Loading Now

Azure Storage innovations: Unlocking the future of data

Whether you’re enhancing the resilience of crucial workloads or upgrading older systems, Azure Storage has the right solution for you.

Microsoft is transforming what’s achievable in the public cloud and steering the next wave of AI-driven change for organisations. Whether you’re innovating with AI, boosting the resilience of mission-critical workloads, or modernising legacy systems with cloud-native alternatives, Azure Storage has something to offer.

At Microsoft Ignite 2025 and KubeCon North America last month, we highlighted the newest advancements in Azure Storage that power your workloads. Here’s a summary of these new releases and developments.

Innovating for the Future with AI

Azure Blob Storage provides a unified storage foundation for the entire AI lifecycle, from data ingestion and preparation to model deployment and checkpoint management.

To help customers quickly train, refine, and deploy AI models, we updated the Azure Blob Storage architecture to support exabytes of capacity, tens of Tbps of throughput, and millions of IOPS for GPUs. In this video, you can witness a single storage account scaling to over 50 Tbps in read throughput. Azure Blob Storage also forms the backbone that enables OpenAI to train and deliver models at an exceptional speed and scale.

Diagram of AI training and fine-tuning stages.
Fig 1. Storage-centric view of AI training and fine-tuning

For customers managing terabyte- or petabyte-scale AI training data, Azure Managed Lustre (AMLFS) offers a high-performance parallel file system that delivers outstanding throughput and parallel I/O to keep GPUs consistently fed with data. AMLFS 20 (preview) supports 25 PiB namespaces and up to 512 GBps throughput. Integration with Hierarchical Storage Management (HSM) improves AMLFS scalability, allowing smooth data movement between AMLFS and your vast datasets in Azure Blob Storage. The auto-import feature (preview) enables you to bring only the necessary datasets into AMLFS, while auto-export sends trained models to long-term storage or for inference.

Rakuten is speeding up the training of large Japanese language models on Microsoft Azure, utilising Azure Managed Lustre, Azure Blob Storage, and Azure Kubernetes Service to enhance GPU usage and simplify scaling.

Natalie Mao, VP, AI & Data Division, Rakuten Group

After you train and refine models, the focus shifts to inference, providing immediate predictions and insights. Azure Blob Storage serves as top-tier storage for Microsoft AI services, including Microsoft Foundry Agent Knowledge (preview) and AI Search retrieval agents (preview). This lets customers use their own storage accounts for full flexibility and control, keeping enterprise data secure and ready for retrieval-augmented generation (RAG).

Additionally, Premium Blob Storage offers consistent low latency and up to three times faster retrieval speeds, essential for RAG agents. For those who prefer open-source AI frameworks, Azure Storage has introduced the LangChain Azure Blob Loader, which provides tailored security, memory-efficient loading of millions of objects, and up to five times faster performance than earlier community implementations.

Diagram of AI inference with enterprise data.
Fig 2. Storage-centric view of AI inference with enterprise data

Azure Storage is evolving into an intelligent, integrated AI-enhanced platform, simplifying the management of exabyte-scale AI data. Storage Discovery and Copilot collaborate to help you analyse and understand the changes within your data estate over time, using user-friendly dashboards and natural language queries. With Storage Discovery and Storage Actions, you can optimise costs, safeguard your data, and manage large datasets with hundreds of billions of objects used for training and refinement.

Optimising Modern Applications with Cloud Native

Today’s modern cloud-native applications demand agility. Two key principles always stand out: elasticity and flexibility. Your storage should scale seamlessly with dynamic workloads—without added operational costs. The innovations below are built for the cloud, letting you auto-scale, optimise costs intelligently, and deliver the performance modern applications demand.

Azure Elastic SAN offers cloud-native block storage built for scaling, with tight Kubernetes integration to enable rapid scaling and multi-tenancy that reduces costs. With auto-scaling support, Elastic SAN will automatically adjust resources as needed, making it easier to manage storage across workloads. Early next year, we will make Kubernetes integration via Azure Container Storage for Azure Kubernetes Service (AKS) generally available (GA). These updates enable you to maintain familiar hosting environments while integrating cloud-native functionalities.

Agility in cloud-native systems is also crucial for modern applications based on object storage, as costs and performance must be optimised for fluctuating, unpredictable traffic patterns. Smart Tier (preview) on Azure Blob Storage continuously monitors access patterns, automatically transitioning data between tiers.

New data starts in the hot tier. After 30 days of inactivity, it shifts to the cool tier, and after 90 days, it moves to cold storage. If you access an object again, it moves back to hot, ensuring your data stays in the most cost-effective tier automatically. This approach lets you manage costs without sacrificing performance, simplifies data management at scale, and lets you focus on development.

Hosting Mission-Critical Workloads

Today’s enterprises run mission-critical workloads that rely on stable block storage and deliver predictable performance to ensure business continuity. Azure Ultra Disk is our premier block storage solution, designed specifically for applications like high-frequency trading, ecommerce platforms, transactional databases, and electronic health records systems that require extraordinary speed, reliability, and scalability.

