DATA PROTECTION
Entertainment and media-focused storage supplier 45Drives announced a major expansion of SnapShield, its server-side cybersecurity platform. SnapShield uses real-time behavioral analysis at the storage server to recognize ransomware-like activity. New Data Exfiltration Protection identifies suspicious file-access behavior that may signal attempted data theft, while a new Centralized Management System gives enterprises and managed service providers a single interface for protecting and monitoring multiple servers, locations and customer environments. More information here.
…
Lakehouse analytics supplier Cloudera has a partnership deal with French AI lab Mistral to combine Cloudera’s hybrid data and AI platform with Mistral’s sovereign AI models to give enterprises a secure path to build, customize, and run AI using their own private data. Mistral’s suite of frontier models and tools will be integrated with Cloudera’s hybrid data and AI platform, enabling customers to run their own custom AI models and applications across 30 exabytes of data. Enterprises will have the deployment flexibility across public, private, on-premises, and fully air-gapped environments, while maintaining consistent governance and control over their context. The partnership will span Mistral’s broad portfolio of frontier AI models and tools, including reasoning, chat, coding, document intelligence, and voice, giving Cloudera customers greater choice in how they apply AI to their enterprise data.
…
Commvault has appointed James Hayes as its first Chief Public Affairs Officer to lead Commvault’s government and public affairs strategy across federal, state, and international governments and regulatory bodies. Hays joins Commvault from IonQ and Tenable, where he led federal advocacy, cybersecurity policy, and engagement on emerging technology standards. He also led EMC’s Government Affairs Office in Washington, D.C. prior to the Dell acquisition.
…
CoreWeave has launched Physical AI Field Engineering, an offering that pairs customer teams with domain specialists to build, validate, and deploy AI across the full engineering lifecycle, from research and development through in-field operations. CoreWeave engineers work alongside a customer's own team, building models from data the customer already owns, like test bench results, simulation output, production sensors, and live telemetry. Each model is validated against the real physics of the customer's systems until it holds up in practice. This involves;
- Strategy: Identifies which problems are actually worth solving with AI, and which data is worth building on.
- Simulation infrastructure: Stands up the GPU, storage, and simulation stack a specific use case needs. For full infrastructure design and scale beyond that, field engineers connect customers into CoreWeave's broader physical AI platform.
- Real-world data: Turns scattered test, sensor, and production data into a model that predicts an outcome, catches an anomaly, or explains a failure, instead of leaving that signal buried and unused.
- Agentic learning: Turns what a model finds into something that changes the physical world - a system recalibrated to run better, a fault caught and corrected before it becomes a failure, or a robot executing a trained skill a customer's team built and deployed.
…
The Information reports China’s CXMT has begun producing HBM3E memory chips in small quantities and plans to expand production in 2027, a major advance for China's chip industry.
…
d-Matrix today announced a new collaboration with Nvidia that includes a multi-year product roadmap, providing d-Matrix Raptor XPU entry into the widely deployed Nvidia AI factory ecosystem. The company (d-Matrix) makes custom processors (XPUs) for the decode phase of infringing while Nvidia GPUs run the profile phase. Raptor uses SRAM and stacked 3D DRAM on top of the compute. The available pool of SRAM memory scales elegantly up to a certain point that falls short of handling the largest reasoning models and the general shift toward significant token consumption.
A stacked 3D DRAM configuration lives in the die-to-die interconnectivity space, which allows it to target 10x better memory bandwidth and 10x better energy efficiency using 3DIMC over HBM4 configurations. 3DIMC (3D digital in-memory compute) stacks a compute/logic die face-to-face on a custom DRAM die using microbumps, eliminating the conventional PHY, interposer, and long-distance data movement of HBM. d-Matrix says “We took passive DRAM we use for capacitors and converted that to an active stack, By doing that we can expose every small bank in the DRAM and directly bring it to the compute engine. With 3D DRAM, we now have the entire 3D surface area to connect, and the signals can be run at the DRAM base which is a few hundred Mhz. By doing that we can get 20 TB/s bandwidth per stack—10x what HBM4 can achieve—at a power consumption of .3-.4 pJ/bit bit, compared to 3-4 pJ/bit.
d-Matrix Raptor is expected to tape-out before the end of the year, and is designed for integration with NVLink Fusion and the MGX rack-scale ecosystem. Raptor is being actively evaluated at AI hyperscalers and frontier labs for its unique memory stacking solution and is backed by more than 100 patents.
