Data exchange automator **Adeptia **says that the idea that AI agents will replace SaaS apps reflects a fundamental misunderstanding and oversimplified explanation of what enterprise SaaS actually does. AI agents can help query data, but can’t fix malformed EDI files or guarantee HIPAA compliance. The future isn’t agents replacing SaaS, but rather intelligent SaaS platforms that leverage AI while respecting the realities of how enterprises operate.

…

Data protector Arctera announced new AI-powered features in Arctera InfoScale to identify and respond to ransomware indicators in real time. By learning behavioral patterns across applications, storage and infrastructure, Arctera InfoScale is now able to instantly recognize the traits of a successful ransomware breach as soon as it happens, automatically triggering responses to contain the impact of an attack. It correlates anomalies to specific applications, volumes, or clusters, helping teams quickly isolate and investigate issues. More info here.

…

Secure data management supplier Axoflow launched its Security Data Layer (SDL) to fix problems in log management, SIEM optimization, pipeline reliability, and data accessibility. Axoflow’s pipeline automatically classifies, reduces, normalizes, and enriches data. SDL enables security teams to store, query, and process security data in diverse deployment scenarios, from storing temporal data right on the processing nodes – with federated search across the nodes – to tiered data lakes to meet specific use cases. It automates data collection, transformation, and tiered storage across hybrid environments. The architecture allows organizations to decouple storage from their SIEM, run analytics and AI models on local data, and send only what is needed to external systems. This provides greater cost control and reduces vendor lock-in. More info here.

…

CelerData says its open-source StarRocks 4.0 data lakehouse delivers up to 60 percent faster performance year-over-year through deep engine optimizations, keep systems stable under heavy concurrency with smarter spill handling, and make JSON analytics fly using Flat JSON V2 so teams can query logs and clickstreams at 3 to 15 times the previous speed without flattening or pipeline rewrites.

…

Data protector **Commvault **announced conversation has become the new interface for managing enterprise resilience. Its Model Context Protocol (MCP) server, a policy-based bridge between enterprise systems and popular GenAI assistants like OpenAI’s ChatGPT Enterprise and Anthropic’s Claude. Users can interact with Commvault Cloud in simple natural language to configure, manage, and execute resilience tasks – from setup to status checks and corrective actions. A conversation between the user and the GenAI assistant with Commvault’s MCP server running in the background, could look something like this:

• User: “Is my instance of Docusign backed up?”
• GenAI response: “You don’t have a Docusign backup set up yet. Would you like me to set that up so that you have the necessary configuration in place?”
• User: “Yes.”
• GenAI response: “Perfect, I’ll set that up for you now” and the AI assistant may follow up with the job ID and additional questions like, “Would you like me to back this up now and set up the backups on a repeatable schedule?”

This same conversational experience applies across all workloads protected by Commvault Cloud, not just a single application, giving teams a consistent, voice- or text-driven way to manage protection at scale. For more on how this works and to see this in action, watch this demo and read a blog.

…

**Commvault **introduced Data Rooms, a secure environment that enables enterprises to safely connect their trusted backup data to the AI platforms they rely on, or to their own AI initiatives, such as internal data lakes. By combining governed, self-service access with built-in classification and compliance controls, Data Rooms bridges the gap between data protection and data activation, helping organisations transform backup data into AI-ready assets without adding new risk or complexity.

Authorised users can locate and prepare data directly from backup repositories across on-premises and cloud environments. Built-in governance helps maintain control so that approved, access policy-compliant datasets can be safely shared and exported, with classification, sensitivity tagging, and audit trails automatically applied. Additionally, Data Rooms operate within Commvault Cloud’s zero-trust architecture, leveraging role-based access controls (RBAC) and encryption at rest and in transit. These safety measures can give organisations confidence that their data remains protected, governed, and traceable from backup to activation.

To learn more about Commvault’s Data Rooms offering, read this blog.

…

Coveware by Veeam unveiled its Q3 ransomware report, revealing that ransom payments have fallen to their lowest level on record, pushing attackers to get creative, from mass low-cost campaigns to bribing insiders for access. Key findings include:

• Ransom payment rates dropped to just 23%, the lowest ever recorded by Coveware
• Average payments fell 66%, while median payments dropped 65% to $140,000
• The Akira RaaS group dominated, driving attack volumes through a low-cost, high-volume model
• Insider bribery emerges as a new tactic, with the Medusa gang offering employees a 15% cut of ransom proceeds
• Large enterprises increasingly refuse to pay, squeezing profits and pushing cybercriminals toward more targeted, high-cost operations

See the full findings here.

…

Search AI supplier Elastic announced Streams, agentic AI-powered SW, which uses AI to automatically partition and parse raw logs to extract relevant fields, greatly reducing the effort required of Site Reliability Engineers (SREs) to make logs usable. Streams also automatically surfaces significant events such as critical errors and anomalies from context-rich logs, giving SREs early warnings and a clear understanding of their workloads, enabling them to investigate and resolve issues faster. Streams in Elasticsearch is available today in both serverless and version 9.2. Read this blog to find out more.

