Storage news ticker – February 2

Data trust company Ataccama has been named a Leader in The Forrester Wave: Data Quality Solutions, 2026. Ataccama received the highest score in the Strategy category among all evaluated vendors. Get the report here.

…

Research house Coldago released a new edition of its Modern Data Protection Map.

The 2025 edition selected and evaluated 20 vendors and identified 4 market leaders: Cohesity, Commvault, Rubrik and Veeam. Companies listed by alpha order are: Acronis, Arcserve, Atempo, Bacula, Cobalt Iron, Cohesity, Commvault, Dell, Druva, Eon, HYCU, IBM, Kaseya, Keepit, N-able, NinjaOne, OpenText, Rubrik, Salesforce and Veeam. More info here. You can purchase the document (€7,990) by emailing [email protected].

…

Coldago Research has released its final 2026 Gem List, highlighting five standout companies selected this year. Since the program began, Coldago has identified a total of 85 companies. The Coldago Gem List is the outcome of an annual study that spotlights five players to watch in the months ahead. Companies are selected based on their vision, disruptive potential, product innovation, value proposition, business momentum, recent announcements, and execution – with a strong emphasis on technology and product advances. For 2026, the five companies joining the Gem group are: 9LivesData, InfoScale, Lucidity, StorWare and VergeIO.

…

Commvault has expanded its Google Cloud integration with new capabilities that deliver immutable protection, rapid recovery of data and cloud applications, as well as archiving for compliance support:

• Advancing security and supporting compliance with Commvault’s Air Gap Protect (AGP): Enables immutable, indelible backups that can be stored directly in an isolated, virtually air-gapped location in Google Cloud, helping to isolate data from production systems and to defend against ransomware and insider threats. A new compliance-ready Archive Tier supports regulatory requirements and optimises long-term retention.
• Rebuilding cloud apps and accelerating recoveries with Commvault Cloud Rewind: With this announcement, Google Cloud customers can also use Commvault Cloud Rewind to rapidly and cleanly rebuild cloud applications. This is an important advancement as these applications, and associated dependencies, often power enterprise data and are essential to keeping business running, even during ransomware events or major deployment errors.
• Simplifying compliance search for Google Workspace environments: Expands on existing Google Workspace protection, first announced last year, by adding advanced compliance-search capabilities for Commvault’s eDiscovery offering. Customers can now locate and export relevant data, such as email messages and files, across their Google Workspace backups faster for audits, litigation, or investigations. This expanded support is designed to help organisations reduce the time and costs associated with eDiscovery and support their legal and regulatory obligations.

…

Commvault has hired Dale Smith as Senior Director, Channel Sales for North Europe. HE has joined from Juniper Networks, where he held the position of UK&I Channel Director, before taking on the position of Senior Channel Director, EMEA in 2022.

…

While MIT found last year that 95 percent of generative AI pilots fail to get into production, new data from Databricks, drawing from more than 20,000 customers worldwide, has shed light on how businesses are actually making AI agents work in 2026. Its 2026 State of AI Agents report reveals that businesses are moving from single chatbots to multi-agent systems – use of the latter on the Databricks platform grew by 327 percent in just four months. The data, from fully aggregated, anonymised customer telemetry, found: 77 percent of customers used two or more LLM families (GPT, Claude, Gemini etc.), determining which model works best for different use casesCompanies actively using AI governance tools put 12x more AI projects into production, while companies using evaluation tools on their models put 6x more into production 40 percent of the top AI use cases focus on practical customer concerns such as customer support, advocacy and onboarding. Get the report here.

…

Descope announced updates to its Agentic Identity Hub, an identity platform for AI agents and MCP servers, designed to address the production-deployment gap that is holding back over half of enterprises from adopting AI. While 88 percent of enterprises are using or planning to use AI agents, only 37 percent have moved beyond the pilot phase. Descope’s Agentic Identity Hub 2.0 provides organizations with dedicated identity infrastructure for AI agents and MCP servers, enabling them to manage AI agents as first-class identities alongside human users, add OAuth 2.1 and tool-level scopes to MCP servers, and govern agent access with enterprise-grade policy enforcement. The platform integrates agentic identity management, comprehensive MCP auth, credential vaults, policy controls, and logging capabilities to streamline AI agent deployment while maintaining security standards.

