Google Cloud Billing added spend cap budgets in preview

Eligible services can now be tied to budgets that automatically pause usage when estimated costs go above a configured cap.

That’s a strong guardrail if you’re trying to stop surprise spend spikes before they turn into real charges. It gives teams a more direct control point for managing unplanned usage.

Cloud Billing also introduced early anomaly signals for AI workloads

These near-real-time cost estimates are designed to flag AI spend surges before finalized invoices arrive.

This is especially useful for higher-variance services like Gemini and Vertex AI. Getting early signals means you can react sooner and keep AI costs from drifting unexpectedly.

Compute Engine increased Hyperdisk Balanced throughput for C4D machine types

Larger attached volumes can now reach up to 12,500 MiB/s of throughput.

That gives compute-optimized workloads more storage performance headroom. It can also help you match storage provisioning more closely to actual workload needs instead of overbuying just to keep up.