Microsoft 365 Copilot actually costs $69/month (not $30): renegotiate your EA before your next renewal

August 2, 202616 minMicrosoft, Copilot, M365, enterprise pricing, Latam SaaS, unit economics

Short answer (60 seconds): Microsoft raised the M365 base suite on July 1, 2026 (E3 USD 36→39, E5 USD 57→60) after killing volume discounts on November 1, 2025. Combined with Copilot at USD 30/user, the real all-in is USD 69/user/month with E3 or USD 90 with E5. For a 1,000-seat company, that is USD 828K/year (E3) or USD 1.08M/year (E5) — without anyone having asked for more software. Just the seat uplift from the suite change (no Copilot) adds USD 36K/year in E3 for that same company. The new M365 E7 bundle at USD 99/user exists as Microsoft's attempt to "absorb" the price hike. The new rates take effect at each customer's next EA anniversary after July 1. If you sell SaaS to enterprise clients, their M365 budget shrank 8-10% in real terms: the "value" conversation just got harder this week.

The "Copilot at USD 30" headline got replicated everywhere in July. It is a real number, but misleading: the all-in cost depends on which base suite you have, how many seats, and whether you still get the volume discounts (which Microsoft eliminated in November 2025). This post is the one I wished a CFO had on a Monday morning when the Microsoft notice landed. It has four parts: the pricing table so you do not have to build it, the real all-in calculation across three company sizes, the E7 bundle vs the separate combination decision, and a checklist for SaaS founders selling to enterprise.

Disclosure: we are not a Microsoft partner and do not resell licenses. The numbers in this post come from Microsoft's public pricing guides, the coverage by infosprint, llmgateway and samexpert (cited below), and our own calculations using the formula seats × monthly price × 12. Where I quote a figure without an exact source — particularly the 100, 1,000 and 5,000-seat scenarios — those are explicit estimates against list pricing. Before making contracting decisions, validate the numbers with your reseller or your Microsoft account manager.

What happened this week (and why it matters)

Microsoft executed four pricing moves between June and July 2026 that, taken together, change the total cost of ownership (TCO) of any company running Microsoft 365 with Copilot:

1. Base suite price hike on July 1, 2026. E3 went from USD 36 to USD 39/user/month, E5 from USD 57 to USD 60. The change applies to new contracts and renewals; contracts in flight keep the old price until their next anniversary (source: Microsoft Licensing update 2026).

2. Volume discounts eliminated on November 1, 2025. A 1,000+ seat company could previously negotiate 10-20% off list. After November 1, 2025, the maximum discount sits at 2-5% and only applies to multi-year commitments with prepayment, per the Microsoft Learn pricing changes and the Samexpert analysis of the EA in 2026.

3. End of the "Copilot for just USD 30 more" upsell. Copilot list price stays at USD 30/user/month as an addon, but the suite base went up and volume discounts disappeared. The all-in for a company buying today is USD 69 with E3 or USD 90 with E5 (source: Atonement Licensing — M365 Copilot pricing 2026).

4. New M365 E7 bundle at USD 99/user. Launched as a response to the noise. Includes E5 + Copilot + Agent 365 (low-code agents) + Entra Suite (ID P2 + Governance) + Purview upgraded (eDiscovery Premium tier). Targets companies already paying E5 + Copilot and the ones Microsoft wants to move from E3 + Copilot to a "premium" bundle without selling feature by feature (source: LLM Gateway — Microsoft Copilot enterprise pricing).

The macro context: the new July 1 rates take effect at each customer's next EA anniversary. EAs do not all renew on the same day — each contract has its own cycle (typically 3 years from signature). If your EA expires between July 2026 and July 2027, you will renegotiate the new list price during that window with no volume discount.

