It took a couple of weeks for Microsoft to put out a blog post that outlines what they already “leaked out” in the January 2026 Power Platform licensing guide update (see my earlier post). On January 14, we got two blog posts that are both viewable for the entire online audience. The customer version is shorter, so let’s instead look at the partner alert:
“As part of Microsoft’s efforts to increase value for enterprise-wide use of Microsoft Power Apps and reduce licensing complexity, we are announcing that the Power Apps per App SKU was removed from customer price lists effective January 2, 2026. Existing customers with this SKU will not be impacted and can continue to access and renew their existing licenses, as there are no end-of-life plans for this SKU.”
Taking away the cheaper $5 license option selectively from new customers is not my idea of how to reduce complexity of Power Platform licensing, but okay. Whatever. As discussed earlier, the $10 option of per app pay-as-you-go meter and the full $20 Power Apps Premium remain on the price list for everyone.
Existing customers: End-of-sale vs. End-of-life
These are two commercial concepts that Microsoft has published definitions on. Under “Manage end-of-sale offers”, we see that end-of-sale (EOS) offers are “no longer available for purchase,” while existing subscriptions aren’t affected and “services continue to work” after moving to EOS. The intention is explicitly to disallow new purchases but still support renewals.
End of life (EOL) means offers are “no longer available for purchase” (same as EOS) and “existing subscriptions are not renewed.” That last part is the crucial difference here when it comes to Power Apps per app licenses. A license that goes EOL will only work until the end of your subscription period.
If you currently have Power Apps per app in use and are coming to the renewal date of your Microsoft contract, these won’t disappear. EOS does not mean you’d need to freeze your license count either. Existing Microsoft customers commonly can add seats/quantity on what they already own, at least within their agreement/subscription rules.
How about new customers?
If you were just planning to go live with Power Apps based solutions and hadn’t yet bought the per app licenses, tough luck! Or is it?
Microsoft’s customer-facing blog post clearly says: “Effective January 2, 2026, the Power Apps per app SKU is no longer available for purchase by new customers.” Which is a bit strange, given how the product has remained in the catalog during January. Declaring that you’ve retroactively stopped selling something is, at least to me, a sign of things not proceeding according to plan.
Since the Microsoft 365 admin portal’s Marketplace still showed the Power Apps per app license as an option to purchase in my tenant, I decided to give it a go. The following video demonstrates how things were on January 15, one day after the blog posts:
What can we learn from all this? At least the fact that if you want to become an existing customer in the eyes of Microsoft, it’s not a bad idea to go and purchase your first Power Apps per app plan license “while stocks still last”!
Is there a future for per app licensing?
It doesn’t give us much comfort how Microsoft says “there are no end-of-life plans for this SKU”. Because up until January 2026, there weren’t any plans for the end-of-sale either. There were plenty of rumors, yes. We didn’t know exactly when the EOS date would be upon us yet those who follow the business apps products closely were able to assume this was likely to happen at some point.
Grandfathering is a concept that Microsoft has applied to Power Apps licensing already before. Back in 2019 when the SQL Server connector was classified as a premium connector, many customers who were using the seeded Office licenses for Power Apps and Power Automate faced a similar situation. The decision Microsoft made was as follows:
“In addition, apps and flows created prior to October 1, 2019, that are using these connectors will receive an extended transition period until October 1, 2024. During this time, these qualifying apps and flows will be exempt from the Premium connector licensing requirements for the reclassified connectors, custom connectors and on-premises data gateways.”
The customers were given a five-year exemption from the premium classification. Specifically for the apps and flows they had created earlier. While new solutions in the tenant required proper Premium licenses, the specific resources from earlier were whitelisted from licensing enforcement.
While it’s not exactly an identical scenario, that gives us a reference point of a five-year transitioning period. Who knows, perhaps the existing Power Apps per app customers could still be on this plan in 2031 then? All we can do is speculate at this point, since “there are no EOL plans” according to the latest guidance.
If the Power Apps product offering had been “reimagined” rather than merely “simplified”, we might have seen something similar as what Microsoft did with Power BI Premium capacity. In March 2024, it was announced that existing customers were able to renew their Power BI Premium subscriptions until the beginning of February 2025. From that point onward, the only option given was to move to Microsoft Fabric capacity instead.
