Storage news ticker - 13 March

Entertainment and media-focused storage supplier 45Drives announced a strategic partnership with Lime Technology to integrate Unraid OS directly into its 45HomeLab hardware to deliver a fully integrated, builder-grade homelab platform- factory pre-installed, validated, and tested in its North American manufacturing facilities. For the first time, Unraid users will be able to purchase premium 45HomeLab systems with the OS tightly validated on the hardware from day one- no guesswork, no compatibility questions, no DIY burn-in.

…

Ceramic-coated tablet archive technology supplier Cerabyte, working with the Institute of Materials Science and Technology at TU Wien (Technical University Vienna) produced the world’ smallest QR code. It covered an area of just 1.98 square micrometers — smaller than most bacteria and 37 percent the size of the previous world record holder. The record has now been verified and officially entered into the Guinness Book of Records.

…

Object storage supplier Cloudian ran an independent survey of 203 enterprise IT decision-makers to reveal that the era of cloud-first AI strategies is giving way to a more deliberate, workload-driven approach. It found that 93 percent of enterprises have already repatriated some AI workloads from public cloud, are in the process of doing so, or are actively evaluating repatriation. Nearly four in five (79 percent) have already moved workloads, while 73 percent plan to further shift toward on-premises or hybrid infrastructure over the next two years. Data sovereignty is the top driver of on-premises adoption. Cloud AI costs are exceeding expectations. Performance demands favor on-premises deployment. AI budgets are growing to fund this shift. The full Enterprise AI Infrastructure Survey 2026 report, including detailed methodology, charts, and analysis across all findings, is available for download at Cloudian.com.

…

Cubbit, the first geo-distributed cloud storage enabler, and Worldstream, a Dutch cloud infrastructure provider operating its own European data centres and network, today announced a new Business Alliance Partnership to equip Dutch enterprises with fully sovereign, secure, cost-predictable, and flexible cloud storage services. Worldstream is integrating DS3 Composer — Cubbit’s software-defined object storage — across its three data centres in the Netherlands to create and offer its own S3 cloud storage service, fully hosted and operated by Worldstream on its strategic assets. Worldstream is launching the service with an initial capacity of 1 petabyte, with further expansion planned in the coming years. Key use cases range from backup and active archiving to content delivery.

…

AI analytics supplier Databricks announced Genie Code, an autonomous AI agent designed to help data teams move from AI-assisted coding to agent-driven development. Databricks has also acquired Quotient AI, whose technology will help evaluate and improve AI agents once they are running in production. Quotient automatically monitors agent performance — measuring answer quality, catching regressions early, and pinpointing failures — feeding a reinforcement learning loop that keeps agents improving over time. Genie Code features:

• Agent-driven data work: As agentic coding tools reshape software engineering, Genie Code brings the same approach to data teams - enabling them to move from prompting a copilot to delegating tasks like building pipelines, developing ML models, debugging failures and shipping dashboards.
• Learns how teams work: The system improves over time, updating its internal instructions based on past interactions and coding preferences.
• Built for enterprise data: Powered by Unity Catalog, it understands enterprise data, semantics and governance policies, and can proactively maintain production systems.
• Quotient AI acquisition: The acquisition aims to advance continuous evaluation and reinforcement learning for AI agents, strengthening key products like Genie, Genie Code and Agent Bricks.

…

HPC and AI storage supplier DDN has hired Regis Paquette as its SVP, Partner Ecosystem, to lead its partner ecosystem strategy in a pivotal new role designed to help DDN and its AI Data Intelligent platform become the global accelerator of AI adoption in the market. Most recently at Canonical, he built the company’s partner ecosystem to pivot the revenues the from a direct sales driven model to a highly scalable, partner-led, highly profitable business, across public clouds, OEMs, ISVs, Global System Integrators and channels. DDN is on a bit of an exec hiring spree.

…

MRAM supplier Everspin announced its UNISYST MRAM family, built as a natural extension of Everspin’s existing PERSYST MRAM platform. UNISYST is a unified code-and-data MRAM architecture that bridges traditional configuration memory and higher-density persistent storage, extending MRAM into traditional NOR flash applications where superior performance, endurance and reliability are valued.

Everspin will initially offer the UNISYST family in densities ranging from 128 megabits to 2 gigabits, using a standard xSPI interface operating up to octal SPI at 200MHz. The devices are planned to feature AEC-Q100 Grade 1 qualification and minimum 10-year data retention at extreme temperature, supporting demanding environments across automotive, aerospace, industrial and edge AI applications.

Engineering samples of UNISYST are expected to be available in the fourth quarter of 2026, with additional densities and options to follow. Read more here.

…

Hammerspace is partnering with Secuvy to deliver an integrated “Data-First” approach that keeps data continuously AI-ready, turning raw data into secure AI outcomes. Hammerspace provides the performance and orchestration layer so AI pipelines can reach distributed file and object data in place and move only what’s needed to the right compute at the right time. At the same time, Secuvy adds the intelligence layer, continuously identifying sensitive data and associated risks so privacy and governance controls can be applied consistently across hybrid and multi-cloud environments. Together, the companies unify data discovery, classification, privacy, and AI governance without the need to duplicate data. Find out more here. 

