AIStor Tables
AIStor Tables is an enterprise-grade data store from MinIO that embeds the full Apache Iceberg V3 REST Catalog API directly into object storage, unifying structured and unstructured data for AI and analytics workloads.
Publisher review
AIStor Tables is an enterprise-grade data store from MinIO that embeds the full Apache Iceberg V3 REST Catalog API directly into object storage, unifying structured and unstructured data for AI and analytics workloads. Announced generally available on February 3, 2026, it targets organizations running on-premises, private, sovereign, or hybrid environments that need to eliminate silos between tabular data (tables, transactions) and unstructured data (images, audio, logs) without relying on external catalog services. The product is designed for data engineers, AI/ML teams, and lakehouse architects who require a single, high-performance store for both analytics and agentic AI workloads, with support for multiple query engines including Spark, Trino, Dremio, and Starburst.
AIStor Tables works by integrating the Iceberg V3 REST Catalog API as a native first-class feature within MinIO's object store, removing the need for a separate catalog database or service. Key capabilities include multi-table transactions with atomic guarantees, version control for roll-back and time-series operations, and Iceberg Views—virtualized tables defined by SQL queries that are portable across engines like Spark, Trino, and Hive. The store supports multiple protocols (S3, S3 Express, MCP, SFTP, and Iceberg Catalog) and claims to reduce storage costs by up to 40 percent compared to hyperscaler-controlled table services, as per MinIO's published benchmarks.
MinIO positions AIStor Tables as the first object-native on-premises store with a built-in Iceberg catalog, directly competing with AWS S3 Tables (a managed cloud service) and traditional on-prem solutions that require separate catalog databases (e.g., Hive Metastore, AWS Glue). The product has been validated by early adopters like Nomura, Japan's largest investment bank. However, MinIO's licensing model—reportedly requiring a minimum of 400 TB—has drawn criticism from community users who find pricing prohibitive for smaller deployments, with some Reddit posts describing the company as having "mental" prices and a "terrible" reputation for past UI changes.
The honest trade-offs: AIStor Tables eliminates external catalog complexity but introduces vendor lock-in to MinIO's ecosystem, especially given that the Iceberg catalog is embedded rather than pluggable. The 400 TB minimum license makes it impractical for small-to-medium businesses, and community backlash over MinIO's licensing changes (e.g., stripping the free UI to an object browser) raises trust concerns. Additionally, while the product promises cost savings over hyperscalers, actual total cost depends on existing infrastructure and data gravity, and the lack of a free tier or transparent pricing tiers may deter evaluation. For teams already invested in open-source Iceberg catalogs (e.g., Nessie, Polaris), the proprietary integration may not justify the migration effort.
How it works
-
Native Iceberg V3 Catalog API
Embedded the full Apache Iceberg V3 REST Catalog API directly into the data store, eliminating the need for a separate catalog service.
-
Iceberg Views support
Supports virtualized tables defined by SQL queries that are portable across Spark, Trino, Hive, and other engines without duplicating storage.
-
Multi-table transactions
Provides atomic, multi-table transactions with version control for roll-back, time-series, and operational consistency at the storage layer.
-
Unified structured/unstructured storage
Combines tables, transactions, images, audio, and logs under a single Iceberg-based fabric for AI and analytics workloads.
-
Multi-protocol support
Supports S3, S3 Express, MCP, SFTP, and Iceberg Catalog protocols for flexible data access across tools and pipelines.
-
Cost reduction up to 40%
Claims to reduce storage costs by up to 40 percent compared to hyperscaler-controlled table services, per MinIO benchmarks.
-
On-premises and hybrid deployment
Available across on-premises, private, sovereign, and hybrid environments, with no dependency on public cloud catalog services.
Strengths and trade-offs
Strengths
- First in the industry to embed the full Apache Iceberg V3 REST Catalog API directly into the data store, removing external catalog overhead.
- Unifies structured and unstructured data under a single Iceberg fabric, enabling AI to access all enterprise knowledge without silos.
- Supports multi-table transactions with atomic guarantees and version control, ensuring data correctness for operational and analytical workloads.
- Reduces storage costs by up to 40 percent compared to hyperscaler-controlled table services, as claimed by MinIO benchmarks.
Trade-offs
- Requires a minimum license of 400 TB, making it prohibitively expensive for small-to-medium businesses and evaluation deployments.
- Vendor lock-in to MinIO's proprietary Iceberg catalog integration, limiting flexibility to switch to open-source catalog alternatives like Nessie or Polaris.
- Community backlash over past licensing changes (e.g., stripping free UI to object browser) raises trust and long-term commitment concerns.
- No free tier or transparent pricing tiers publicly available, hindering hands-on evaluation before significant financial commitment.
