Dynamics 365 2026 wave 2: what's coming, and why there's no release plan

If you went looking for the "2026 release wave 2 plan" document this summer and came up empty, you didn't miss it. There isn't one, and there isn't going to be one. In August 2026 Microsoft announced that Dynamics 365, Power Platform, and Dataverse roadmap content is moving to the company-wide AI at Work roadmap, and that the twice-yearly release plan documents are being retired. The 2026 wave 1 plan was the last of its kind.

This guide covers what actually changed, what still ships this October, and how to run your update planning now that the ritual you built it around is gone.

The wave model, briefly

Since 2018, Dynamics 365 and the Power Platform have shipped on a public two-wave rhythm: wave 1 covering April through September, wave 2 covering October through March. Each wave came with a formal release plan published months ahead, an early access window roughly two months before general availability, and a mandatory-update model that meant every environment got the wave eventually whether you opted in early or not.

The model had real value: you could read, in one document, everything Microsoft planned to ship across Sales, Customer Service, Finance, Supply Chain, Business Central, and the Power Platform, with target dates and preview/GA status per feature. Implementation teams timed regression testing to the early access windows. It was predictable, and predictability is worth a lot in ERP.

What Microsoft announced in August 2026

The changes, concretely:

• No more release plans. From September 2026, new Dynamics 365, Power Platform, and Dataverse capabilities are published to the AI at Work roadmap — the successor to the Microsoft 365 roadmap — instead of wave plan documents. Disclosure becomes continuous rather than twice-yearly.
• The Release Planner retires on November 15, 2026. Existing roadmap items are being migrated to the new roadmap before then.
• Old plans stay up for reference. The published release plans, including 2026 wave 1, remain on Microsoft Learn as historical documents.

Microsoft's stated reason is one destination for roadmap content across business applications, productivity, and AI. The less charitable read — which plenty of the partner community voiced immediately — is that continuous disclosure means no single moment where a thin wave invites side-by-side comparison with the last one. Both things can be true.

What still ships in October

Here is the part that matters operationally: the update cadence has not changed. The retirement is a change to how Microsoft communicates the roadmap, not to how the products ship.

• Business Central version 29 is still expected in October 2026, on the usual April/October major-release rhythm. Based on what Microsoft has signalled and partner reporting, expect the emphasis to continue where version 28 left off: Copilot and agent capabilities, supply chain and warehousing refinements, and compliance/e-invoicing coverage. Treat any specific feature list you read before the release hits early access as provisional — without a wave plan, third-party summaries are doing more guessing than they used to.
• Finance and Supply Chain Management continue on the One Version continuous-update model, which was already monthly-ish and never really fit the wave framing anyway.
• The customer engagement apps and Power Platform keep shipping previews and GA features continuously, now announced through the AI at Work roadmap as they land.

Early access windows for the October updates still exist per product; they're just announced per product rather than as one coordinated wave event.

Hype versus shipped: how to read the new roadmap

The wave plans had a discipline problem in their final years, and it's fair to carry that scepticism forward: features were announced with preview dates that slipped, "planned" features quietly vanished between plan revisions, and the Copilot/agent items in particular tended to be described in the plan as more finished than they were in the product. The 2025 and 2026 wave 1 plans were heavy on agentic AI framing; the features that reliably arrived in GA form were the unglamorous ones — compliance, performance, warehousing, platform governance.

Continuous disclosure makes this pattern easier to hide and harder to audit, so adjust how you read:

• Trust GA status, not announcements. A roadmap entry means intent. Only "generally available" means you can build a project plan on it.
• Preview features are demos. Useful for evaluation, not for training material or cutover plans. This was true under the wave model and is truer now.
• Watch the message center, not the marketing blog. Admin-impacting changes (deprecations, mandatory changes, behaviour switches) come through the Microsoft 365 message center regardless of the roadmap format.

What to change in your own process

If your organisation timed testing and training around the two wave moments, you need a replacement rhythm, because Microsoft no longer provides one:

1. Keep the April/October beats for BC. Major Business Central releases still land then, so your regression-test calendar can stay put.
2. Set a roadmap review cadence. Someone should scan the AI at Work roadmap, filtered to your products, monthly or quarterly. The information still exists; it just no longer arrives as one readable document.
3. Subscribe to per-product release notes. Business Central's "what's new," the F&O One Version release notes, and the customer engagement release announcements are now the primary sources rather than summaries of a plan you already read.
4. Budget review time you used to get for free. Reading one wave plan twice a year was cheap. Continuous monitoring is a standing task — small, but real, and worth assigning to a named person.

The honest summary

Nothing about the software you run changes in October 2026. Business Central 29 ships, F&O keeps updating, the CRM apps keep shipping features. What's gone is the twice-yearly document that let you see the whole roadmap in one sitting and hold Microsoft to it six months later. That's a genuine loss for planning discipline, dressed as a convenience. Plan your own cadence, read GA status cynically, and keep your regression calendar anchored to the release dates that still exist.