…

Descope and Skyflow announced the first natively integrated identity and data control layer for organizations building MCP servers, addressing identity and data privacy. The integration combines Descope’s authorization capabilities with Skyflow’s runtime data security to enable organizations to deploy MCP servers with user authentication and consent, scope and tool-based access control, and enterprise-grade privacy guardrails. With Descope controlling user and agentic identity and Skyflow controlling data access, organizations can make runtime access control decisions without rebuilding their security infrastructure. You can see the integration in action in this video demo.

…

SaaS data protector Druva launched Threat Watch, a zero-touch, automated cloud-native offering for proactive threat monitoring of backup data. Threat Watch is designed to continuously scan backup snapshots to identify dormant threats and indicators of compromise (IOCs). It’s designed to deliver continuous, peace-time monitoring of backup data and complements threat hunting activities that typically ramp up during an incident. It doesn’t require additional hardware or agents. This scan in-place approach avoids the delays of moving data to separate security tools and enables Druva to offer the industry’s only Data Movement Latency SLA. As a result, detection occurs in near real-time without impacting production performance or increasing infrastructure costs.

It Uses a curated and customer-configurable IOC library, including indicators from CISA, Google Mandiant Threat Intelligence, and Druva ReconX Labs, with support for customer-provided IOCs via upload or API. Built on Dru MetaGraph, Druva’s graph-powered foundation for real-time data intelligence Threat Watch will be able to output threat signals into DruAI to help teams prioritize risk, understand impact, and act with greater confidence. Read more here.

…

Dutch SMB e-commerce design and marketing partner Vistaprint is using Fivetran to power secure, automated data pipelines and reverse ETL. Vistaprint relies on data to enhance personalisation, optimise marketing performance, and streamline operations. It uses AWS, Snowflake, and analytics tools like Statsig to understand how small businesses engage with its platform. Vistaprint has moved from on-prem systems to a cloud-first architecture and adopted a data mesh strategy that treats data as a shared resource across teams. Fivetran automates data movement from Vistaprint’s core systems into Snowflake and activates that governed data back into operational tools using reverse ETL. In paid channels, Vistaprint uses reverse ETL to enrich platforms such as Google and Meta with first-party data, enabling more precise targeting and improved return on ad spend in programmatic advertising.

…

Hammerspace announced the promotion of Tony Asaro to Chief Business Officer. In this expanded role, Asaro will lead Hammerspace’s global revenue organization — including sales, alliances, channel and go-to-market strategy. Hammerspace enters 2026 with strong sales momentum, driven by strategic partner expansion, substantial VAR channel growth (with just under 200 resellers), and international expansion. Over the past year, the company launched its Asia headquarters in Singapore and scaled engagement across China and South Korea, while building new regional coverage for India and the Middle East from Dubai—extending field capacity, partner reach, and customer delivery for sovereign AI and GPU-intensive deployments.

…

IBM has introduced its Storage Scale Multi-Tenancy Reference Architecture v2.04, a comprehensive guide for building secure, scalable, and efficient multi-tenant environments with IBM Storage Scale and IBM Storage Scale System 6000. A blog by Mathias Dietz, RAS Architect at IBM Germany, introduces it. Download the RA here.

…

Salesforce unit Informatica has released its annual global CDO study, revealing a “trust paradox” as organisations accelerate AI initiatives, emphasising the urgent need to improve data management and upskill workforces. The “CDO Insights 2026: Data governance and the trust paradox of data and AI literacy take center stage” report surveyed 600 data leaders, with highlights for Europe and the UK including:

• Nearly seven in ten European businesses are racing to become agentic by the end of Q1 2026.
• UK businesses are moving at a slower pace, with just 61 percent planning to start the shift to becoming an agentic enterprise.
• A lack of data and AI literacy has caused a clear ‘trust paradox’: while European data leaders believe most employees (61 percent) trust their organisations’ AI data, 96 percent say staff need more training to use it responsibly.
• AI visibility and governance isn’t keeping pace with AI adoption – 77 percent of European leaders say their company’s AI visibility and governance has not kept pace with employee use of AI.
  