Any way you look at it, the per app passes that were originally a tempting way for customers to dip their toes into low-code applications are now on borrowed time.
More than ever, it’s about the Platform
In the partner message, Microsoft is clear about the problems that these low-commitment licenses for ad-hoc Power Apps premium feature use (such as Dataverse and model-driven apps) presented to their overall story:
“In addition to licensing complexity, the Power Apps per App SKU had the potential to introduce scalability and governance challenges that slowed innovation and reduced overall platform return on investment for customers. In place of this SKU, we recommend that you position Power Apps Premium SKU for enterprise customers.”
The Managed Environments initiative has been used to gently nudge customers away from trying to make do with the seeded apps and flows capabilities in Microsoft 365 licenses for almost four years now. It has been the key argument for why enterprise customers should buy “proper” licenses for their users. The governance, administration and ALM capabilities that are locked behind the Managed Environments gate have been steadily growing. While not fully replacing tools like the CoE Starter Kit for governance scenarios, most Power Platform developers today would likely recommend enabling the ME switch for solutions they build.
For an up-to-date overview of what the Managed Platform story consists of, you can watch this Ignite 2025 session where Ryan Jones and his crew demonstrate the powerful capabilities behind the simple looking apps and flows:
Where the messaging around the per app EOS falls a bit short is that the Power Apps per app license does allow the use of Managed Environments today. So, any argument you hear from MS or partners about “we need Power Apps Premium licensing to leverage the managed platform features” can be rightfully challenged. Give them a link to this blog right here, for example.
The fundamental issue with per app, the way I see it, is that it positions Power Apps against vibe-coded React apps. That can lead customers to question why do they even need a GUI-based app creation toolkit like Power Apps anymore. Especially when the alternative would be a custom app that has no per-user or per-app price tags attached to it.
Microsoft is addressing this scenario with Code Apps. Currently still in preview, this new type of Power Apps is aimed at the code-gen path where the apps aren’t built with a dedicated GUI studio experience (like canvas apps) but rather via tools like GitHub Copilot or Claude Code. Power Platform then serves as the management layer that provides functionality not easily available in pure vibe code apps: identity, security, connectors to enterprise data sources, observability, and so on.
Remember that these preview products are usually pending for the final licensing model to be published by Microsoft. Once Code Apps becomes generally available, it would not surprise me if there are some new licensing variables introduced. For instance, we should not expect these Code Apps to support the EOS per app model anymore. Which would once again be a further argument for why customers should invest in the full Power Apps Premium licenses with a per-seat model.
So, what SHOULD customers do then?
There has been a huge push for Copilot in the past couple of years, yet the adoption rate remains far too low for the vision that Microsoft has in generating revenue from AI. While tech vendors claimed that 2025 would be the year of AI agents, it turned out to be more like the year of vibe code. Which was powered by AI coding agents. Claude Code was leading the way to what is now turning out to be a power user paradigm for working with much more than merely code artifacts.
At times like these, I don’t believe it makes sense to commit to a specific technical pattern or license type. Microsoft 365 Copilot faces a long road to broad adoption and its included features keep evolving and shifting. Meanwhile, Copilot Studio carries a part of the Power Virtual Agents chatbot era legacy with it. When LLMs have become good enough in code generation for many professional developers, the value of user interfaces that Microsoft tries to stitch together from AI models and existing business apps is… Well, it’s not guaranteed to be the winning ticket.
It’s more likely that the business application UIs that customers have built for themselves, for their own processes, remain relevant for many years. Replacing them with unproven AI chat experiences could lead you to just paying for both the old app and the new agent. Meanwhile, the platform level services behind Power Apps are a key differentiator for Microsoft cloud. They are the reason why vibe coding experiments to build a V2 of the business apps aren’t necessarily going to be as easy as they seem at first.
All in all, holding on to Power Apps and seeing what another year in the LLM race brings to the world sounds like the smartest move. New Microsoft capabilities like Code Apps will get to demonstrate their viability during 2026. PAYG model for consumption-based licensing could become more natural as Copilot Credits need to be purchased anyway. As always, there’s a lot of moving parts to consider. If you feel like you could use one-to-one guidance on these topics, be sure to reach out.
