Zilliz Cloud Update: Tiered Storage, Business Critical Plan, Cross-Region Backup, and Pricing Changes
Today, we’re excited to announce one of the biggest updates yet to Zilliz Cloud—featuring a fully rebuilt storage system that substantially reduces costs and a new Business Critical plan built for teams with stringent security and compliance needs. We’re also standardizing storage pricing across all major cloud providers and introducing transparent data-transfer policies to make costs easier to predict as customers expand globally.
These updates are the result of months of collaboration with customers building real-world AI and vector-search systems—from RAG pipelines to recommendation and multimodal applications. Together, they make Zilliz Cloud not just faster or more cost-efficient, but more adaptable to how modern teams actually build.
Let’s take a closer look at what’s new in this release.
New Tiered Storage, Plus Lower Prices for Compute and Storage
We’re introducing a next-generation Tiered Storage solution and upgrading all existing Extended-Capacity clusters to Tiered Storage clusters, designed to make large-scale vector workloads both faster and more economical. In this architecture, your entire dataset resides in object storage (e.g., AWS S3), while each cluster’s local SSD and memory act as a cache to accelerate queries and data access.
Under the hood, the new architecture is powered by several key innovations:
Zilliz Auto Index with multi-level quantization to minimize I/O during retrieval.
Predictive prefetching that learns query behavior and preloads likely-needed data.
Deep tuning across heterogeneous storage media—memory, SSD, and object storage—maximizes throughput and minimizes latency across all layers.
Intelligent data migration pipelines automatically manage data across three tiers: Hot (memory), Warm (SSD), and Cold (object storage). Frequently accessed data stays in memory for instant response, regularly used data remains on local SSD, and infrequently accessed data is stored cost-efficiently in object storage.
As workloads evolve, the system continuously promotes and demotes data between tiers, ensuring that performance adapts to real usage patterns. In production testing, Tiered Storage achieves cache hit rates of over 90%, meaning that most queries are served directly from the fast tiers—combining the cost efficiency of object storage with the responsiveness of in-memory search.
25% lower compute, 87% lower storage
We didn’t just rebuild the architecture—we also redefined the cost structure. With this release, Tiered Storage compute pricing is reduced by 25%, and storage costs drop from 0.04 per GB per month—an 87% reduction.
For a 10TB dataset, monthly storage costs fall from roughly \400. More importantly, the total cost of ownership now approaches that of raw S3 storage while maintaining high performance for active workloads.
Business Critical Plan: Built for the Highest Levels of Security and Compliance
Over the past year, we’ve been in close conversations with teams across healthcare, finance, legal services, and other highly regulated industries. The message has been consistent: these organizations want the convenience of managed vector search but need additional layers of compliance, data protection, and operational control that go beyond standard SaaS offerings.
While Zilliz Cloud already meets HIPAA, SOC 2 Type II, and other key industry standards, some customers operate in environments that demand even stronger safeguards. To support them, we’re introducing the Business Critical plan — our highest level of security and compliance offering within Zilliz Cloud, purpose-built for teams that cannot compromise on data protection, regulatory compliance, or uptime.
What does the Business Critical Plan include?
HIPAA-ready for healthcare and other regulated workloads.
Global Cluster support for globally distributed applications — deploy clusters across regions so data stays close to users, improving performance and meeting local data-access needs.
Multi-region replication for high availability — automatically maintain read replicas in secondary regions and enable fast failover if the primary region becomes unavailable.
One-click failover for automated disaster recovery.
Point-in-time recovery (PITR) to protect against data loss or corruption.
Together, these capabilities provide the control and resilience needed to operate securely at a global scale—without the operational overhead of managing infrastructure in-house.
If you’re unsure whether the Business Critical Plan meets your organization’s needs, please contact us to discuss your specific compliance and deployment requirements.
Cross-Region Backup for Dedicated Clusters
Zilliz Cloud now supports Cross-Region Backup for Dedicated Clusters, giving teams a stronger foundation for business continuity and disaster recovery. This capability ensures that even in the event of a complete cloud region outage, your operations remain protected and recoverable.
Key capabilities include:
Automated replication: Configure your backup policy once, and Zilliz Cloud continuously handles replication to your chosen destination region.
Geographic redundancy: Keep backup copies in a physically separate region to protect against localized failures or outages.
Rapid recovery: Restore data from a cross-region backup to a new cluster in minutes, reducing downtime and improving your Recovery Time Objective (RTO).
Cross-Region Backup adds another layer of reliability to Zilliz Cloud’s managed platform—ensuring that your most critical vector workloads remain available, no matter what happens in a single region. For more information, check the documentation.