…

Interesting Nikkei article on Kioxia and Sandisk JV about recent announcement extending JV arrangements. Here’s a snippet:

…

Kioxia announced the EXCERIA PRO G2 SDXC UHS-II Memory Card Series, available in a faster V90 version for professional sports and wildlife photographers as well as in a more economical V60 version for everyday content creators.The V90 has capacities up to 512GB for high end 8K photo or video shoots and offers reads up to 310 MB/s and writes up to 300 MB/s^. V60 Series delivers reads/writes up to 300 MB/s and 250 MB/s, as well as capacities up to 512 GB.

^…

Kioxia has begun sampling new UFS Ver. 4.1 embedded memory devices with 4-bit-per-cell, quadruple-level cell (QLC) technology. Designed for read-intensive applications and high-capacity storage needs, the new devices are powered by Kioxia’s 8 generation BiCS FLASH 3D flash memory technology. It boosts sequential write speeds by 25 percent, random read speeds by 90 percent, and random write speeds by 95 percent compared to the previous generation (UFS 4.0/BiCS6 QLC UFS). Write Amplification Factor (WAF) is also improved by max. 3.5× (with WriteBooster disabled). Available in 512 GB and 1 TB capacities. There is a reduced package size compared to the previous QLC UFS: 11×13 mm → 9×13 mm.

…

NAND flash shortages are no longer a temporary hurdle—they are a fundamental challenge to data center growth, with procurement timelines now stretching into months. If your projects are stalled waiting for specific vendor SKUs, it’s time to change the game. The latest blog from tech evangelist and VP at Lightbits Labs, Carol Platz, offers 4 software-defined storage strategies to create a resilient architecture that insulates you from supply chain volatility.

…

Pete Paisley’s LTO Show Podcast is up and running. Check it out on YouTube here.

…

In response to MySQL community relationship building developments and adding vector capability at Oracle, MariaDB Chief Product Officer Vikas Mathur said: “We are pleased to see the MySQL ecosystem evolving to include more of the innovations MariaDB has championed for years. Our goal has always been to push the boundaries of what’s possible in open source, and seeing others follow that lead is a testament to the path we’ve paved. While Oracle treated MySQL as a ‘little database’ side project, MariaDB has been busy building the future (and hiring the best of the MySQL talent that Oracle let go).

“MariaDB has spent years delivering innovations that forced Oracle to update MySQL – from columnar analytics to parallel advanced replication to our launch of native vector search last year. We didn’t wait until it was convenient to open the doors to these innovations; we built them into the core of our server because that’s what a modern open-source database requires.

“MySQL users now face a clear choice: remain with a vendor that only innovates when their hand is forced, or join MariaDB, which is 100 percent focused on what’s next. Since MariaDB is the easy button when migrating out of MySQL, future-proofing your stack is now just a click away.”

…

Micron has started construction of an advanced wafer fabrication facility located within the company’s existing NAND manufacturing complex in Singapore. It’s a planned investment of approximately $24 billion over 10 years and is designed to ultimately provide 700,000 square feet of cleanroom space. Wafer output is scheduled to begin in the second half of cy2028. Micron’s previously-announced HBM advanced packaging facility, also located in the same Singapore manufacturing complex, is on track to contribute meaningfully to Micron’s HBM supply in cy2027. The company expects opportunities for synergies between NAND and DRAM production.

