Is the Microsoft Enterprise Agreement Still Worth It in 2026?

The 23% renewal budget hit nobody warned you about.

For most of the last twenty years, an Enterprise Agreement (EA) renewal was a structured negotiation inside a known framework. You had a Microsoft Business and Services Agreement (MBSA), an Enterprise Enrollment, possibly a Server and Cloud Enrollment, and a Customer Price Sheet (CPS) that fixed your prices for three years. The variables were the discount, the product mix, and a handful of negotiated concessions. The destination, a renewed EA, was rarely in doubt.

That world ended on 1 November 2025. And the next twelve months are going to make the change uncomfortably visible to every procurement and IT leader whose agreement renews in 2026 or 2027.

The change is not one event. It is four overlapping shifts, each individually defensible from Microsoft's point of view, and collectively the most significant rewrite of enterprise licensing economics in a generation.

What Actually Changed, and What's Still in Motion

Before we get to strategy, let's lay out the facts. There are four moving pieces, and they compound.

1. Volume discounts on Online Services are gone

On 1 November 2025, Microsoft eliminated programmatic Levels B, C, and D pricing for Online Services across the Enterprise Agreement, the Enterprise Subscription Enrollment, the Server and Cloud Enrollment, and the Microsoft Products and Services Agreement (MPSA). The levels still exist in name. In practice every customer pays Level A, which is to say Microsoft.com list price, at next renewal or when adding a service not already on the Customer Price Sheet.

For Level B customers this is roughly a 6% increase. For Level D, it's around 12%. On-premises licensing is unaffected. So is US Federal, State and Local Government, and Education. Everyone else is paying list.

2. Microsoft 365 list prices rise on 1 July 2026

On 4 December 2025, Microsoft announced that Microsoft 365 E3 will rise from $36 to $39 per user per month (an 8.3% increase) and E5 from $57 to $60 (5.3%). Frontline SKUs see steeper percentage rises, up to 33% on M365 F1. The new pricing takes effect at next renewal after 1 July 2026.

Read those two changes together. A 25,000-user E5 customer who was on Level D pricing prior to November 2025 was paying around $50 per user per month effectively. At their next renewal after July 2026, they're paying $60. That's a 20% effective increase on their largest software line item, without anything changing in their estate. Unified Support, priced as a percentage of Microsoft spend, follows the same curve.

3. Microsoft 365 E7, the "Frontier Suite", launched 1 May 2026

Announced on 9 March 2026, M365 E7 bundles E5, Microsoft 365 Copilot, Microsoft Entra Suite, and Agent 365 (Microsoft's new AI-agent governance plane) into a single SKU at $99 per user per month. Bought separately under July 2026 pricing those four components total roughly $117, so the bundle saves about 15% if you actually use everything in it.

E7 is generally available on EA, Enterprise Agreement Subscription (EAS), MCA-E, and CSP from 1 May 2026. CSP launch promotions through 31 December 2026, including a 15% triennial discount, can drive the effective rate down to around $84 per user per month for qualifying seat counts. After that, the discount window closes.

4. The agreement stack has fragmented

The EA isn't the only outcome at your next renewal anymore. Five are now in play, and not all of them are yours to choose:

• EA renewal on negotiated terms: still possible, particularly above 2,400 users, but Microsoft is selective about who gets one.

• CSP (Cloud Solution Provider): partner-managed, Microsoft Customer Agreement (MCA)-based, no minimum seat count.

• MCA-E (Microsoft Customer Agreement for Enterprise): Microsoft direct, also MCA-based, no Licensing Solution Provider (LSP) in the room.

• MS-EDS (Microsoft Enterprise Direct Services), also called the S500 programme: you keep an EA contract stack, but the LSP is removed and Microsoft manages the relationship directly.

• Forced MCA migration: for organisations below Microsoft's evolving EA eligibility thresholds, EA renewal is simply declined. From 1 March 2026, Microsoft has been moving EA customers on Azure Consumption Commitment (MACC) plans to MCA-E ahead of their renewal date.

The first two you can pursue. The last three Microsoft can impose. Knowing which conversation you're walking into is now the first piece of preparation, not a footnote at the end.

Why the EA Was Worth Defending, and Why It Still Is for Some

I want to be careful here, because most of the noise online is partner content telling you to move to CSP. Some of it is reasonable. Most of it isn't catered to you; it's catered to the partner's commercial model.