…

SaaS data protector HYCU announced Legal 360, a resilience offering built specifically for law firms and legal departments. The company also previewed a new business continuity and resilience capability that enables authorized lawyers to securely access protected iManage Cloud matter documents during cloud outages. Legal 360, built on the R-Cloud platform, protects iManage Cloud, Microsoft 365, Entra ID, Docusign, and supporting infrastructure as part of a single resilience model. Find out more here.

…

SaaS data protector Keepit has signed a strategic agreement with Hammer Distribution to enhance its offering in the UK and Ireland. Through local data centers, Keepit and Hammer will support customers in staying compliant with evolving UK regulations. Keepit’s cloud backup and recovery platform currently secures fourteen SaaS applications and more are coming.

…

Next week at NVIDIA GTC 2026, Kioxia America will showcase its latest high-performance memory and solid-state drive (SSD) products.

• Optimizing Memory Footprint While Improving Scalability for Vector Databases with AiSAQ Technology - Featuring Kioxia CM9 Series Enterprise NVMe SSDs. 
• Super High IOPS SSD Emulator for AI-Ready Storage - SSD emulator capable of delivering over 100 million IOPS, designed to support demanding environments such as SCADA. 
• High-Capacity QLC Storage for Fast Retrieval of Large Datasets - Featuring 245.76 TB LC9 Series Enterprise NVMe SSDs. 
• Kioxia Enterprise and Data Center SSD Portfolio - A comprehensive look at NVMe SSD solutions engineered for servers and storage systems, offering the performance, scalability, and reliability required for modern data center workloads. 

…

Open-source object storage supplier MinIO announced AIStor Table Sharing allowing enterprises to securely share on-prem data directly with the Databricks platform for instant access to fresh data for real-time analytics and intelligence. AIStor natively integrates Delta Sharing open protocol directly into the object store. 

…

Nutanix’ eighth annual Enterprise Cloud Index (ECI) survey and research report showed that 85 percent of respondents report that AI is accelerating their adoption of containers to improve speed, reliability, and scalability. Organizational silos create new AI risk. Shadow IT is creating AI challenges. Agents unlock enormous potential for organizations. Data sovereignty is non-negotiable. Containers are the foundation of modern applications, with AI as the key driver. Download the full eighth annual ECI report here.

…

CXL semiconductor supplier Panmnesia has signed a strategic partnership with SK Telecom, South Korea’s largest telco and a leading AI company to address data center inefficiencies by the two jointly developing a CXL-based next-generation AI data center architecture. They propose a disaggregated architecture in which computing resources are separated by type and flexibly composed as needed. Instead of being confined within servers, CPUs, GPUs, and memory are interconnected at the rack level through a CXL Fabric Switch, operating as a unified system. They will will eliminate Ethernet involvement in computational paths and transition to CXL.

At the core of this architecture is the Panmnesia Link Controller, an electronic component that can be integrated into CPUs, GPUs, AI accelerators, and memory devices. Within each device, it enables direct communication over CXL, replacing data transfer that previously required multiple data copies into simple memory access operations. Furthermore, the architecture enables GPU-to-GPU and GPU-to-memory communication without software intervention, significantly improving processing efficiency. As a result, AI data centers can deliver higher performance without increasing the number of GPUs.

The two companies plan to validate the next-generation AI data center architecture by running real AI models and comprehensively evaluating GPU and memory utilization, latency, and throughput by the end of this year. Following this, they intend to conduct proof-of-concept deployments in large-scale AI data center environments and pursue commercialization and business expansion.

…

Panmnesia has signed a strategic partnership with Openchip, a European AI accelerator design company. The two aim to enhance performance, scalability, and resource efficiency in AI accelerator and infrastructure environments. Openchip will tackle the data movement bottlenecks by combining RISC-V accelerators with Panmnesia's advanced link technologies.

…

Parallel filesystem supplier Quobyte has appointed Andrew Perry as VP of Sales.Incumbent Mark Lawrence is transitioning out. We’re told “Perry joins Quobyte from ScienceLogic, where he served as VP of Sales for the Americas, consistently driving record-breaking revenue and expanding the company's footprint across the enterprise and public sectors. His career also includes senior leadership roles at WEKA, SpringPath and Violin Memory, where he was instrumental in scaling high-performance sales teams and navigating the transition from legacy hardware approaches to modern, cloud-native software architectures.” 

The vastly experienced Perry has been around the LinkedIn block. “His immediate focus will be on expanding Quobyte’s presence in key verticals such as Neoclouds, Financial Services, Media & Entertainment, Life Sciences, and Research—sectors where Quobyte’s ability to deliver linear scalability and 100 percent uptime is a critical competitive advantage.” 

… 

SW RAID supplier Xinnor has released new benchmark results showing how xiRAID Opus, its NVMe-over-RDMA fabric. combined with ZenFlow, helps overcome GPU memory limitations in large-scale AI model training. xiRAID Opus aggregates local and networked NVMe volumes into high-throughput virtual volumes, delivering what Xinnor believes is the fastest NVMe-oF fabric available. The full benchmark analysis, including detailed step-time data, I/O characterization, and per-model-size results, is available on the Xinnor blog.

…