…

EnterpriseDB (EDB), the Postgres data and AI supplier, announced its inclusion in the Nvidia AI Factory for Government reference design, a blueprint that enables government agencies and high-assurance enterprises to deploy sovereign, secure, and scalable AI infrastructure. EDB Postgres AI leverages Nvidia NeMo Retriever and NIM microservices to deliver a secure, high-performance path to production AI. EDB is actively working toward FedRAMP Authorization, leveraging the Game Warden platform to accelerate accreditation. EDB PG AI supports compliance with STIG, FIPS 140-3, and delivers Iron Bank–hardened containers for U.S. agency.

…

HighPoint Technologies unveiled the Rocket 7624A PCIe Gen 5 x16 to 2-MCIOx8 NVMe RAID Adapter. It delivers PCIe Gen 5 x16 bandwidth and professional-grade NVMe RAID capabilities at an unprecedented sub-$1,000 price point — making it the industry’s most Cost-Performance PCIe Gen 5 x16 NVMe RAID adapter to date. Read more here.

…

IBM says its watsonx.data developer edition is a free, locally deployable version of IBM’s open data lakehouse platform. It is designed for developers, data engineers, and data scientists to rapidly explore, learn, and prototype. The developer edition includes core lakehouse features such as multi-engine query capabilities, Apache Iceberg table format, and sample datasets. It is optimized for single-user, local use cases, with a simplified installation process and no entitlement requirements. Download it here.

…

IBM herald good news for all home lab fan’s 😉 as Storage Scale Developer Edition Trial has been updated to v6.0 Code Level. It’s available on Red Hat Enterprise Linux on x86_64. This edition provides all the features of the Data Management Edition (DME) and is limited to 12 TB per cluster. This edition is made available to customers free of charge to try IBM Storage Scale features in test setups: https://www.ibm.com/account/reg/us-en/subscribe?formid=urx-41728 . The download ZIP archive is single file Storage_Scale_Developer-6.0.0.0-x86_64-Linux.zip (1.7 Gb)Storage Scale.

…

Financial industry benchmarker STAC ran a STAC-M3 benchmark audit on a stack featuring KX’s KDB+ 4.1 database system sharded across six Supermicro Storage SuperServer SSG-222B-NE3X24R servers. Each system contained 24x 12.8 TB Micron 9550 MAX NVMe SSDs and two Intel Xeon 6767P 64-core CPUs. Unlike many previous tests using a shared disk array, this SUT used local storage across all nodes. This system outperformed all publicly disclosed results on 19 of 24 Kanaga mean-time tests, 3 of 3 Antuco 50-user and 100-user benchmarks, and 3 of 5 Kamaga throughput benchmarks. The benchmark report, including full test results, is available to download from the STAC website.

…

Nvidia launched its BlueField-4 data processing unit (DPU), with 800 Gbps throughput. at its GTC event in Washington. It combines an Nvidia Grace CPU and ConnectX-9 networking to deliver 6x the compute power and support AI factories up to 4x larger than possible with BlueField-3, accelerating gigascale AI infrastructure. It features multi-tenant networking, rapid data access, AI runtime security and cloud elasticity with native support for DOCA microservices — containerized services that secure, scale and simplify AI deployment and operations. It supports a multiservice architecture with native service function chaining, enabling seamless integration and management of multiple network, security and storage services within a single, unified framework.

Cisco, DDN, Dell Technologies, HPE, IBM, Lenovo, Supermicro, VAST Data and WEKA have planned adoption of BlueField-4, with the DPU entering early availability as part of Vera Rubin platforms in 2026.

…

PNY announced the CS3250 M.2 2280 form factor NVMe PCIe Gen5 x4 SSD with fast NVMe PCIe Gen5 x4 performance. With sequential read speeds of up to 14,900 MB/s and write speeds up to 14,000 MB/s, and 1 and 2 TB capacities, it is designed for AI developers, gamers, content creators, and performance-driven professionals. The drive has a 5-year limited warranty or TBW3 and costs: 1TB: $125.99, 2TB: $219.99.A 4 TB version is coming soon.

…

The Danish Centre for AI Innovation (DCAI ) which operates the Gefion AI supercomputer, has launched of one of Europe’s first fully sovereign, end-to-end encrypted AI infrastructure that unifies high-performance compute and storage in a single platform. The storage is 140 petabytes (PB) of usable storage capacity—including 80 PB of performance-optimized NVMe SSDs running on WEKA’s NeuralMesh storage software. Gefion now provides a complete AI factory where European organizations can train models, manage data, and deploy intelligent systems without compromising on sovereignty, security, or performance.

…