2 months ago
60
2025 was a pivotal twelvemonth successful Azure Storage, and we’re heading into 2026 with a wide absorption connected helping customers crook AI into existent impact.
2025 was a pivotal twelvemonth successful Azure Storage, and we’re heading into 2026 with a wide absorption connected helping customers crook AI into existent impact. As outlined successful past December’s Azure Storage innovations: Unlocking the aboriginal of data, Azure Storage is evolving arsenic a unified intelligent level that supports the afloat AI lifecycle astatine endeavor standard with the show modern workloads demand.
Looking up to 2026, our investments span the afloat breadth of that lifecycle arsenic AI becomes foundational crossed each industry. We are advancing retention show for frontier exemplary training, delivering purpose‑built solutions for large‑scale AI inferencing and emerging agentic applications, and empowering cloud‑native applications to run astatine agentic scale. In parallel, we are simplifying adoption for mission‑critical workloads, lowering TCO, and deepening partnerships to co‑engineer AI‑optimized solutions with our customers.
We’re grateful to our customers and partners for their spot and collaboration, and excited to signifier the adjacent section of Azure Storage unneurotic successful the twelvemonth ahead.
Extending from grooming to inference
AI workloads extend from large, centralized exemplary grooming to inference astatine scale, wherever models are applied continuously crossed products, workflows, and real-world decision making. LLM grooming continues to tally connected Azure, and we’re investing to enactment up by expanding scale, improving throughput, and optimizing how exemplary files, checkpoints, and grooming datasets travel done storage.
Innovations that helped OpenAI to run astatine unprecedented standard are present disposable for each enterprises. Blob scaled accounts let retention to standard crossed hundreds of standard units wrong a region, handling millions of objects required to alteration endeavor information to beryllium utilized arsenic grooming and tuning datasets for applied AI. Our partnership with NVIDIA DGX connected Azure shows that standard translates into real-world inference. DGX unreality was co-engineered to tally connected Azure, pairing accelerated compute with high-performance storage, Azure Managed Lustre (AMLFS), to enactment LLM research, automotive, and robotics applications. AMLFS provides the champion price-performance for keeping GPU fleets continuously fed. We precocious released Preview enactment for 25 PiB namespaces and up to 512 GBps of throughput, making AMLFS champion successful people managed Lustre deployment connected Cloud.
As we look ahead, we’re deepening integration crossed fashionable archetypal and third-party AI frameworks specified arsenic Microsoft Foundry, Ray, Anyscale, and LangChain, enabling seamless connections to Azure Storage retired of box. Our autochthonal Azure Blob Storage integration wrong Foundry enables endeavor information consolidation into Foundry IQ, making blob retention the foundational furniture for grounding endeavor knowledge, fine-tuning models, and serving low-latency discourse to inference, each nether the tenant’s information and governance controls.
From grooming through full-scale inferencing, Azure Storage supports the full cause lifecycle: from distributing ample exemplary files efficiently, storing and retrieving long-lived context, to serving information from RAG vector stores. By optimizing for each signifier end-to-end, Azure Storage has performant solutions for each signifier of AI inference.
Evolving unreality autochthonal applications for agentic scale
As inference becomes the ascendant AI workload, autonomous agents are reshaping however unreality autochthonal applications interact with data. Unlike human-driven systems with predictable query patterns, agents run continuously, issuing an bid of magnitude much queries than accepted users ever did. This surge successful concurrency stresses databases and retention layers, pushing enterprises to rethink however they designer caller unreality autochthonal applications.
Azure Storage is gathering with SaaS leaders similar ServiceNow, Databricks, and Elastic to optimize for agentic standard leveraging our artifact retention portfolio. Looking forward, Elastic SAN becomes a halfway gathering artifact for these unreality autochthonal workloads, starting with transforming Microsoft’s own database solutions. It offers afloat managed artifact retention pools for antithetic workloads to stock provisioned resources with guardrails for hosting multi-tenant data. We’re pushing the boundaries connected max standard units to alteration denser packing and capabilities for SaaS providers to negociate agentic postulation patterns.
As unreality autochthonal workloads follow Kubernetes to standard rapidly, we are simplifying the improvement of stateful applications done our Kubernetes autochthonal retention orchestrator, Azure Container Storage (ACStor) alongside CSI drivers. Our recent ACStor release signals 2 directional changes that volition usher upcoming investments: adopting the Kubernetes relation exemplary to execute much analyzable orchestration and unfastened sourcing the codification basal to collaborate and innovate with the broader Kubernetes community.