Comparative M365 + Copilot pricing (August 2026, USD per user/month)

Consolidated table with prices effective August 2, 2026. The "all-in" column is what a company pays today if you add base suite + Copilot, assuming list with no volume discount (which is the new default):

Three quick reads from the table:

• E3 vs before: +USD 3/user/month (+8.3%). At 1,000 seats that is +USD 36K/year just from the suite uplift, with no Copilot.
• E3 + Copilot vs before: used to be USD 66 (USD 36 + 30), now USD 69. An "apparent" 4.5% hike that hides the loss of the volume discount.
• E7 vs E5 + Copilot separate: USD 99 vs USD 90. You pay USD 9 more for the bundle, but you add Agent 365 (low-code agents), Entra Suite (ID P2 + Governance) and Purview upgraded (eDiscovery Premium tier). Defender for Identity and Purview eDiscovery already come with E5. Microsoft sells it as "cheaper if you buy everything."

The data point almost nobody looks at: the E3 + Copilot combination at USD 69/user/month is what gets replicated in press releases as "Copilot costs USD 30", but in practice a company coming from E3 pays USD 69, not USD 30. And the one coming from E5 pays USD 90. The magic number of "USD 30" is only the delta on top of a suite that just went up (source: Infosprint — Copilot price increase on idle seats).

What "Copilot at USD 30" means for your unit economics

The headline number is real but misleading. Here is the math across three company sizes, all against the new list price (no volume discount):

100-seat company (SMB / scale-up)

Reading: at 100 seats, the difference between E3+Copilot and E7 is USD 36K/year. If your company will not use Agent 365 or Entra Suite, E7 is not justified. The E3 + Copilot Cowork mix (60% of users on seat, 40% consuming per task) can lower the invoice 20% — but it requires instrumenting which tasks run from which account (source: Infosprint).

1,000-seat company (mid-market)

Reading: at 1,000 seats, E3 + Copilot costs USD 828K/year and E5 + Copilot costs USD 1.08M/year. Just the seat uplift from the E3 increase (no Copilot) is USD 36K/year extra — what the company pays "even though they did not add anything." This is the figure that hurts most in the CFO QBR: you are paying more for the same thing.

5,000-seat company (enterprise)

Reading: at 5,000 seats, bundle decisions move the budget in a USD 1.8M/year range. Here the conversation is no longer "what is best" but "what justifies each of the 9 extra dollars per seat between E7 and E3+Copilot." Companies in this range used to get multi-year prepaid commit discounts that Microsoft no longer offers, which puts aggressive renegotiation on the table for any EA expiring in the next 12 months (source: Samexpert).

Calculation disclaimer: the numbers assume list price with no volume discount, the new default since November 1, 2025. Companies with 3-year EAs already in flight may have different prices until their next anniversary. The Copilot Cowork scenarios assume a 60/40 split between seat and task-based — the real ratio depends on per-user adoption. Run your own calculation with your seat mix and adoption before committing.

What to audit in your SaaS this week

Operational checklist to run Monday through Friday. If you sell SaaS to enterprise clients in EU/US — or if your own CFO is asking for the "real all-in" on your internal tools — this is the list:

1. When does your client's EA expire? EAs do not all expire at the same time: each has its own anniversary date. If your enterprise client has 500+ M365 seats and an EA expiring between July 2026 and July 2027, they are in the renegotiation window with the new list price. Ask your account manager or reseller. If the EA auto-renews, the new list price applies automatically and your budget goes from USD 36 to USD 39 (E3) without anyone having signed an increase.

2. Does your own pricing assume an M365 seat uplift from the client? If your SaaS charges per M365 seat, per active M365 user, or per a feature that requires the client to pay more for M365, the change hits you twice: (a) your client pays more for the base, (b) you get paid less because the budget for your SaaS shrank. Review your pricing page and your proposals: if you say "X dollars per M365 seat", add an indexing clause to the Microsoft pricing guide so you do not get trapped in a margin that evaporates.

3. Can you offer Copilot Cowork as a low-cost upsell? If your product integrates with M365 (increasingly common in B2B SaaS sold to enterprise), Copilot Cowork at USD 0.70–6 per task (light/medium) gives you a low-cost entry for clients that do not yet have Copilot. Instead of "I sell you my Copilot integration, you need the seat at USD 30", you can say "start with Cowork, pay per use, and if ROI justifies, scale to seat." It is the enterprise equivalent of "freemium": you lower the entry barrier and let ROI prove itself.

