Google Data Cloud’s roundup of recent announcements spans several months, covering new features, previews, and events. Stateful processing for BigQuery continuous queries enters preview, enabling JOINs, aggregations, and windowing functions directly in streaming queries to compute metrics like 30-minute averages. The synthetic data generator tool for Managed Service for Kafka is now generally available, letting users send mock data to clusters in three clicks. Dataflow pipeline updates introduce a stop-and-replace option (parallel pipeline) to reduce disruption during migration, plus a timeout on drains to control costs; both are generally available.

In the Lakehouse area, managed tables for Apache Iceberg enter preview, providing Google-managed tables with native multi-engine read/write interoperability across BigQuery and open-source engines, along with automated compaction and partition tuning. Google Cloud’s NoSQL Database portfolio—Bigtable, Firestore, Memorystore—is featured in a new series, with events like the ‘Beyond the Query’ webinar for AI agents and hands-on Cloud Engineer’s AI Toolkit Workshops. A 10-day free trial for Bigtable (1 node SSD cluster, up to 500GB) is offered, plus $300 in free credits for new customers.

Managed Service for Apache Airflow now includes general availability of Airflow 3.1, AI-powered agentic troubleshooting, a managed Airflow MCP Server for custom agent integration, and declarative YAML-based orchestration pipelines. New Google-built ODBC and JDBC drivers for BigQuery are in preview, providing high-performance, open-source connections. Google also reintroduced Data Studio for AI-era use cases, announced BigQuery Graph preview for scalable graph analytics, and expanded Conversational Analytics to Looker Embedded environments, enabling natural language queries in custom apps. Looker adds self-service Explores for ad-hoc analysis with governed data.

Cloud SQL introduces autoscaling read pools (multiple read replicas via a single endpoint). The Gemini assistant in BigQuery Studio evolves into a fully context-aware analytics partner. Managed and remote MCP support for AlloyDB, Spanner, Cloud SQL, Bigtable, and Firestore is announced to power AI agents. Firestore’s Enterprise edition gains pipeline operations with over a hundred new query features, index-less queries, and new index types. Cloud SQL for SQL Server integrates with Microsoft Entra ID for centralized identity management in multi-cloud environments, now in public preview. Cloud Composer adds Gemini Cloud Assist investigations for automated root cause analysis of failed Airflow tasks.