With Azure Ultra Disk, we can confidently scale our platform globally, knowing that performance and resilience meet enterprise needs. This consistency enables our teams to focus on AI innovation and workflow automation rather than infrastructure.

Charles McDaniels, Director of Systems Engineering Management for Global Cloud Services, ServiceNow

We understand that performance, cost, and business continuity are paramount for our customers, and we’re aiming higher in all these areas:

  • Performance: We’ve further improved the average latency for Azure Ultra Disk by 30%, achieving average latency below 0.5ms for small IOs on virtual machines (VMs) with Azure Boost. A single Azure Ultra Disk can deliver leading performance of 400K IOPS and 10 GBps throughput. Additionally, the new Ebsv6 VMs allow both Premium SSD v2 and Azure Ultra Disk to achieve industry-leading performance levels of 800K IOPS and 14 GBps throughput for the most demanding applications.
  • Cost: Flexible provisioning for Azure Ultra Disk can lower total ownership costs by up to 50%, enabling independent scaling of capacity, IOPS, and MBps with more precision.
  • Business Continuity: Instant Access Snapshots (preview) provide you with the ability to back up and restore workloads immediately with superb performance in rehydration. This unique experience for Azure Premium v2 and Ultra Disk stops the usual hassles of monitoring snapshot readiness or pre-warming resources, and cuts recovery, refresh, and scale-out times from hours to seconds.

Azure NetApp Files (ANF) is engineered to offer low latency, high performance, and scalable data management. The capacity for large volumes has significantly expanded, providing over a threefold increase in single volume capacity to 7.2 PiB, and a fourfold increase in throughput to 50 GiBps. Cache volumes bring data and files closer to where users need quick access in a space-efficient format. This makes ANF ideal for various high-performance computing tasks, like Electronic Design Automation (EDA), Seismic Interpretation and Visualisation, Reservoir Simulations, and Risk Modelling. Microsoft is not just positioning ANF for mission-critical applications; we’re also using it for in-house silicon design.

Breaking Barriers—Migrating Your Storage Infrastructure

Every organisation’s cloud journey is unique. Whether you want to transition existing environments to the cloud with minimal fuss or embark on a complete transformation, Azure Storage has solutions tailored for you. The Storage Migration Solution Advisor in Copilot can offer insights to simplify your migration decisions.

Azure Data Box and Storage Mover streamline the migration process from on-premises and other cloud platforms to Azure. The next-generation Azure Data Box is now generally available. Storage Mover is our fully managed data migration solution that is secure, efficient, and scalable, now supporting on-premises NFS shares to Azure Files NFS 4.1, on-premises SMB shares to Azure Blob storage, and cloud-to-cloud transfers. The Storage Migration Solution Advisor in Copilot speeds up your migration decision-making.

For those ready to migrate their NAS facilities, Azure Files now simplifies this process more than ever. We have launched a new management model to make file shares easy to use and cost-effective. Furthermore, Azure Files allows you to eliminate complicated on-premises Active Directory or domain controller setups, with Entra-only identities for SMB shares. With this cloud-native identity support, you can now handle user permissions directly in Azure, extending to external identities for applications like Azure Virtual Desktop (AVD).

Entra-only identities support with Azure Files alters SLB’s Petrel workflows by removing on-premises domain controller dependencies, simplifying identity management and storage infrastructure for distributed teams engaged in complex exploration and reservoir characterisation. This cloud-native design allows customers to access SMB shares securely and easily, without the need for multifaceted VPN or hybrid setups.

Swapnil Daga, Storage Architect for Tenant Infrastructure, SLB

ANF Migration Assistant makes relocating ONTAP workloads from on-premises or other clouds to Azure straightforward. Behind the scenes, the Migration Assistant utilises NetApp’s SnapMirror replication technology for efficient, full-fidelity block-level incremental transfers. This lets you handle large datasets without impacting production workloads.

For customers with on-premises partner solutions that want to migrate to Azure using the same partner technology, Azure recently launched Azure Native offerings with Pure Storage and Dell PowerScale.

To facilitate migrations, Azure Storage’s Migration Program connects you with a wide range of specialists and tools. Trusted partners like Atempo, Cirata, Cirrus Data, and Komprise can expedite the migration of SAN and NAS workloads. This program aims to provide secure, low-risk transfers of files, objects, and block storage to assist enterprises in fully realising the potential of Azure.

Start Your Next Chapter with Azure Storage

The age of AI-enabled transformation is upon us. Start your journey by exploring Azure’s sophisticated storage solutions and migration tools designed to accelerate AI integration, cloud transitions, and modernisation. Take the next step today and unlock new opportunities with Azure Storage as the backbone for your AI projects.

For any inquiries, feel free to reach out at [email protected].

Share this content:


Discover more from Qureshi

Subscribe to get the latest posts sent to your email.

Discover more from Qureshi

Subscribe now to keep reading and get access to the full archive.

Continue reading