Microsoft Ignite lands in San Francisco on November 18–21, 2025, and the data stack is poised for some long‑awaited “flip the switch” moments. If even half of the current previews in Fabric and Power BI cross the finish line, day‑to‑day analytics work gets simpler, faster, and easier to govern. Here’s what we’ll be watching most closely.
Below are the five GA announcements we’re hoping to hear on stage—and why they matter in practice: Cosmos DB Mirroring, Materialized Lake Views, Lakehouse Schemas, Calendar‑based Time Intelligence in Power BI, and one more preview with outsized impact: OneLake Security.
1) Cosmos DB Mirroring
Why it matters – Fabric Mirroring is already the “no‑ETL” bridge that lands operational data into OneLake as Delta, enabling Direct Lake analytics without copy‑heavy pipelines. As of September, Cosmos DB mirroring is still labeled preview; GA would bring the SLA, scale, and support story enterprise teams need.
What it unlocks:
– Near–real‑time replication of Cosmos data into OneLake as Delta tables (v‑order optimized), ready for Spark, SQL, and Direct Lake.
– Cross‑database queries that join mirrored tables with warehouses and lakehouses in plain T‑SQL.
What we’ll verify on day one:
– Supported Cosmos APIs and region parity at GA. Throughput/RU impact and failover behavior for multi‑region accounts. Monitoring, retention defaults, and API surface for CI/CD.
– Support for Private Link
2) Materialized Lake Views
Why it matters – Materialized Lake Views (MLVs) bring a declarative layer for medallion pipelines: you define intent; Fabric handles orchestration, incremental logic, and data quality enforcement. Today, they’re preview—GA would formalize refresh semantics, lineage, and production SLAs.
What it unlocks:
– Optimal refresh chooses full, incremental, or no‑op intelligently.
– Built‑in rule checks and quality reporting across your bronze → silver → gold flow.
What we’ll verify on day one:
– Failover/retry guarantees and backfill behavior. End‑to‑end lineage in the OneLake catalog. How MLV scheduling and dependencies surface in deployment pipelines.
3) Lakehouse Schemas
Why it matters – Schemas are the missing organizing principle for lakehouses—grouping tables for discovery and access control while aligning with warehouse‑style patterns. Docs and community threads still mark them preview, and even the portal has been inconsistent. GA would stabilize governance and tooling against schema‑qualified objects.
What it unlocks:
– Cleaner ACL scoping (schema‑level) and clearer object naming.
– Better compatibility with MLVs, which target schema‑enabled lakehouses.
What we’ll verify on day one:
– Parity with warehouses for GRANT/REVOKE‑like controls. Data pipeline support for writing to specific schemas. Git/deployment pipeline diffing for schema objects.
4) Power BI Calendar‑based Time Intelligence
Why it matters – This is a foundational rethink of time intelligence. “Calendars” let you define the time attributes that matter—fiscal, retail 4‑4‑5, lunar, ISO weeks—and use them directly in DAX, including new week‑aware functions. It’s preview today; GA would standardize modeling patterns that teams now patch together with custom code.
What it unlocks:
– Multiple calendars per model; sparse dates allowed.
– New functions (e.g., TOTALWTD) and cleaner semantics for period calculations.
What we’ll verify on day one:
– Migration guidance from “Mark as date table” patterns. Web modeling/TMDL parity for defining calendars in source control. Performance deltas vs. classic time‑intelligence functions.
5) OneLake Security
Why it matters – Unified, engine‑independent data security in OneLake—down to row and column level—has been the most requested governance feature of Fabric. It’s public preview now, with the official roadmap pointing to Q4 2025 GA. Ignite is the perfect stage to lock the date and the scope.
What it unlocks:
– Define roles once; enforce across Spark, SQL endpoints, and Direct Lake.
– Row‑level and column‑level security anchored to the lake, not just the semantic layer.
What we’ll verify on day one:
– Interop with Direct Lake on SQL vs. on OneLake flavors. Behavior across shortcuts and mirrored datasets. Admin APIs for role management and policy automation.
Pulling it together
Tell them what we told them. At Ignite, the most meaningful wins won’t be flashy demos—they’ll be the GA badges that reduce your risk: Cosmos DB Mirroring to eliminate brittle ETL, Materialized Lake Views to standardize medallion transformations, Lakehouse Schemas for cleaner governance, Calendar‑based Time Intelligence to end time calcs duct tape, and OneLake Security to make access control portable across engines. If these five land, Fabric’s #lakehouse story becomes simpler, sturdier, and more production‑ready for 2026.
EduDataSci - Educating the world about data and leadership 