The news
Microsoft (MSFT) made Fabric IQ generally available in Copilot Chat and Cowork, grounding Copilot's answers in companies' Power BI data. Microsoft Fabric IQ is on by default for Fabric and Power BI customers, the company said in blog posts dated September 28 and 29.
The announcements came at the European Microsoft Fabric + SQL Community Conference in Barcelona. Microsoft describes Fabric IQ as a shared intelligence layer. It links data in OneLake, Fabric's central data store, with metrics from Power BI semantic models (the definitions and relationships behind a company's reports) and with ontologies, which map business concepts and how they relate. In Copilot, Microsoft said, a question about the latest sales forecast can be answered from the same model that powers the company's official reports.
Microsoft said the feature preserves existing access controls and governance, and that Copilot customers get Fabric insights without additional AI token costs. Integration with Copilot's new Code experience is coming through Microsoft's Frontier program. A related feature, IQ sharing, entered preview to let organizations share governed data and business context with customers and partners; Microsoft named EY, S&P Global and TomTom among companies building integrations.
On the database side, Microsoft said SQL Server on Azure Local, which runs SQL Server and Azure services in a customer's own facilities, is generally available. Its posts disagreed on the version for sites without cloud connectivity: one described SQL Server on Azure Local with disconnected operations as in preview, while the Azure blog called it generally available. Database Hub, a central console for managing SQL Server, Azure SQL, PostgreSQL and Cosmos DB databases, entered preview.
Microsoft said Fabric has more than 40,000 customers and that SQL Server supports enterprise workloads at more than 90% of the Fortune 500. CIO.com reported that Arun Ulagaratchagan, executive vice president of Azure Data, also cited 2.6 million active Fabric users when opening the conference, which drew more than 5,000 people.
The numbers
- Fabric customers
- More than 40,000 (Microsoft)
- Active Fabric users
- 2.6 million (Microsoft executive, per CIO.com)
- Power BI customers
- More than 425,000 (Microsoft)
- Fortune 500 companies using SQL Server
- More than 90% (Microsoft)
- Extra AI token cost for Fabric insights in Copilot
- None, Microsoft says
Why CEOs should care
The default setting is the first thing for CISOs and data leaders to check. With Fabric IQ on by default, employees using Copilot Chat can get answers drawn from Power BI models without anyone launching a project. Microsoft says existing permissions carry over, so Copilot's answers are only as well governed as those permissions. Audit who can see which semantic models, confirm that sensitive data is restricted at the row and model level, and ask Microsoft what controls administrators have over the default.
For CFOs, the promised upside is fewer meetings spent arguing over whose number is right, because Copilot's answers would come from the same model as official reports. That only holds if the semantic models are accurate and maintained, so the value depends on the quality of existing data-modeling work. Confirm the no-extra-token-cost claim against your own licensing before assuming Copilot usage of Fabric data is free at scale.
For CIOs, this raises the stakes of where governed business data lives. Copilot's usefulness is being tied to Fabric and Power BI, and IQ sharing extends that reach to partners. Companies that keep their semantic layer on other platforms should ask how Copilot will treat that data. Regulated businesses eyeing SQL Server on Azure Local should get written confirmation of the support status of the disconnected version before planning production use, given Microsoft's conflicting descriptions.
The bigger picture
Microsoft's argument is that as powerful AI models become widely available, advantage will come from a company's own data and context. Jessica Hawk, corporate vice president of Azure product marketing, wrote that this should work without making an organization dependent on any single model provider. The flip side is dependence on Microsoft's data platform: once Copilot, agents and apps draw their business meaning from Fabric, moving off Fabric gets harder.
Microsoft also said it is backing Apache Ossie, an incubating, community-led effort to set a vendor-neutral standard for exchanging semantic metadata, working with Snowflake and others. How far that standard goes will shape whether business definitions built in Fabric can travel to other tools.
What’s next
Microsoft said it will expand on these announcements at Ignite in San Francisco, November 17 to 20. Watch for documentation on administrator controls for Fabric IQ's default setting, the Code integration through the Frontier program, a single stated support status for disconnected SQL Server on Azure Local, and native IQ sharing support for Fabric IQ ontologies and Power BI semantic models, which Microsoft lists as on its roadmap.
What “Fact-checked” means
Fact-checking means testing a story’s facts against the evidence before it is published. This story went through at least two separate checks before this version was published.
- What we checked
- Its names, figures, dates, job titles, quotes and who said what were checked against the story’s sources, including its main source where it could be opened. The headline was checked for accuracy and overstatement.
- How
- A first check reviewed the whole story. If it passed, a second, skeptical check went back to the sources to look for mistakes in the most important facts. If a check flagged the story, it was edited to fix the problems found, and a separate re-check then reviewed the whole story again.
- Who
- The checks are made by our newsroom, as steps kept separate from the writing, under rules set by our editor, Hussein Mukhtar. A story the checks still flag is held for the editor, who decides whether it is fixed, published or dropped.
- If something is wrong
- “Fact-checked” does not mean error-free. If a material error is found after publication, we correct the story and add a note saying what changed. Report an error
Companies in this story