4. Does your "value" narrative survive a tighter budget question? Your enterprise client's CFO will receive the Microsoft notice this month. The first reaction is "what can I cut". If your SaaS does not have a clear ROI per feature or per user, you are on the list. Prepare a one-pager on "what your company loses if you drop me" with concrete numbers, not adjectives.

5. Are you instrumenting which Copilot features your client uses? Microsoft is moving from seat-based to task-based. If your SaaS depends on Copilot for something specific (a workflow, an integration), instrument which Copilot tasks each client runs. You use it to (a) demonstrate ROI per workload, (b) defend your pricing with adoption data, (c) detect under-utilized clients before they themselves ask for a downgrade.

When NOT to use Copilot (neither seat nor Cowork)

Copilot is useful but not for everyone. Three cases where you should stay where you are or pick something else:

1. Companies with legacy volume licenses (Open Value, Select Plus). If your enterprise client is in a legacy 3-year contract with volume discount that runs through 2027, the July 1 uplift has not hit them yet (the increase applies to renewals, not contracts in flight). Buying Copilot at USD 30 today is paying premium for a suite that still has the old price. Wait for renewal and renegotiate everything together: suite + Copilot + E7 bundle. Negotiating incrementally is losing leverage.

2. Companies where Teams is not the primary work hub. Copilot monetizes through Teams, Outlook, Word, Excel, PowerPoint. If your client runs on Slack + Notion + Google Workspace (the typical pattern in tech-forward B2B SaaS), Copilot becomes "an assistant in another tool" that the user does not open. Typical adoption in this case is <15% at 6 months, and the CFO deinstalls it at the first renegotiation. Better: Google Workspace Duet AI or a direct LLM API depending on the use case.

3. Companies with strict data sovereignty (finance, healthcare, government). Copilot processes data through Microsoft data centers. If your client is subject to regulation that requires local data residency (LFEA in Argentina, LGPD in Brazil, HIPAA + specific BAAs in US healthcare, FedRAMP in US public sector), Copilot may not comply. E7 with Purview helps for eDiscovery, but the residency of training data or embeddings still belongs to Microsoft. Evaluate Azure OpenAI with dedicated tenant if you need the enterprise option.

When NOT to stay on Copilot out of inertia: if your internal adoption is <20% of seats after 6 months, you are paying for licenses that do not get used. Microsoft does not prorate: you pay 100% of the seat even if the user never opens it. This is the metric the CFO will look at in the October QBR, so better have it clear now.

Three operational practices that apply wherever you are

Regardless of your company size or which vendor you use, three optimizations apply to the enterprise licensing cycle that started this month:

1. Negotiate the EA before the anniversary date, not after. The leverage window is 90-120 days before expiration. Find out when your EA expires (ask your account manager if you do not know) and start the conversation 4 months ahead. Microsoft knows the customer is "trapped" near the hard date and tightens the discount accordingly. Companies that get the best EA results renew 60-90 days early, not on the anniversary date. With volume discounts gone since November 2025, the lead time matters even more.

2. Model TCO over 3 years, not 1. The July 1 price change is the first of an expected series. The cumulative 2024-2027 increase on the M365 suite (excluding Copilot) sits in the 25-35% range per public pricing guides. If you plan your M365 budget year by year, the shock hits every anniversary. If you plan it over 3 years with an 8-12% annual increase assumption, you can absorb it without renegotiating each time.

3. Measure adoption per user, not just total cost. The "every Copilot seat at 100% of price" pattern hides that 60-70% of users do not open Copilot after the first month. You pay USD 30 × 1,000 seats = USD 360K/year and have 350 active users. Microsoft does not solve this with automatic downgrade. The practice: monitor Copilot MAU/WAU per tenant, identify inactive users, and move them to Copilot Cowork (task-based) or remove the seat. It is the enterprise equivalent of cloud infra "right-sizing".

Conclusion

The headline number — USD 69/user/month with E3 + Copilot or USD 90 with E5 — is the real all-in a company pays today at list price. It is 4-5% more expensive than two months ago, and the volume discount elimination (November 2025) means the increase does not get recovered at renewal. For a 1,000-seat company, the seat uplift alone (E3 $36→$39, E3+Copilot $66→$69) is USD 36K/year extra "without having bought anything new".

