Portable Data, Competitive Execution: Iceberg + ANSI SQL Are Rewriting the Data Platform Playbook

Data teams are entering a new phase of the warehouse-to-lakehouse era: the differentiator is moving away from where data lives and toward how portable that data remains. Platform buyers are signaling fatigue with one-way migrations and proprietary dialects. Vendors are responding by turning “openness” into a first-class product narrative, and by shipping migration accelerators that reduce switching costs.

Snowflake’s recent messaging makes the point explicitly. Snowflake is promoting Spark pipeline migration into Snowflake via Snowpark Connect and the CoCo “spark-migration skill” to streamline rewrites and claim better price-performance for existing Spark workloads (Snowflake, “Migrate Apache Spark to Snowflake with CoCo”). In parallel, Snowflake is leaning into Apache Iceberg as the interoperability layer, positioning an “open, AI-ready lakehouse” with Google Cloud that reduces lock-in by standardizing table formats (Snowflake, “Apache Iceberg Lakehouse: Snowflake & Google Cloud”).

Databricks is playing the same game from the other side. Databricks is pushing conversion from proprietary warehouse code to open ANSI SQL with Genie Code, framing SQL portability as a practical migration wedge (Databricks, “Convert proprietary code to open ANSI SQL with Genie Code”). Databricks is also publishing migration narratives like NBCUniversal’s analytics modernization to validate that migrations are now repeatable programs, not bespoke projects (Databricks, “NBCUniversal’s Seamless Migration”). The pattern across both ecosystems is consistent: make the storage and semantics more standard, then compete on execution engines, governance, and operational ergonomics.

CTOs should interpret the “open table format + portable SQL” push as a structural change in vendor leverage. Iceberg (and, more broadly, open table formats) turns the data layer into an asset that can outlive any single query engine. ANSI SQL portability reduces the tax of moving BI and transformation logic. Migration tooling (CoCo, Genie Code, and similar) is becoming a competitive weapon because it compresses timelines and lowers risk, which changes procurement dynamics.

Actionable takeaways for engineering leaders:

1. Treat Iceberg (or an equivalent open table format) as a strategic boundary. Keep raw and curated datasets in an open format where possible, then let engines compete above it.
2. Invest in SQL portability as an internal platform capability. Standardize on ANSI-first patterns, add linting and CI checks for dialect drift, and isolate vendor-specific features behind well-defined interfaces.
3. Evaluate vendors on “migration operations,” not just benchmarks. Ask for automated conversion coverage, validation workflows, lineage preservation, and rollback plans. Migration speed and correctness are becoming product features.
4. Align the operating model with portability. A portable data layer still needs clear ownership for schema governance, cost controls, and cross-engine observability, or portability will degrade into fragmentation.

Data portability is turning into the new baseline. The competitive frontier is shifting to execution quality, governance, and the ability to move safely, quickly, and repeatedly.

Sources

1. https://www.snowflake.com/en/blog/migrate-spark-to-snowflake/
2. https://www.snowflake.com/en/blog/snowflake-google-cloud-open-lakehouse/
3. https://www.databricks.com/blog/convert-proprietary-code-open-ansi-sql-genie-code
4. https://www.databricks.com/blog/nbcuniversals-seamless-migration-unlocking-scalable-analytics-databricks