Is Power BI dead?

Peter Verrykt is Business Director of Professional Services and Business Unit Lead for Data, AI & Automation at Xylos. He attended FabCon 2026 in Barcelona and shares what he saw there.

You might wonder whether Power BI still matters now that Fabric and AI are so central. That question might sound extreme, but it’s not entirely unfounded.

My answer after FabCon: yes, it still matters. But what Power BI is and how you work with it is changing.

Yes. Semantic models are becoming mainstream and are now the foundation of Fabric. That’s good news for anyone who’s been working with the Power BI data model for a long time.

At FabCon, there was also a strong focus on ontologies that you click together and on open standards for connecting to numerous external data sources. We were already pitching that idea 15 years ago; now it finally seems to be becoming a reality.

FabCon Europe 2026 in Barcelona.

Copilot efficiently and instantly solves complex DAX queries for you today, even when context transitions and filter contexts are involved. You can put together dashboards in just a few minutes. And the best part of all: AI in Power BI fits neatly within your M365 license in terms of capacity usage.

However, oversight is still necessary. You need to have a good understanding of what’s being built and how. If you start with a clear context and data you’re familiar with, you’re already helping yourself immensely. The time you save by no longer having to search for the right formulas can be invested in validating what’s been built.

If you’re purely a dashboard builder, I see a challenge. If the data is in good shape (because, yes, data is more important than ever), your dashboard will be ready in just a few minutes, complete with all the options for drill-through, parameters, and more.

Maybe it’s time to embrace AI and discover where you can reinforce each other’s strengths. Or to take a look at Fabric apps.

With Fabric apps, you can put together a complete app, complete with write-back, visuals, functions, and more. It’s never been easier. A blessing for some, a nightmare for others (read: DBAs, DPOs).

What makes me happy: you publish an app just like you publish a Power BI report. The only difference is that a full-fledged app goes live, with all the bells and whistles. The catch: control and manageability.

Personally, I see cost as the biggest risk. If you blindly rely on your agent (or agents, whether orchestrated or not), your capacity costs will quickly spiral out of control.

So pay attention to how much you interact with your data, because that’s where the big difference lies between an app and Power BI. Take an interactive map, for example: in Power BI, you can slice and dice the data at no extra cost, whereas a Fabric app goes back to the data with every click. If the app has a lot of users, you’ll quickly see that reflected in your capacity usage.

More and more is coming together in Fabric. The semantic layer seems to be becoming a fundamental part of the architecture. Will you stay loyal to Power BI or take the plunge into Fabric apps? It all depends on how you’re organized. Or rather, how you’re configured.

So Power BI isn’t dead. It’s just gained a much larger family. Those family members are all coming to visit, because they all want something from you. So take a moment to figure out who you’ll send to the wine cellar and who you’ll have help out in the kitchen, before it all turns into one big mess and everyone is sprawled all over your house.

My advice: start with data you’re familiar with. Decide in advance who’s allowed to use which data and who’s allowed to build what. And keep an eye on your capacity costs from the very beginning.

Our experts will help you map out your data environment and sketch the initial outline of your path to AI. Contact us for an exploratory conversation.