…
Fortanix announced a partnership with CrowdStrike to help organizations secure AI across the entire lifecycle. By combining CrowdStrike’s Falcon platform with Fortanix Confidential AI, the integrated system will deliver end-to-end AI security, from threat detection and response, to protecting AI workloads, sensitive data, and proprietary models while actively running in memory. Fortanix Confidential AI creates trusted execution environments where AI workloads execute in isolated, hardware-protected memory. Sensitive enterprise data, prompts, model weights, and inference outputs remain encrypted and protected while in use, preventing exposure to cloud operators, privileged administrators or malicious actors. Together, CrowdStrike and Fortanix will provide organizations with holistic security that protects the AI workload, from data to models and agents, before, during and after execution.
…
SaaS data protector Keepit as surpassed $100 million in annual recurring revenue (ARR), saying its AI Truth Cloud strategy, which reframes data protection as the trusted foundation AI-enabled enterprises run on, rather than simply insurance against loss, is driving customer adoption at scale.Since growth equity investor One Peak's initial investment in September 2020, the company has grown more than 13 times —a greater than 55 percent compound annual growth rate (CAGR) — serving organizations of every size, from single-seat small businesses to global enterprises, across its independent, vendor-neutral cloud infrastructure.
"Backup used to answer one question: Can we get the data back?" said Frederik Schouboe, Keepit co-founder and chief visionary officer. "AI is making companies ask something bigger: Can we trust the data our business is running on? This means data protection is no longer a cost of doing business, but rather the foundation of doing business you can trust.” The same infrastructure that helps mitigate the effects of ransomware and accidental loss now gives enterprises a source of truth they can run AI on with confidence.
…
Micron is offering Taiwan employees payouts worth 35 to 68 months of salary for fiscal 2026, with a minimum cash floor of NT$1.7 million (US$53,809), according to Reuters. This follows local unions saying 80 percent of Micron’s workers there supported a strike as the existing bonus structure does not reflect the company’s record-breaking, AI-fueled revenues. This follows Samsung and SK Hynix making similar moves.
…
Open source database supplier Percona announced a partnership with open source observability company Coroot to bring Coroot Percona Edition to Percona customers. Percona and Coroot can provide full stack insight and observability for cloud-native workloads, rather than looking only at database or cloud infrastructure issues on their own. Coroot can automatically trace an issue across that stack to its root cause, and can directly route teams into a Percona Support case for any database-related findings.
…
Vector database and search company Pinecone has announced full-text search is generally available in Pinecone Database. It adds BM25 keyword ranking and text-match filters to the same index as dense vectors, so agents get the exact order number or error code embeddings miss, with no separate system to tune and maintain. If your search, recommendation, RAG, or agent workload mixes questions with identifiers and exact phrases, Full-Text Search gives you:
- Exactly correct instead of confidently close on the queries embeddings get wrong: SKUs, error codes, case numbers, order IDs, and exact quoted phrases.
- BM25 keyword ranking across multiple text fields per index, with Lucene query syntax, fuzzy matching, and tokenization and stemming in 18 languages.
- Text-match filters that restrict the results of a semantic search to the documents matching the filter. One query both retrieves based on the identifier and ranks the rest by meaning.
- One index and one schema for text fields, dense vectors, and sparse vectors, queried through the Documents API, with no cluster to tune, patch, or upgrade.
- Usage-based capacity by default, metered in the same read units and write units as vector indexes. Dedicated Read Nodes are available for provisioned read capacity, and BYOC for running inside your own cloud.
Learn more here. …
SK Hynix CEO Kwak Noh-Jung has said: "There are no clear signs of a [memory] downturn, so this situation will continue through the end of 2030...Even if a downturn comes, it will be closer to a gradual decline or flat trend rather than a steep drop. After 2030, I believe supply and demand will find balance and the entire AI industry can grow together.”
…