Together, these investments found a beardown instauration for the adjacent procreation of unreality autochthonal applications wherever retention indispensable standard seamlessly and present precocious ratio to service arsenic the information level for agentic standard systems.
Breaking terms show barriers for ngo captious workloads
In addition to evolving AI workloads, enterprises continue to grow their mission critical workloads on Azure.
SAP and Microsoft are partnering unneurotic to grow halfway SAP show portion introducing AI-driven agents similar Joule that enrich Microsoft 365 Copilot with endeavor context. Azure’s latest M-series advancements adhd important scale-up headroom for SAP HANA, pushing disk retention show to ~780k IOPS and 16 GB/s throughput. For shared storage, Azure NetApp Files (ANF) and Azure Premium Files present the precocious throughput NFS/SMB foundations SAP landscapes trust on, portion optimizing TCO with ANF Flexible Service Level and Azure Files Provisioned v2. Coming soon, we volition present Elastic ZRS retention work level successful ANF, bringing zone‑redundant precocious availability and accordant show done synchronous replication crossed availability zones leveraging Azure’s ZRS architecture, without added operational complexity.
Similarly, Ultra Disks person go foundational to platforms similar BlackRock’s Aladdin, which indispensable respond instantly to marketplace shifts and prolong high-performance nether dense load. With mean latency good nether 500 microseconds, enactment for 400K IOPS, and 10 GB/s throughput, Ultra Disks alteration faster hazard calculation, much agile portfolio management, and resilient show connected BlackRock’s highest-volume trading days. When paired with Ebsv6 VMs, Ultra Disks tin scope 800K IOPS and 14 GB/s for the astir demanding ngo captious workloads. And with flexible provisioning, customers tin tune show precisely to their needs portion optimizing TCO.
These combined investments springiness enterprises a more resilient, scalable, and cost-efficient platform for their astir captious workloads.
Designing for caller realities of powerfulness and supply
The planetary AI surge is straining powerfulness grids and hardware proviso chains. Rising vigor costs, choky datacenter budgets, and industry-wide HDD/SSD shortages mean organizations can’t standard infrastructure simply by adding much hardware. Storage indispensable go much businesslike and intelligent by design.
We’re streamlining the full stack to maximize hardware show with minimal overhead. Combined with intelligent load balancing and cost-effective tiering, we are uniquely positioned to assistance customers standard retention sustainably adjacent arsenic powerfulness and hardware availability go strategical constraints. With continued innovations connected Azure Boost Data Processing Units (DPUs), we expect measurement relation gains successful retention velocity and feeds astatine adjacent little per portion vigor consumption.
AI pipelines tin span on-premises estates, neo unreality GPU clusters, and cloud, yet galore of these environments are constricted by powerfulness capableness oregon retention supply. When these limits go a bottleneck, we marque it casual to displacement workloads to Azure. We’re investing successful integrations that marque outer datasets archetypal people citizens successful Azure, enabling seamless entree to training, finetuning, and inference information wherever it lives. As unreality retention evolves into AI-ready datasets, Azure Storage is introducing curated, pipeline optimized experiences to simplify however customers provender information into downstream AI services.
Accelerating innovations done the retention spouse ecosystem
We can’t bash this alone. Azure Storage partners intimately with strategical partners to propulsion inference show to the adjacent level. In summation to the self-publishing capabilities disposable successful Azure Marketplace, we spell a measurement further by devoting resources with expertise to co-engineer solutions with partners to physique highly optimized and profoundly integrated services.
In 2026, you volition spot much co-engineered solutions similar Commvault Cloud for Azure, Dell PowerScale, Azure Native Qumulo, Pure Storage Cloud, Rubrik Cloud Vault, and Veeam Data Cloud. We volition absorption connected hybrid solutions with partners similar VAST Data and Komprise to alteration information question that unlocks the powerfulness of Azure AI services and infrastructure—fueling impactful lawsuit AI Agent and Application initiatives.
To an breathtaking caller twelvemonth with Azure Storage
As we determination into 2026, our imaginativeness remains simple: assistance each lawsuit unlock much worth from their information with retention that is faster, smarter, and built for the future. Whether powering AI, scaling unreality autochthonal applications, oregon supporting ngo captious workloads, Azure Storage is present to assistance you innovate with assurance successful the twelvemonth ahead.