Pricing context
Pricing is not explicitly disclosed in public sources; community reports indicate a minimum license of 400 TB, with costs described as 'mental' and 'insane' by some users, suggesting enterprise-level pricing with no free tier.
Getting started with AIStor Tables
-
Sign up for AIStor Tables
Contact MinIO sales to initiate the licensing process, as AIStor Tables requires a minimum 400 TB license with no free tier. Provide your organization details and deployment environment (on-premises, private, or hybrid) to receive access credentials and installation instructions.
-
Connect your object storage
Deploy AIStor Tables on your infrastructure using the provided installer. Configure the connection to your existing MinIO object store or set up a new one, ensuring S3-compatible endpoints are accessible. Verify network connectivity and storage capacity meets the 400 TB minimum.
-
Configure the Iceberg catalog
Enable the embedded Apache Iceberg V3 REST Catalog API within AIStor Tables. Set catalog properties such as warehouse location and namespace defaults. No external catalog service is needed; the catalog runs natively inside the object store.
-
Load structured and unstructured data
Use the Iceberg REST Catalog endpoint to register tables and load structured data (e.g., CSV, Parquet) via Spark, Trino, or Dremio. Simultaneously, store unstructured data like images or logs in the same bucket using S3 protocols, unifying both types under the Iceberg fabric.
-
Run a multi-table transaction
Execute an atomic multi-table transaction using your query engine (e.g., Spark SQL) against the Iceberg tables. Perform updates, inserts, or deletes across multiple tables, then verify version control by rolling back to a previous state using the catalog's time-series capabilities.
Frequently Asked Questions
What is AIStor Tables from MinIO?
AIStor Tables is an enterprise-grade data store from MinIO that embeds the full Apache Iceberg V3 REST Catalog API directly into object storage. It unifies structured and unstructured data for AI and analytics workloads, eliminating the need for a separate catalog service.
How does AIStor Tables integrate Apache Iceberg?
AIStor Tables integrates the Iceberg V3 REST Catalog API as a native first-class feature within MinIO's object store. This removes the need for a separate catalog database or service, enabling direct support for query engines like Spark, Trino, Dremio, and Starburst.
What are the key features of AIStor Tables?
Key features include native Iceberg V3 Catalog API, Iceberg Views for portable SQL-defined tables, multi-table transactions with atomic guarantees and version control, unified storage for structured and unstructured data, and multi-protocol support including S3, S3 Express, MCP, SFTP, and Iceberg Catalog.
How does AIStor Tables pricing work?
AIStor Tables pricing is not publicly disclosed, but community reports indicate a minimum license of 400 TB. Users have described costs as 'mental' and 'insane', suggesting enterprise-level pricing with no free tier, making it impractical for small-to-medium businesses.
What are the trade-offs of using AIStor Tables?
AIStor Tables eliminates external catalog complexity but introduces vendor lock-in to MinIO's ecosystem due to the embedded Iceberg catalog. The 400 TB minimum license is prohibitive for smaller deployments, and past licensing changes have raised community trust concerns.
How does AIStor Tables compare to AWS S3 Tables?
AIStor Tables is an on-premises store with a built-in Iceberg catalog, competing with AWS S3 Tables, a managed cloud service. MinIO claims up to 40% cost savings over hyperscaler table services, but AIStor Tables requires a minimum 400 TB license and lacks a free tier.
Alternatives
How AIStor Tables compares
Direct head-to-head against 2 competitors. Picked by 7wData.
AIStor Tables
- Pricing
- Pricing is not explicitly disclosed in public sources; community reports indicate a minimum license of 400 TB, with costs described as 'mental' and 'insane' by some users, suggesting enterprise-level pricing with no free tier.
- Target
- AIStor Tables is an enterprise-grade data store from MinIO that embeds the full Apache Iceberg V3 REST Catalog API directly into object storage, unifying structured
- Strength
- First in the industry to embed the full Apache Iceberg V3 REST Catalog API directly into the data store, removing external catalog overhead.
- Watch for
- Requires a minimum license of 400 TB, making it prohibitively expensive for small-to-medium businesses and evaluation deployments.
AWS S3 Tables
- Pricing
- Storage, requests, object monitoring fees
- Target
- Enterprise-scale analytics and AI
- Deployment
- Cloud
- Strength
- Deep AWS ecosystem integration
- Watch for
- Costs escalate with scale
Databricks Delta Lake
- Pricing
- Custom/Contact sales
- Target
- Unified analytics and AI workloads
- Deployment
- Cloud, hybrid
- Strength
- Native Delta Lake format integration
- Watch for
- Complex pricing structure
User reviews
No user reviews yet. Be the first to write one.
Sources
Reporting on this tool draws on these publicly available sources.