Three out of four CSP Copilot Business buyers will receive their first-ever Azure subscription auto-provisioned on first purchase, a security and FinOps event most organizations are not watching for.
• The CSP transacting partner carries residual invoice liability for unexpected consumption, which is why the three written questions matter before November 30.
• Eleven countries are not on the December 1 guest list, which changes rollout sequencing for multi-national buyers.
• For a specific subset of organizations, zero is the correct opening ceiling, and default is not a synonym for recommended.

The organizations that enter 2027 with controlled Copilot consumption will be the ones that treated the 50-day window as a decision window, not a notification window. BabyBots helps mid-market CIOs, CFOs, and CSP partners run the three-phase playbook end-to-end, from contract clause review through spending-policy design and partner reconciliation alignment. The playbook works whether you run it with us or without us. What does not work is treating default as a plan.