Grasp AI Logo
Back to Blog
Cloud Strategy

Fabric February 2026: CMK Encryption, Python %run, and 30+ Platform Updates

Leon Godwin
11 March 2026

The Challenge

Monthly Fabric feature summaries are dense. There's always a long list of incremental improvements, and the real question for most teams is: which of these actually matter for our workloads? February 2026 brings over 30 updates across the platform — from notebook encryption to Data Factory performance tuning. Let me cut through the noise and highlight what's worth your attention.

What's Changed

Data Engineering Gets Serious Security and Modularity

Customer Managed Key encryption for notebooks is the headline here. Enterprise teams running Fabric in regulated environments — financial services, healthcare, public sector — now get notebook content and metadata encrypted at rest using customer-owned keys in Azure Key Vault. This has been a blocker for some organisations. It's no longer a blocker.

Python notebooks now support %run, which sounds minor until you've spent time maintaining duplicate utility code across 15 notebooks. You can reference and execute other notebooks within the same session, reusing functions and variables directly. It's the modular development pattern Python developers expect.

Full-size mode lets you expand a single cell to fill the entire notebook view. Small quality-of-life improvement, but if you've ever tried to refactor a 200-line cell while squinting at a narrow editor panel, you'll appreciate it.

Data Factory: Smarter, Faster Data Movement

Data Factory gets a cluster of practical improvements this month:

  • Adaptive Performance Tuning (preview) — Fabric now automatically optimises data movement operations based on workload characteristics. Think of it as auto-tuning for copy activities.
  • Parallel CSV reads — Large CSV datasets are now read in parallel, which matters more than you'd think when legacy systems still dump everything to CSV.
  • Column mapping for CDC in Copy Jobs — Change data capture workflows now support column mapping, making incremental loads more flexible.
  • Service Principal and Workspace Identity auth for Copy Job activities — no more shared credentials for automated pipelines.

Real-Time Intelligence: Private Network Streaming

The new Eventstream connectors for private networks let you stream real-time data into RTI from resources that aren't publicly accessible. For enterprises running VNet-integrated services, this removes a significant architecture workaround. Combined with dashboard performance improvements, the real-time analytics story in Fabric gets more practical for production use.

Platform-Wide: OneLake Catalog Completeness

With Workspace Apps (V2) now indexed in the OneLake Catalog, every item type in Fabric is discoverable from a single place. The catalog also surfaces richer metadata and lineage visualisation for all items, whether accessed from the catalog directly or from workspaces.

The Fabric Identity governance changes are worth noting too. The default limit jumps from 1,000 to 10,000 identities per tenant, and admins can set custom limits via the REST API. At scale, identity management matters.

Getting Started

If I had to pick three actions from this release:

  1. Test CMK encryption if you're in a regulated industry. Enable it on a non-production workspace first, verify your Key Vault integration, then roll out. Documentation here.

  2. Refactor your notebook utilities using %run. Pick your most-duplicated helper functions, extract them to a shared notebook, and reference them. Five minutes of setup saves hours of maintenance.

  3. Enable Adaptive Performance Tuning on your heaviest Data Factory pipelines. It's in preview, so test on non-critical workloads first, but the potential for automatic optimisation is worth exploring.

For the full list of changes, the official feature summary covers every item in detail. And with FabCon 2026 running March 16-20 in Atlanta, many of these features will be demonstrated live — use code FABCOMM to save $200 on registration.

What This Means

February's update is a maturity release. CMK encryption, identity governance at scale, private network streaming — these aren't flashy features, but they're the ones that get Fabric approved for production in enterprises that were previously holding back. The platform is steadily closing the gaps between "good for analytics" and "ready for regulated, at-scale deployment."

The Data Engineering improvements (Python modularity, VS Code MCP support, ODBC preview) also signal that Microsoft is investing in developer experience alongside the data platform. That's the right balance.

Leon Godwin, Principal Cloud Evangelist at Cloud Direct