…

Nvidia and CoreWeave announced an expansion of their relationship to enable CoreWeave to accelerate the buildout of more than 5 gigawatts of AI factories by 2030 to advance AI adoption at global scale. Nvidia has invested $2 billion in CoreWeave Class A common stock at a purchase price. The companies intend to:

• Build AI factories developed and operated by CoreWeave using Nvidia technology
• Leverage Nvidia’s financial strength to accelerate CoreWeave’s procurement of land, power, and shell to build AI factories.
• Test and validate CoreWeave’s AI-native software and reference architecture, including SUNK and CoreWeave Mission Control, to unlock deeper interoperability and work toward inclusion within Nvidia’s reference architectures for Nvidia’s cloud partners and enterprise customers.
• Deploy multiple generations of Nvidia infrastructure across CoreWeave’s platform through early adoption of Nvidia computing architectures, including the Rubin platform, Vera CPUs, and BlueField storage systems.

Good news for BlueField-4 storage software and chassis suppliers.

…

OpenAI scaled Azure PostgreSQL to power 800 million ChatGPT users. It says, “Even as our infrastructure has evolved, PostgreSQL has remained unsharded, with a single primary instance serving all writes. … We added nearly 50 read replicas, while keeping replication lag near zero, maintained low-latency reads across geo-distributed regions, and built sufficient capacity headroom to support future growth. … We consistently deliver low double-digit millisecond p99 client-side latency and five-nines availability in production. And over the past 12 months, we’ve had only one SEV-0 PostgreSQL incident (it occurred during the viral launch of ChatGPT ImageGen, when write traffic suddenly surged by more than 10x as over 100 million new users signed up within a week.) Interesting read.

MariaDB Chief Product Officer Vikas Mathur has blogged about this. What if OpenAI Used MariaDB Instead of PostgreSQL to Handle 800 Million Users? He says: “Interestingly, many of the significant hurdles OpenAI encountered while pushing Postgres to its limits are areas where MariaDB’s architecture and feature set offer immediate, built-in mitigation. For organizations facing scaling demands, understanding these differences is crucial… For modern, rapidly scaling businesses where high write throughput, low operational latency, massive concurrency, and agile operations are non-negotiable, MariaDB offers an architectural shortcut around some of the most frustrating scaling challenges faced by PostgreSQL users.”

…

Vector database company Pinecone has an ETL function called Pinecone Assistant n8n node. With it you can upload files (PDFs, docx, text, JSON, Markdown) and start querying immediately. The entire pipeline is handled automatically: chunking, embedding, vector search, query planning, and reranking. Find out more here.

…

Pure Storage sales AE Daniel Gibney has written an excellent LinkedIn post: When Flash Gets Scarce: How SSD Shortages Expose the Weak Links in Enterprise Storage, discussing how Pure’s DFMs make better use of NAND chips than SSDs, delivering more usable capacity to customers. Obviously he praises Pure’s DFMs but, even so, it’s worth 5 minutes of your time reading this.

…

RAIDS AI, an AI safety monitoring platform, has announced its commercial launch. Following a successful beta testing phase launched in October 2025, the platform is now live to support all organizations deploying or using AI. Co-founded by Nik Kairinos and Brett Stonefield, the RAIDS AI platform detects and alerts rogue AI behaviour in real time. It works by monitoring AI behaviour and highlighting deviations before they cause failures, bias or regulatory exposure. The platform enables the safe and responsible deployment of AI and gives its users the confidence to scale AI without compromising on safety.

…

Samsung has raised DRAM prices by 70 percent in Q1 and is now hiking NAND flash prices by 100 percent, according to reports from South Korea’s local press). More here.

…

Australian Neocloud SharonAI Holdings announced the expected deployment of a 1K Nvidia B200 cluster at NEXTDC’s Tier IV M3 Data Center in Melbourne, Australia. The deployment is delivered under Sharon AI’s existing Lenovo TruScale agreement, supporting the scalable rollout of the 1,000-unit B200 cluster. Lenovo’s servers are integrated with VAST Data storage. This expansion significantly bolsters Sharon AI’s existing GPU fleet, which currently includes Nvidia A40, L40s, H100, H200 and now B200 architectures. Looking further into 2026, Sharon AI remains committed to the cutting edge of accelerated computing, with plans to deploy B300 and GB300 systems to maintain its position as a primary infrastructure provider for hyperscale, enterprise, government and research customers in Asia-Pacific.

