Google launches 2nd generation of its Cloud SQL service, with up to 10TB per instance
Context & Ripple Effects
This launch caps a year in which Google rebuilt its database stack for enterprise buyers: after opening its fast NoSQL store to customers with the Cloud Bigtable launch in May, it is now raising the ceiling on its managed relational offering to 10TB per instance — the size class where mid-market production databases actually live.
Why it matters: scale limits were the practical reason enterprises kept Cloud SQL out of serious deployments. The move positions Google to hold those workloads until the portfolio matures, which the related coverage confirms happened with the general-availability release of Cloud SQL, Bigtable, and Datastore the following summer.
First-order effects
- Enterprise customers running MySQL-class workloads can now keep larger single instances on Cloud SQL instead of sharding or leaving Google Cloud Platform, removing the most common disqualifier cited against the service.
Second-order effects
- A scaled-up managed SQL tier forces the rest of Google's database lineup to compete as a coherent suite rather than point tools — the path that led Google to commit to supporting Microsoft's SQL Server on Cloud SQL three years later.
Third-order effects
- If the pattern holds — bigger instances, more engines, then general availability — managed databases stop being a side bet and become core GCP infrastructure, culminating in the modern portfolio where even Spanner carries graph and vector search for AI workloads.
The trend: Google Cloud is scaling its managed database lineup from experimental services toward an enterprise-grade, AI-ready portfolio built over a decade.