Google’s cloud storage gets faster and smarter for AI

Google announced AI-focussed data delivery speed increases and better metadata generation for its Google cloud storage at its Next ’26 conference event.

It’s providing higher-performance object and Lustre storage infrastructure, automated metadata annotation, and AI agent connectivity via MCP, streamlined data management through zero-configuration dashboards, aggregated activity views, and enhanced batch operations. And there are expanded capabilities across Google Cloud NetApp Volumes, Filestore for GKE, and its backup and data protection portfolio.

Sameet Agarwal, VP/GM, Storage, Google Cloud and Asad Khan, Senior Director of Product Management, Google Cloud, say in a blog-like article: “We are announcing innovations across every layer of our storage stacks — performance, intelligence, and management — to ensure your data is as fast and as useful as the AI models, apps and agents you are building.”

Google is announcing:

• Cloud Storage Rapid with Rapid Cache and Rapid Buicket components
• Google Cloud Managed Lustre and new Dynamic Tier
• Smart Storage
• Storage Intelligence improvements

Cloud Storage Rapid

The Cloud Storage Rapid service provides “extreme throughput, frequent I/Os, and ultra-low latency,” using object storage with “industry-leading durability, massive distributed scale, and cost-effective auto-tiering.” It integrates with PyTorch and JAX, to support the “the most popular AI/ML ecosystem frameworks” and there are two variants;

Rapid Bucket which uses Google’s Colossus distributed storage system “to deliver more than 15 TB/s of bandwidth, 20 million requests per second, and sub-millisecond latency in a single zonal bucket.” It has access “via high-performance gRPC and S3-compatible APIs.” Compared to traditional object storage, checkpoint restores are 5x faster and checkpoint writes are 3.2x faster. Also it “increases accelerator utilization for multi-modal training with 50% reduced GPU blocked time and 2.5x faster data loading” compared to regional storage. 

Google notes “Regional storage will give you the best price for performance possible, as long as you’re processing and serving your data in the same region. However, there are tradeoffs to regional storage in the areas of availability and resiliency.”

Rapid Cache, the renamed Anywhere Cache, accelerates bandwidth for bursty workloads with an aggregate read throughput of 2.5 TB/s for existing buckets, with no code changes. A new ingest-on-write feature provides up to 2.2x faster checkpoint restores.When you configure a cache to ingest data on write, data is written to the cache as it gets uploaded to the bucket and allows your workloads to benefit from an immediate cache hit on the very first read.

Google Cloud Managed Lustre

This is its parallel filesystem offering, built on DDN’s Lustre and EXAScaler, and is claimed to be “the most performant managed Lustre offering available in any cloud. … [It] delivers up to 10 TBps of throughput — a 10x increase since last year and 4–20x higher than managed Lustre offerings from other hyperscalers for a single instance.”

It’s based on new C4NX VMs and Hyperdisk Exapools, and Managed Lustre writes and restores checkpoints 2.6x faster when compared to other Google Cloud storage offerings.

There is a new Dynamic Tier, priced at $0.06/GB-month, delivering low-latency performance by “serving data from persistent disk rather than relying on object-based caching. … [and eliminating] a performance cliff…. A single SKU provides simple, predictable billing without the hidden complexity of traditional data tiering.”

Managed Lustre can be used as a shared KV-cache for AI inference, improving performance and economics. It reduced the mean time to first token by more than 40 percent compared to using KV Cache in host memory alone. DDN says it improves total inference throughput by 75 percent, citing a Google Cloud blog. The bloggers say: "We believe that Google Cloud Managed Lustre should be your primary storage solution for external KV Cache."

Smart Storage

New Smart Storage capabilities enable ML teams to select training datasets from semantic criteria without building retrieval pipelines. They provide:

• Automated annotations, including image annotations, as content is ingested. “You pay to annotate the data once at write time, and every downstream system can use those annotations immediately for the life of the object.”
• A Cloud Storage MCP server lets you read, write, and analyze Cloud Storage data using the standard MCP protocol.

Storage Intelligence

Google’s existing Storage Intelligence services are being enhanced with:

• Zero-configuration dashboards to instantly surface cost anomalies and integrate Security Command Center’s Data Security Posture Management (DSPM) data governance feature, to detect critical security vulnerabilities across Cloud Storage; no setup required. 
• New object events and bucket activity tables in Insights Datasets now drive deeper cost analysis and accelerate operational tasks. 
• Enhanced batch operations make it simpler to act on billions of objects with new change ACL and storage class operations, and support for multi-bucket operations.

There are three more enhancements;

• Google Cloud NetApp Volumes: With the launch of Flex Unified, NetApp Volumes provides a unified enterprise storage platform that bridges the data center and the cloud, provisioning both block (iSCSI, NVMe/TCP) and file (NFS/SMB) on the same storage pool. New ONTAP-mode lets you bring your automation (Terraform, Ansible) and ONTAP APIs directly to NetApp Volumes.
• Filestore for GKE: Developers building AI workloads on Google Kubernetes Engine (GKE) can start small, with shares as small as 100 GiB, and scale capacity and IOPS independently. Tighter integration to the Colossus distributed file system provides more scale and enterprise capabilities.
• Google Cloud Backup and DR now features agentic AI capabilities that can autonomously audit your backup estate and remediate coverage gaps, with new GA integrations for AlloyDB and Filestore.

NetApp is deepening its partnership with Google Cloud by adopting Gemini Enterprise internally to power AI-driven operations across sales and product development.

To learn more, visit the Google Cloud Storage console to find out about the new features, read more about Cloud Storage Rapid, or explore the Next '26 storage sessions.