The traditional EA still has structural advantages that nothing in the MCA-based world replicates:

• Three-year price lock. The July 2026 increase doesn't apply to EA customers whose contracts started before then. MCA-E and CSP annual subscriptions have no such protection.

• Software Assurance (SA) benefits: License Mobility, dual-use rights, and most importantly, Azure Hybrid Benefit. SA is not available in CSP at all, and is not a standard MCA-E feature.

• From-SA licensing. If you migrated from Office on-premises to Microsoft 365 on an EA, you're likely entitled to Office Professional Plus on top, via the From-SA discount. CSP does not offer that. Move to CSP without addressing it and you have a compliance gap.

• Core Infrastructure Suite (CIS). The CIS bundle of Windows Server Datacenter and System Center, with its associated discount, doesn't exist in CSP.

• Mid-term reduction at anniversary. EA lets you reduce subscription quantities at each annual anniversary within the three-year term. CSP New Commerce Experience (NCE) annual subscriptions lock you in for the full twelve-month commitment; you can only adjust quantities when the subscription comes up for renewal, not during it.

• Rights to run on-premises Office Servers. Microsoft 365 E3 and E5 in EA grant the right to run unlimited on-premises Exchange and SharePoint for licensed users. The same SKUs in CSP do not.

• Unified Support negotiated alongside the EA. The Unified Support angle deserves its own section below.

A word on Unified Support

Unified Support is technically channel-agnostic. It is available to customers on EA, MCA-E, CSP, and Azure-only agreements. So strictly speaking, it is not an "EA-only" benefit.

In practice, however, the relationship runs much deeper inside an EA. Unified Support is priced as a graduated percentage of your annual Microsoft spend (starting at 7.5–10% and falling with scale), and Microsoft routinely co-terms the Unified Support agreement with the EA renewal date. That co-termination is the key. It makes Unified Support a negotiable lever inside the EA conversation: pricing, scope, response service level agreements (SLAs), and proactive service hours can all be put on the table as part of the broader commercial discussion.

When customers move to MCA-E, that bundling tightens in ways that disadvantage the customer. In our experience, multi-year Unified Support agreements tied to MCA-E typically include a true-up clause that recalculates support cost upward when Azure or licence consumption grows beyond a small buffer (commonly 5%). There is no equivalent true-down if your spend drops. The ratchet only turns one way. Inside an EA, the renewal cycle gives you a structured, three-year point at which to re-bid Unified Support, including against established third-party support alternatives recognised by Gartner. Outside the EA, that periodic reset is much harder to engineer.

For large enterprises that consume Unified Support seriously, the co-termination is an underappreciated reason to keep the EA in play. The EA is not just where your licences live; it is where you negotiate your support relationship from a position of leverage.

Same products. Same support model. Different commercial mechanics.

For a large, hybrid enterprise still running meaningful on-premises infrastructure, with serious Azure Hybrid Benefit savings to protect and a substantial Unified Support footprint, the EA is frequently still the right vehicle in 2026, if you can secure one and if you negotiate the discount that the programmatic levels used to provide automatically.

Why CSP Now Deserves a Seat at Your Negotiation Table, Even If You Stay on EA

You don't need to migrate to CSP to benefit from understanding it. You need to understand it because it's now the best alternative to a negotiated agreement (BATNA) that you can present opposite Microsoft.

Three reasons.

First, the price gap has closed. The volume-discount advantage that justified the EA on commercial grounds alone is gone for Online Services. Above 2,400 users, EA still typically wins on negotiated discount and the three-year lock. Below that, the per-seat price difference between EA Level A and CSP from a competent enterprise CSP partner is often within negotiating distance.

Second, CSP educates you on the MCA, and the MCA is the contract foundation under MCA-E. Two of the five renewal outcomes I listed above sit on the MCA, not the MBSA + EA + Enrollment + Product Terms stack you've been operating under. The MCA is evergreen, has no fixed term, no true-up, no Customer Price Sheet, and product use rights that Microsoft can update multiple times per year. If Microsoft initiates an MCA-E conversation with you in the next twelve months, you do not want that to be the first time you've read an MCA. Working through a serious CSP evaluation with a qualified partner forces your legal and procurement teams to read the contract that will govern half your possible futures.

