Apache Iceberg Will Become the Default Multi-Cloud Data Interoperability Standard, Ending Proprietary Lakehouse Lock-In by 2028
The Claim Apache Iceberg's adoption as the open standard for cross-cloud lakehouse interoperability will effectively end proprietary data format lock-in, enabling enterprises to run analytics across AWS S3, Azure, Databricks, Snowflake, and BigQuery without data migration or custom pipelines within three years.
The Convergence Signal The most strategically underrated announcement at Google Cloud Next '26 was not a new model or a chip — it was a data format commitment. Google's Cross-Cloud Lakehouse uses Apache Iceberg as its interoperability standard, and the integration list reads like a who's who of enterprise data platforms: AWS S3, Databricks Unity Catalog, Snowflake Polaris, and BigQuery — four platforms that compete fiercely with each other — are all connecting via the same Iceberg REST Catalog standard.
This matters because format lock-in has been the primary instrument of cloud vendor lock-in for the past decade. If you built your lakehouse on Delta Lake, you stayed on Databricks. If you built on Parquet with AWS Glue, migration was an expensive multi-year project. Iceberg changes the calculus: it's table format plus a catalog API, and when all major platforms implement it natively, data becomes portable without extraction and re-loading.
What Google Is Offering and What It Costs Cross-Cloud Interconnect delivers sub-second latency for cross-cloud data queries and — critically — Google is offering zero egress fees for this access. Egress fees have been the silent enforcer of cloud lock-in: moving petabytes of data between clouds is economically prohibitive even when it's technically feasible. Removing that cost fundamentally changes the architecture conversation.
BigQuery's performance improvements compound the value: 35% faster query speeds and 40% cost reduction year-over-year means that running analytics on cross-cloud data via BigQuery is not just technically viable but economically attractive.
The Realpolitik None of the competing platforms confirmed full Iceberg parity from their side in this corpus. Google's announcement is Google's position — Databricks, Snowflake, and AWS all have their own competitive interests that may lead to partial implementation or proprietary extensions that erode the standard's practical interoperability over time. This is the standard fate of open standards in competitive markets.
But the signal is still meaningful: when competitors with mutual interest in lock-in converge on a single format standard simultaneously, the standard has typically won. The pattern mirrors what happened with container orchestration (Kubernetes), columnar storage (Parquet), and REST APIs.