This week’s data analytics news is dominated by a data-collaboration deal that finally closed, plus a cluster of governance and AI features landing on the big lakehouse platforms. Publicis Groupe completed its roughly $2.2 billion purchase of LiveRamp on November 2, while Snowflake and Databricks both pushed out updates that matter to anyone running shared data, Iceberg tables or secrets-heavy pipelines. Note that the Snowflake items below span October 29 to November 2, because the last 24 hours were light on major primary-source announcements.
Data Analytics News: Publicis Closes $2.2B LiveRamp Acquisition
On November 2, Publicis Groupe announced it had completed its acquisition of LiveRamp, the data collaboration and identity platform. The deal was first reported in May as an all-cash transaction valued at $2.2 billion. Alongside the close, Publicis launched what it calls a plug-and-play AI growth engine that combines LiveRamp’s data collaboration platform with Epsilon’s identity technology.
The pitch to clients is that data stays in their own environment, they own the AI assets they help create, and they avoid platform lock-in. According to PPC Land’s coverage, an FTC consent order accepted on October 19 requires LiveRamp to stay commercially neutral toward rival agencies’ clients. For anyone following data analytics news, the takeaway is that first-party data pipelines and clean-room style collaboration are now strategic assets worth billions.
Snowflake Takes Open Data Sharing to General Availability
Snowflake’s release notes for November 2 show that Open Data Sharing is now generally available and no longer a preview feature. The same day, Snowflake improved its AI_EXTRACT function, which now handles complex documents including forms, handwriting, dense tables and charts with better accuracy.
If your team shares datasets across organizations, GA status removes the main objection to building on it in production. Better document extraction also means fewer custom parsing scripts sitting between a PDF and a table. If you are still sharpening the querying side, our walkthrough of Snowflake’s recent revenue results gives useful context on where the platform is heading.
Snowflake Previews Cortex Training for Open-Weight LLMs
On October 30, Snowflake listed Cortex Training in Preview, a serverless service for customizing open-weight large language models from inside a workspace without managing GPU infrastructure. The same release batch added declarative secret and environment variable management in DCM Projects, and UUID support as a key type in hybrid tables.
The practical angle is that fine-tuning moves closer to the data warehouse, which reduces the data movement and access-control headaches that normally come with training on governed data. It is still a preview, so treat it as something to test rather than rely on.
Iceberg Governance Gets Tri-Secret Secure and Horizon Catalog Access
Two Iceberg items reached general availability in Snowflake’s notes. On October 30, Tri-Secret Secure for Iceberg tables became GA, letting customers use their own managed keys to encrypt Snowflake storage for Iceberg tables on AWS and Azure. On October 29, access to externally managed Iceberg tables through Horizon Catalog’s REST interface also went GA, alongside telemetry that records catalog API requests from outside engines into an event table.
For data engineers running multi-engine lakehouses, customer-managed encryption plus request logging addresses two of the most common audit questions: who holds the keys, and who touched the data.
Databricks Raises Unity Catalog Secret Limits
Databricks’ November 2 release notes include higher limits and BROWSE support for Unity Catalog secrets. Schemas can now hold up to 1,000 secrets and metastores up to 10,000, and the BROWSE privilege on a catalog lets users discover secrets and their metadata without being able to read the values. The same page lists a Beta managed warehouse for MLflow, so traces stored in Unity Catalog can be viewed in the MLflow UI without selecting a SQL warehouse.
These are incremental changes in the day’s data analytics news, but they matter for teams with many pipelines and credentials: bigger limits mean fewer workarounds, and separating discovery from access is a cleaner permission model.
What This Means for Analysts
The common thread in today’s data analytics news is control: who owns the data, who holds the keys, and who can see that a secret exists. Before your next planning cycle, check whether any of these features removes a workaround in your own stack, and read the primary notes linked below rather than relying on summaries. For hands-on practice with the querying side, see our recent piece on Polars out-of-core group-by.