CSP isn't a migration target — it's the contract you must read to negotiate the contract you actually want.

Third, a documented alternative changes the conversation with Microsoft's commercial team. A documented CSP quote from an enterprise-grade partner, not a price card from your reseller's CSP arm but a real proposal with transition methodology, pricing, and references, is a negotiation lever that did not exist for most customers five years ago. You don't have to use it. You need it in your back pocket.

Mixing Programmes: A Hybrid Strategy Worth Considering

Most of the EA-versus-CSP-versus-MCA-E debate online is framed as a binary. It isn't. Microsoft permits, and many large customers actively use, a hybrid approach: pin your stable, organisation-wide commitments inside the EA, and place the volatile, uncertain, or pilot products in CSP on annual subscriptions where you can true them down at each anniversary.

The practical consequence is immediate. The EA's biggest weakness has always been that you cannot reduce subscription quantities mid-term. If you commit 5,000 seats of Power BI Pro for three years and adoption stalls at 1,800, you pay for 5,000 until renewal. CSP annual subscriptions, by contrast, allow non-renewal at each twelve-month anniversary. A product you bought twelve months ago can be reduced or removed entirely at the next anniversary, with no penalty beyond the term you already paid for.

A practical scenario. Take a 6,000-user enterprise renewing its EA in late 2026:

• In the EA: Microsoft 365 E3 for the full 6,000-user base. Demand at this scale is stable, the discount is meaningful, the three-year price lock is valuable, and the full Software Assurance construct (Azure Hybrid Benefit, From-SA, License Mobility) is native to the EA. CSP server subscriptions carry partial equivalents from April 2026, but From-SA and the broader SA benefit set do not exist outside the EA.

• In CSP, on annual subscriptions: Microsoft 365 Copilot for 600 users (10% pilot population), Power BI Pro for 1,200 users, Visio Plan 2 for 400 users, Microsoft 365 Copilot for Sales for a 150-user pilot in the commercial team.

The Copilot pilot is the cleanest example. Adoption forecasts for Copilot remain unreliable across the industry. Microsoft itself acknowledged roughly 3% of M365 business subscribers had bought Copilot seats as of early 2026. Committing 600 Copilot seats for three years inside an EA at $30 per user per month is roughly $648,000 of three-year exposure. If adoption proves real, you fold those seats into the EA at the next true-up or renewal. If it doesn't, you exit at the twelve-month CSP anniversary and you've paid for one year of pilot, not three.

The same logic applies to Visio, Project, Power BI Pro, Dynamics 365 modules, and the various Copilot variants for Sales, Service, and Finance. These are products where adoption uncertainty is genuine, where Microsoft is most aggressive about committing you to multi-year volumes for "discount reasons", and where a CSP annual subscription gives you a clean exit point each year.

There are caveats, and they need flagging:

• Customer Price Sheet discipline. Anything you might want to add to the EA later should be on the CPS now, while you can still negotiate price. Once a product is omitted, adding it later means full Level A list price.

• Avoid double-counting. A user licensed for Microsoft 365 Copilot in CSP cannot also be counted in an EA Copilot order. Track assignments centrally.

• Different terms in different channels. The same SKU bought in EA versus CSP carries different rights. Office Professional Plus rights via From-SA are an EA construct. License Mobility for Windows Server / SQL Server subscriptions in CSP was only explicitly confirmed from 1 April 2026. Verify rights before you split a workload.

• Two channels, two true-ups. EA true-ups happen at anniversary. CSP renewals happen on the subscription anniversary, which may not align. Plan the cadence.

For organisations that previously committed entire pilot populations to the EA "because it was easier", a thoughtful EA + CSP split is one of the few clear cost-optimisation levers left in the post-November 2025 environment. It will not negate the discount-removal hit. It will materially limit the cost of being wrong about adoption.

The Microsoft 365 E7 Wrinkle

E7 is not a renewal of E5. It is a new premium suite that bundles E5 + Copilot + Entra Suite + Agent 365 at $99. For a customer already running E5 with a meaningful Copilot deployment, the bundle maths work in your favour at list prices: you save roughly $18 per user per month versus buying the components à la carte at July 2026 rates. Note that Microsoft offers special Entra Suite pricing for E5 customers, which narrows the gap.

