In November 2023, we published an analysis of Microsoft Fabric as it entered General Availability. We were well-positioned to offer that perspective. A client had trusted us to implement the platform while it was still in Preview, giving our Data Intelligence practice more hands-on experience than most tech consultancies had accumulated at that point.
The piece gained traction on LinkedIn and caught the attention of Microsoft representatives in our regions. Since then, we’ve continued working with Fabric on select projects, delivered significantly more work across other Microsoft technologies, and maintained close attention to how the platform is performing in the market. We constantly supplement our direct experience by talking with current clients, prospects, former team members who’ve moved to other organizations, and even competitors.
It’s been just over 2years since that initial assessment. How have our insights held up? What’s changed, what hasn’t, and where is Fabric actually headed?
What We Said Then
On Platform Fit: “Fabric is an extremely appealing proposition for organizations already invested in Azure and utilizing their tool sets across various services, especially if your presentation and analysis tool of choice for users is Power BI.”
We still believe Fabric is appealing to organizations in this position. Note the important distinction between appealing and correct. Without substantial commitment to the Azure ecosystem already in place, a move to Fabric is difficult to justify. Enterprises using the Microsoft suite primarily through legacy or on-premises technologies often underestimate the actual lift required. At its core, Fabric operates as a fully managed cloud technology. This creates both advantages and constraints.
On Speed and Accessibility: “In our experience, increased performance and reduced time-to-insight has been amazing. As a services firm, it’s critical for us to be stewards of client resources, so completing a project in a fraction of the usual time is a significant benefit.”
This remains a genuine strength of Fabric. Experienced consultants and tech-savvy business users consistently report that a well-configured environment drives impressive speed-to-insight. Many platforms claim to democratize data access. Fabric can actually deliver on that promise in the right organizational context.
On Reliability Concerns: “This has come at the cost of reliability. Not every function works every time, and as with any managed service digging into what exactly went wrong is far more difficult than custom or fully high-code solutions.”
This is where our concerns have not only persisted but grown. Any consumer of cloud services hopes that reliability improves over time as the platform matures. Unfortunately, reliability issues continue surfacing in the field with concerning frequency:
- First-party connectors that previously worked with common data sources suddenly break, often with error messages that provide no actionable information
- Developers inadvertently consume an environment’s entire capacity with a single ingestion job or query, leaving both technical and business teams unable to work
- Unexpected changes in how fully managed configurations or deployment processes behave, sometimes without adequate documentation or warning
Some of these issues stem from intentional design decisions. Some represent user error or misunderstanding. Some are natural consequences of platform evolution. We understand how difficult it is to build a product that balances being “fully managed” with being “actually debuggable.” But for a platform that’s been in General Availability for 2years, the volume of negative feedback from practitioners is notably high.
Is Fabric Ready for Production?
We posed this question in our original analysis. Two years later, the answer remains: not quite, for most organizations.
Only organizations with existing, deep integration into Microsoft and Azure should consider it a genuinely reliable platform for core workloads. And even these organizations should recognize they’re landing on something that’s still actively evolving in ways that may create unexpected friction.
The real test is whether organizations are successfully running production workloads at scale. The evidence is mixed:
Pharmaceuticals company (previously committed to the platform): Abandoned Fabric for Databricks after extended evaluation. Their assessment: “It just doesn’t scale for our needs.”
Manufacturing organization (attempted major enterprise rollout): Pulled back to Azure Data Lake, Databricks, and other Microsoft technologies for core requirements. Their perspective: “It’s still part of our ecosystem and we’ve had success when we’re intentional about where we use it, but even simple operations will consume capacity unpredictably.”
Insurance carrier (recently deployed first production workload): Slowly expanding their engineering capabilities and organizational buy-in. Their view: “We genuinely like it, and the business is seeing value.”
Healthcare administrator (rapid development of small applications): Hit major obstacles after 6 months of work. Their experience: “We were able to get these applications running quickly, but we’ve faced ongoing issues with capacity consumption and DevOps integration.”
Some of these implementations began shortly after the GA announcement. Others started more recently. The mixed results suggest that success depends heavily on specific organizational context, use case characteristics, and realistic expectations about platform maturity.
The Competitive Landscape
Over the past 2 years, Snowflake and Databricks have continued evolving and winning significant business. As direct competitors to Fabric, it’s worth understanding where they stand:
Snowflake continues offering world-class training that enables talented teams to onboard quickly. Our consultants consistently report positive experiences working with the platform. The primary complaint from clients centers on cost management, particularly for organizations that jump in without understanding how to control consumption. Snowpipe capabilities are adequate for simple data movement but not as powerful as alternatives or as elegantly managed as tools like Fivetran.
Databricks remains, in our assessment, the most powerful platform for organizations running AI workloads on high-quality data or those with sophisticated data engineering requirements. Combining Databricks pipeline capabilities with Azure Data Factory has proven to be an effective balance of power and control across multiple client engagements.
Interestingly, we’re also hearing increasingly positive feedback about business users gaining access to Databricks and doing meaningful work within it when proper governance is established. This challenges the perception that Databricks is strictly an engineering platform.
Where Fabric Makes Sense
Based on our accumulated experience and market feedback, an organization should consider Fabric as a core platform if they can answer yes to all of these questions:
- Are we fully committed to Azure and the broader Microsoft ecosystem for the foreseeable future?
- Do we have strong, responsive support from Microsoft representatives and architects when questions or issues arise?
- Do we have an organizational culture that values speed and rapid iteration over absolute reliability?
- Do we have workload requirements that justify paying for F64 capacity or higher (starting at approximately $5,000 monthly)?
- Do we have a team with solid Microsoft expertise but a strong preference for low-code approaches that democratize how the business interacts with data?
For larger organizations requiring more rigorous governance, teams that prefer high-code platforms with more control, or smaller organizations spending less than $60,000 annually on data infrastructure, we would suggest allowing the product to continue maturing before making it central to your data strategy.
Looking Forward
Microsoft has substantial resources and strategic motivation to make Fabric successful. The vision of a unified data platform that serves both technical and business users remains compelling. But vision and current reality are not the same thing.
Organizations evaluating Fabric in 2026 should approach it with clear eyes about its current state: a platform with genuine strengths in specific contexts, but one that still requires patience, expertise, and realistic expectations about reliability and maturity.
The question isn’t whether Fabric will eventually become a robust, production-ready platform for a broad range of use cases. Microsoft’s track record suggests it likely will. The question is whether your organization can afford to be part of that maturation journey, or whether you need something more stable today.