…

TrendForce is forecasting significant memory product price rises, saying “The quarter-over-quarter increase in conventional DRAM contract prices has been revised upward, from a previous estimate of 55-60 percent to now 90-95 percent. Similarly, NAND Flash contract prices are now expected to rise from 33-38 percent to 55-60 percent. Further upward adjustments may still occur.

…

Veeam has made three senior hires:

SVP Worldwide Cloud Sales Brandt Urban, promoted to Chief Business Development Officer (CBDO), reporting directly to CEO Anand Eswaran. Brandt will lead corporate development, mergers and acquisitions, strategic partnerships, and corporate strategy. He will work closely with product and Go-to-Market teams to translate strategy into execution, expanding Veeam’s value proposition and accelerating growth through targeted ventures, alliances and acquisitions. “Brandt will be instrumental in sharpening Veeam’s continued long-term leadership in data resilience and safe AI.”

Tony Colon, Chief Customer Officer (CCO), steps into the newly established CCO role, reporting to CRO John Jester. With more than 20 years of experience leading global customer-facing organizations, he will own the entire customer journey, ensuring every interaction from onboarding to renewal is seamless and aligned, overseeing Renewals, Customer Success, Professional Services and Technical Support teams, partnering closely with sales leaders to maximize customer satisfaction, retention, and growth. “His leadership at ServiceNow, Cisco, and Salesforce has transformed Customer Success teams into billion-dollar revenue engines, and his focus will be on advancing Veeam’s leadership in data resilience, ransomware recovery, and cyber resilience.”

Michael Rau joins as VP of Worldwide Partners, reporting directly to John Jester. Michael will lead Veeam’s global partner business, focusing on developing consistent programs, incentives, and solutions across distribution, resell, SaaS, services, and strategic alliances. He’ll also oversee Veeam’s most critical partnerships, including HPE, Lenovo, Pure Storage, CrowdStrike, Palo Alto, TCS, and Kyndryl. “With nearly a decade at F5 and senior roles at Cisco, Michael brings deep technical and partnership expertise that will further differentiate Veeam’s offerings and unlock even greater value for partners.”

…

Western Digital is getting into quantum computing. CEO Irving Tan told his Q2 FY 2026 earnings call audience: “Last month, we announced a strategic investment in Qolab, which combines our expertise in material science and precision manufacturing with Qolab’s breakthrough approach to quantum hardware design. Working with Qolab, we aim to advance next-generation nanofabrication processes that improve qubit performance, reliability and scalability.”

…

William Blair analysts produced an interesting SSD Memory supply chain diagram:

…

Chinese NAND fabber YMTC is pulling production forward from 2027 at its being-built Wuhan Plant 3 to start mass production wafers in the second half of 2026. IT has 270-layer 3D NAND, its gen 5, and apparently reckons it can build NAND market share whilst competitors, like Samsung (286-layer), SK Hynix (321-layer) and Micron (276-layer) focus more on DRAM and HBM than NAND. Counterpoint Research says YMTC had 10 percent NAND shipment market share in Q1 2025, and 13 percent in Q3 2025, close to that of Micron. It wants 15 percent. More here and here.

…

YMTC is getting into HBM production. At the start of 2025, when SK Hynix was supplying third-generation high bandwidth memory (HBM3E) chips, ChangXin Memory Technologies (CXMT) was still developing HBM2 technology, yet reports surfaced by mid-2025 that it had leapfrogged into HBM3 development. Although there have been no signs that CXMT’s efforts have succeeded, Chinese firms have far more room to learn from trial and error. Read more here.

…

Phison CEO Pua Khein Seng thinks that YMTC could be one of the world’s largest flash suppliers by 2029-2030. Read more in TechRadar Pro.