Many GCCH organizations have built their Power BI environments on Dataflows Gen1, using it as a foundation for data preparation and reporting. With Microsoft’s shift to Dataflows Gen2, understanding how to approach a Dataflows Gen1 to Gen2 migration in GCCH is becoming increasingly important.
This isn’t just a routine upgrade. The move to Dataflows Gen2 through Microsoft Fabric introduces new architecture, governance models and scalability considerations. Successfully navigating a Dataflows Gen1 to Gen2 migration means going beyond a simple lift-and-shift and taking the opportunity to modernize your data strategy.
This is more than a migration; it’s a strategic opportunity to evolve
At first glance, it’s easy to interpret this as a straightforward upgrade: “Move Gen1 dataflows to Gen2 and continue operating as usual.” In practice, that framing will lead to problems.
Dataflows Gen2 is not a one-to-one replacement. It’s built on a fundamentally different architecture, tightly integrated with Fabric, leveraging lakehouse storage, compute separation and more scalable execution patterns.
That introduces both opportunity and risk:
Pipelines behave differently
Storage patterns change (OneLake vs hidden storage)
Refresh orchestration becomes more flexible, but also more complex
Governance and lineage shift from Power BI-centric to platform-centric
Organizations that treat this as a lift-and-shift migration often find themselves reworking things later. The more effective approach is to treat this as a redesign moment.
Why this matters more for GCCH customers
GCCH environments tend to have a few consistent characteristics:
Heavier reliance on Power BI + Dataflows as a primary data platform
This step alone often reduces migration scope by 20–40%.
2. They’re redesigning for Fabric, not recreating Gen1
Instead of asking “How do we move this?”, they’re asking “What’s the right pattern in Fabric?”
That often leads to changes like:
Moving heavy transformation logic into Lakehouse or Warehouse layers
Using Dataflows Gen2 for ingestion/light prep, not everything
Introducing medallion-style architectures (bronze/silver/gold) where appropriate
Leveraging OneLake to eliminate duplication across workspaces
The result is a more scalable, maintainable foundation, not just a migrated one.
3. They’re rethinking orchestration and refresh
Gen1 dataflows often relied on relatively simple refresh patterns. Gen2 introduces more flexibility, but also requires more intentional design:
Dependency management across pipelines
Scheduling strategies across Fabric workloads
Coordination with semantic models
Organizations that get this right see significant performance and reliability improvements. Those that don’t often experience instability early on.
4. They’re aligning governance early
Fabric changes where and how governance happens:
Data is more centralized (OneLake)
Access patterns shift
Lineage becomes more visible, but also broader
Successful teams are:
Defining workspace and domain strategies early
Aligning security models with Fabric architecture
Establishing naming, ownership and life cycle standards
Trying to retrofit governance after migration is significantly harder.
The real risk: waiting too long
The biggest issue we’re seeing isn’t technical, it’s timing.
Because Gen1 still works today, many organizations are delaying action.
But when you factor in:
The effort required to assess and rationalize existing dataflows
The learning curve associated with Fabric
The lag in GCCH feature availability
Competing internal priorities
…it becomes clear that this is not a “quick project.” Organizations that start late will be forced into rushed migrations, and that’s where poor design decisions get locked in.
A practical way to get started
If you’re a GCCH organization heavily using Dataflows Gen1, a strong starting point looks like:
1. Assess your current state
What do you have?
What’s actually being used?
Where are the critical dependencies?
2. Define your Fabric target architecture
How will you use Lakehouse, Warehouse and Dataflows Gen2?
What patterns will you standardize on?
3. Prioritize migrations
Start with high-value, lower-complexity dataflows
Avoid tackling the most complex pipelines first
4. Pilot in Fabric (GCCH environment)
Validate patterns with what’s currently available
Adjust for feature gaps or delays
5. Scale with intention
Don’t mass-migrate without validated patterns
Build reusable approaches as you go
How we can help
This shift is easy to underestimate because it’s framed as a product evolution. In reality, for many GCCH organizations, it’s the moment where Power BI-centric data platforms evolve into something much broader.
Handled well, this is an opportunity to modernize and simplify what has often become a very complex dataflow landscape. Handled poorly, or too late, it becomes a rushed migration with long-term consequences.
If you’re starting to think about what this means for your environment, we’re happy to share what we’re seeing across other GCCH organizations and how they’re approaching the transition.