Three operational recommendations to close out August:

1. Calculate your real all-in this week with your seat count and your suite + Copilot combination. If you do not have it on a single line in a dashboard, this is the week to build it. Your CFO will ask before your next EA anniversary.
2. Decide between the E7 bundle and the separate combination with your real usage mix. E7 makes sense if you actually use Agent 365 or Entra Suite. If not, you are paying USD 9/user/month extra for features you do not open.
3. Try Copilot Cowork before buying the full seat if your adoption is heterogeneous. It is the low-cost entry Microsoft offers exactly for this case. Paying USD 0.70–6 per task (light/medium) and validating ROI per workload is better than buying 1,000 seats and discovering at 6 months that half go unused.

If your SaaS sells to enterprise clients in EU/US and the M365 price hike is hitting your pipeline or your own CFO QBR, book a free 30-minute call — we usually identify 2-3 concrete moves on pricing or narrative in the first session.

---

Read also:

• Claude Sonnet 5 at $2/$10: review your COGS this week — if your SaaS runs on an LLM, the inference price drop shifted unit economics in July.
• 74% of your users are mentally replacing you — the BCG data on shadow AI adoption and why it redefines the ICP.
• 1,100+ OpenAI, Anthropic, Google and Meta employees ask the US government to slow down AI — the regulatory context that will move Copilot pricing over the next 12 months.
• How much does it cost to implement AI in a SaaS startup in 2026 — full project budget: M365 + LLM + infrastructure.
• Back to the blog — all articles.

Frequently asked questions

How much does Microsoft 365 Copilot actually cost per user per month?

The real all-in is USD 69/user/month with M365 E3 + Copilot, or USD 90 with E5 + Copilot, based on the new list prices Microsoft published on July 1, 2026. The "USD 30" figure in most press releases is only the Copilot addon on top of a suite that already went up (E3 USD 36 → 39, E5 USD 57 → 60) and lost the volume discounts that existed through November 1, 2025. The new M365 E7 bundle at USD 99/user includes Copilot and replaces the E5 + Copilot combination if you use the full stack.

When do the new rates apply to my EA?

EAs do not have a single global renewal date: each contract renews on its own anniversary date (the date it was originally signed, typically a 3-year cycle). The new rates from July 1, 2026 take effect at that next anniversary. If your EA expires between July 2026 and July 2027, you have to renegotiate sometime in that window with the new list price and no volume discount. Microsoft notifies 90-120 days before the anniversary; the leverage window is open until that point, not until a fixed calendar date.

Is the M365 E7 bundle at USD 99 worth it vs E3 + Copilot separately?

It depends on which features you need beyond E5. E7 includes E5 + Copilot + Agent 365 + Entra Suite + Purview upgraded for USD 99/user. If you already pay for E5 + Copilot, moving to E7 costs USD 9/user/month more but adds agent capabilities and ID governance. If you are on E3, E7 is 43% more expensive than E3 + Copilot (USD 99 vs USD 69) — only worth it if you actually need the E5 features and the E7 additions.

What is Copilot Cowork and how is it billed?

Copilot Cowork is the "task-based" model Microsoft launched as a complement to traditional seat-based licensing. Instead of paying USD 30/user for every seat, you pay per task executed by the agent based on weight: USD 0.70–2 for a light task, USD 4–6 for a medium task, USD 15+ for a heavy task. It is designed to validate ROI per workload before buying the full seat, and for companies with heterogeneous adoption where 100% seat coverage is not justified.

How does this impact a LATAM SaaS selling to enterprise clients in EU/US?

If your enterprise client pays per M365 seat, their 2026/2027 budget shrank 8-10% in real terms with nothing changing on their side. The "value vs cost" conversation gets harder: the CFO will demand cuts, and your SaaS competes for that same budget. Three practical moves: (1) offer Copilot Cowork as a low-cost entry point instead of full enterprise seats, (2) demonstrate ROI per feature / per user, not per plan, (3) prepare a "Microsoft inflation absorption" narrative if your own pricing is indexed to headcount or client seats.