Index Build Level Tuning: Balance Accuracy and Capacity
Zilliz Cloud introduces a new Index Build Level feature for Milvus 2.6.x clusters, giving you direct control over how your vector indexes are built and optimized. Powered by our next-generation quantization engine, this feature lets you fine-tune the trade-off between storage capacity and search accuracy (recall) to better fit your application’s needs.
When creating an index, you can choose from three levels:
Precision-first: Maximizes search accuracy for mission-critical workloads.
Capacity-first: Optimizes for data density, allowing you to store more vectors while slightly reducing recall.
Balanced (default): Provides an optimal balance between recall and capacity for general use.
For more information, check the index build level documentation.
Pricing Updates for Cross-Cloud Storage and Data Transfer
Starting January 1, 2026, storage pricing will be standardized at $0.04 per GB per month across all major cloud providers, including AWS, Azure, and Google Cloud. This change ensures consistent pricing regardless of where your deployments run, making cost planning more straightforward and more predictable for multi-cloud teams.
As more customers deploy workloads across regions and clouds, we’re also introducing internet egress, cross-cloud, and cross-region data transfer and access capabilities. Data transfer fees will be passed through at our provider’s actual costs—without any Zilliz markup. Each customer will include a set of monthly free credits to cover baseline usage. This pricing change will also take effect on January 1, 2026.
Note: We’ll share more detailed information about these pricing changes ahead of the January rollout.
For a complete list of new features and enhancements in this release, see the Zilliz Cloud release notes.
Getting Started with Zilliz Cloud
Ready to see it in action? Sign in to your Zilliz Cloud and try the new features today.
New to Zilliz Cloud? Sign up for free and get $100 in credits to experience the world’s leading managed vector database.
If your organization operates in a regulated industry or has strict compliance requirements, you can learn more about the Business Critical Plan by contacting your Zilliz account team or our support team.
For questions about any of these updates, check our documentation or contact Zilliz Support—we’re here to help you get the most out of Zilliz Cloud.
Build Without Limits: Zilliz Cloud Key Features at a Glance
With this release, Zilliz Cloud strengthens its position as the most performant, cost-efficient, and secure vector database service—while continuing to deliver the advanced AI search capabilities of Milvus in a fully managed environment:
Elastic scaling & cost efficiency – One-click deployment, serverless autoscaling, and pay-as-you-go pricing.
Advanced AI search – Vector, full-text, and hybrid (sparse + dense) search with metadata filtering, dynamic schema, and multi-tenancy.
Natural language querying – MCP server support for intuitive queries without complex APIs.
Enterprise-grade reliability & security – 99.95% SLA, SOC 2 Type II and ISO 27001 certifications, GDPR compliance, HIPAA readiness, RBAC, BYOC, and now audit logs. See our trust center for more information.
Global availability – Deployments across AWS, GCP, and Azure with sub-100ms latency worldwide.
Seamless migration – Built-in tools to move from Pinecone, Qdrant, Elasticsearch, PostgreSQL, OpenSearch, AWS S3 vectors, Weaviate, or on-prem Milvus.
Taken together, these features make Zilliz Cloud more than just a vector database. It’s a production-ready platform that enables enterprises to build and scale AI applications without limitations.
Fendy FengFendy Feng is the Product Marketing Manager at Zilliz. She has extensive experience developing and enhancing the impact of open-source projects in various global markets by producing high-quality, tailored content. Before joining Zilliz, Fendy worked as a Content Strategist at PingCAP, a fast-growing E-Series startup renowned for its open-source distributed SQL database.
- New Tiered Storage, Plus Lower Prices for Compute and Storage
- Business Critical Plan: Built for the Highest Levels of Security and Compliance
- Cross-Region Backup for Dedicated Clusters
- Index Build Level Tuning: Balance Accuracy and Capacity
- Pricing Updates for Cross-Cloud Storage and Data Transfer
- Getting Started with Zilliz Cloud
- Build Without Limits: Zilliz Cloud Key Features at a Glance
Content
Start Free, Scale Easily
Try the fully-managed vector database built for your GenAI applications.
Try Zilliz Cloud for FreeKeep Reading
The Great AI Agent Protocol Race: Function Calling vs. MCP vs. A2A
Compare Function Calling, MCP, and A2A protocols for AI agents. Learn which standard best fits your development needs and future-proof your applications.
Cosmos World Foundation Model Platform for Physical AI
NVIDIA’s Cosmos platform pioneers GenAI for physical applications by enabling safe digital twin training to overcome data and safety challenges in physical AI modeling.
Producing Structured Outputs from LLMs with Constrained Sampling
Discuss the role of semantic search in processing unstructured data, how finite state machines enable reliable generation, and practical implementations using modern tools for structured outputs from LLMs.