But E7 is also a strategic move by Microsoft. Copilot adoption has lagged. Roughly 3% of Microsoft's 450 million M365 commercial paid seats had bought Copilot as of early 2026. Bundling Copilot into E7 reduces the per-component friction and accelerates Microsoft's AI revenue. Agent 365, the new control plane for AI agents, became generally available on 1 May 2026 and is positioned as a governance requirement for organisations deploying AI agents at scale.

Three things to keep in mind before you sign anything that mentions E7:

1. The 15% triennial CSP promotion expires 31 December 2026. If E7 is in your roadmap, the negotiating window is narrower than your renewal cycle.

2. The bundle only saves money if you'll actually use Entra Suite and Agent 365. If you are on E5 today but your identity stack uses only Entra ID P1 and you have no agent governance pain, E7 is a 65% upgrade in price ($60 to $99) for a bundle discount you won't realise.

3. E7 is being marketed as the simple, modern choice. It is also Microsoft's vehicle for getting more spend per user into your contract before your three-year price lock crystallises. Treat it as a negotiation input, not a default.

For the right customer, E7 is a good deal. For the wrong customer, it's a 65% price uplift dressed as simplification. Run the maths against your actual Copilot adoption rate and your real Entra and agent governance requirements before agreeing to it as the renewal SKU.

The Compounding Cost: A Worked Example

Putting all four shifts together, here's what a 10,000-user E3 customer who was on Level C pricing looks like across the cycle:

Sources: Microsoft licensing news (E3/E5 list prices); discount removal (Level C baseline). Copilot at $30/user/month list.

The discount removal and list-price increase alone produce a 20% increase against the pre-November 2025 baseline. Add even 3% Copilot adoption and the figure reaches 23%, before you consider Unified Support uplift, MACC requirements, or the operational cost of an MS-EDS or MCA-E transition. For a CFO who budgeted Microsoft as a known quantity, the increase is a multi-million-pound surprise spread over two renewal events.

Now apply this exercise to your own organisation, with your actual user mix, before Microsoft's account team applies it for you.

What Preparation Now Actually Looks Like

The traditional EA preparation playbook, utilisation audit, scenario modelling, cross-functional team, a year of lead time, still applies. It's necessary. It's no longer sufficient.

What you need to add for a 2026 renewal:

Read your Customer Price Sheet against November 2025. Anything you might want to add to your agreement before its end-of-term should be on the CPS now. Online Services not already listed on it default to Level A pricing on the next addition. Reviewing the CPS is the cheapest defensive action available, and the window is closing as your renewal approaches.

Build a costed CSP scenario in parallel with your EA renewal model. Not a price comparison from your incumbent reseller. A real evaluation with at least two enterprise-grade CSPs, including migration planning, From-SA gap assessment, CIS replacement plan, and a clean opt-out from your expiring EA. A serious CSP evaluation is also your best route to understanding the MCA-based contract framework if Microsoft pivots to MCA-E mid-cycle.

Design a hybrid programme model. Decide which products belong in the EA (stable, organisation-wide, SA-dependent) and which belong in CSP (volatile, pilot, project-based, Copilot variants, Power Platform add-ons). Build the CSP cadence into your annual operating rhythm.

Run the E7 maths twice. Once at full $99 list, once at the 15% triennial promo rate. Compare both against E5 + Copilot priced separately at July 2026 rates, weighted by your actual Copilot adoption. Then decide whether E7 is a buy, a hold, or a "show me numbers in 2027".

Identify which involuntary path is most likely for you. If you're under 2,400 users, plan for an MCA-based outcome, either CSP or MCA-E. If you're above 2,400 with significant Azure spend, prepare for MS-EDS conversations and decide whether you want to fund independent LSP-style support if Microsoft removes it. If you're in the strategic-account band, expect direct Microsoft engagement with reduced LSP involvement.

Co-term Unified Support with your renewal. It's a percentage of your Microsoft spend; the increases compound. If your Unified Support renewal sits on a different cycle, align it, and benchmark third-party alternatives in parallel.

The Bottom Line

The 2026 EA renewal is not a renewal. It's a programme selection exercise dressed as a renewal, conducted in a pricing environment Microsoft has reset in its own favour, with a new premium SKU launched mid-cycle and several involuntary outcomes Microsoft can impose if you don't show up prepared.

