TLDR: MinIO is significantly upgrading its object storage capabilities by integrating Apache Iceberg tables, creating a crucial linking layer between its high-performance storage and Generative AI models. This move aims to provide structured access to massive unstructured datasets, enabling AI agents to directly interact with data for administration, discovery, and processing, thereby accelerating AI/ML workloads and driving company growth, particularly in private cloud environments.
MinIO, a leader in high-performance object storage, is making a strategic leap into the Generative AI (GenAI) landscape by incorporating Apache Iceberg tables, establishing a vital ‘linking layer’ between its core object store and advanced AI models. This development, highlighted in a recent interview with MinIO co-founder and co-CEO AB Periasamy and CMO Erik Frieberg, underscores the company’s commitment to evolving its offerings to meet the escalating demands of AI workloads.
Periasamy articulated MinIO’s long-standing vision, stating, ‘We always said MinIO object store is a key value store. What is a key value store? It’s a data store. We are closer to Oracle than EMC in the past.’ He emphasized that while object storage traditionally functions as a simple key-value store, it must handle data at ‘massive scale without losing a single object,’ while being ‘transaction ACID-compliant.’ He added, ‘So we always thought of ourselves as a database company, except that we are talking about unstructured data. Things are starting to evolve there because of GenAI.’ This evolution sees MinIO embracing structure through Iceberg tables to better serve the needs of GenAI.
The integration with Iceberg is designed to facilitate a more structured interaction between MinIO’s object store and vector-focused GenAI models and agents. Periasamy noted, ‘The AI is directly talking to the object store through the MCP server. It’s the agents that are interacting with the object store, both in terms of administration as well as the data discovery and dealing with the data itself.’ This direct interaction is crucial for the efficiency and scalability required by modern AI applications.
This strategic pivot is not new for MinIO, which previously launched AIStor, an ‘AI-centered object store,’ in November 2024. At that time, CMO Jonathan Symonds highlighted the dramatic shift in data workloads, stating, ‘We have multiple clients that are over exabyte in terms of data stored on MinIO, and the types of workloads that they’re running against that is totally different than in the past.’ He explained that organizations are now collecting ‘massive amounts of unstructured data’—including video, log files, and telemetry—specifically for building and training AI models. AIStor introduced AI-specific capabilities such as a new S3-compatible API, promptObject, allowing users to ‘talk’ to unstructured data, and a private repository for AI models.
MinIO’s object storage is positioned as a robust solution for various AI/ML use cases. It serves as a high-performance backend for Retrieval Augmented Generation (RAG) setups, storing custom corpuses for domain-specific Large Language Model (LLM) responses. Its compact footprint (less than 100 MB binary) and deployability on any hardware make it ideal for AI at the edge, with features like Bucket Notifications and Object Lambda enabling instant inference on new data. Furthermore, MinIO provides critical data protection features, including erasure coding, site replication, encryption, and Identity and Access Management (IAM), ensuring data redundancy, security, and controlled access for AI workloads.
Also Read:
- DigitalOcean Unveils GradientAI Platform for Streamlined Generative AI Development
- Agentic AI Frameworks Drive Transformative Shift in Enterprise Strategy and Cloud Adoption
This focus on AI is a significant growth driver for MinIO. Periasamy emphasized, ‘This is directly contributing to our company’s growth. Because once you start scaling your data and AI data infrastructure, it points to object store and customers are also looking beyond cloud.’ He concluded that the scale of AI data is pushing customers towards ‘good private cloud and the private cloud object store,’ a market MinIO has actively cultivated outside of major cloud providers, contributing significantly to its aggressive growth in engineering, marketing, and sales.