The customers who will come out of 2026 in good shape are not the ones who fight hardest for the EA they've always had. They are the ones who understood the new agreement structure early enough to negotiate from genuine optionality — EA, CSP, MCA-E, MS-EDS, or some combination — with all the contractual and commercial trade-offs documented in advance.

The customers who will come out badly are the ones who treat this as last cycle's renewal with a slightly higher number on the front page. Microsoft is counting on a large share of customers in that second group. Don't be one of them.

Q&A: 2026 EA Renewal, Common Questions

Q: My EA renews in late 2026. Should I try to renew early to avoid the July 2026 price increase?

A: Possibly, but only if the early renewal is genuinely on terms equal to or better than what you would negotiate at the natural date. Microsoft's account teams are aware of the timing pressure and will frequently propose early renewals on weak commercial terms because they know the customer is anxious. The pre-July 2026 list prices are worth locking in only if the rest of the deal stands up. Run the comparison both ways.

Q: We have fewer than 2,400 users. Is the EA still an option for us?

A: Increasingly not. Microsoft has been declining EA renewals for organisations under 2,400 users since early 2025 and has been moving existing EA customers in this band to MCA-E from March 2026 onwards. If you are between 500 and 2,400 users, your renewal will most likely be a CSP-versus-MCA-E conversation, not a CSP-versus-EA one. Prepare accordingly.

Q: What happens to our Software Assurance investment if we move to CSP or MCA-E?

A: Software Assurance does not exist as a separate construct in CSP or MCA-E. Subscription licences in those programmes include some equivalent rights, version upgrades for example, but the broader SA benefit set (From-SA, License Mobility on legacy volume licences) is forfeited at transition. Run a benefit-by-benefit gap assessment before signing anything.

Q: Can we keep our Azure Hybrid Benefit if we move servers to CSP?

A: Indirectly, yes. CSP now offers Windows Server Datacenter and SQL Server subscriptions whose terms were clarified from 1 April 2026 to confirm they carry Azure Hybrid Benefit and License Mobility equivalents. These are subscription-based equivalents, not the same as perpetual licence + Software Assurance. The EA route remains commercially more attractive in many scenarios where the on-premises footprint is shrinking gradually rather than overnight.

Q: How is Copilot priced under EA versus CSP?

A: Standalone M365 Copilot is $30 per user per month at list across all programmes. The differences are commercial. In EA, you typically commit Copilot quantities for the agreement term. In CSP annual subscriptions, you can true those quantities down at the next anniversary. Microsoft is also running CSP-specific Copilot promotions through 30 June 2026, including 30% and 40% off Copilot for All for qualifying customers, that are not always available inside the EA.

Q: We are being moved to MS-EDS. What changes operationally?

A: The contract stack stays the same: MBSA, EA, Enrollment, Product Terms. What changes is the relationship. The LSP is removed as the contractual partner of record, and Microsoft handles transactional support, contract management, and renewal negotiation directly. Some customers welcome the simplification. Others find the LSP's hand-holding hard to replace, particularly for true-up management and licensing advice. Some former LSPs now offer paid advisory services to MS-EDS customers; that cost is itself a negotiation data point with Microsoft.

Q: Is Unified Support negotiable as part of the EA renewal?

A: Yes, and it should be. Unified Support is priced as a percentage of your Microsoft spend, so as your EA grows, Unified Support grows automatically. Co-terming Unified Support with the EA renewal date gives you a structured point to benchmark pricing, scope, and SLAs, including against third-party support providers recognised as comparable by independent analysts. Large enterprises should consider a request for proposals (RFP) to test the market every renewal cycle.

Q: What is the single most expensive mistake we can make in the next twelve months?

A: Treating the renewal as last cycle with a higher number on it. The structural changes — discount removal, list-price increase, fragmented agreement stack, new premium SKU — mean that doing what you did last time will not produce what you got last time. Start preparation nine to twelve months ahead. Build genuine optionality. Do not let the absence of a credible alternative be the reason you accept Microsoft's first proposal.

Need an independent view of your 2026 renewal options? We don't sell Microsoft licences and we are not a CSP. Talk to us about Microsoft Enterprise Agreement Negotiation or Pricing Research and Deal Benchmarking.