AIC expanded its Nvidia BlueField accelerated storage portfolio with the F2032-G6 2U JBOF, optimized for accelerating AI inference by storing KV cache context. The platform supports up to 32 E3.S/L NVMe SSDs, enabling up to 8 PB of storage in a single 2U chassis using 15, 30, 60, 122 or 256 TB SSDs. It has a High Availability, dual-active node architecture with two or four BlueField-4 DPUs, and can also support ConnectX-9 SuperNICs. It supports DOCA microservices to deliver acceleration for Nvidia’s Inference Context Memory Storage Platform.
…
DapuStor has announced a 245 TB SSD using QLC NAND and a PCIe Gen5 bus. It’s intended for AI data lakes, vector databases, and large storage pools where data stays online and is accessed frequently. Other capacitors include 122 TB. Other 245+ TB SSD actual or intended suppliers are Huawei, Kioxia, Micron, Samsung, Sandisk, SK hynix and subsidiary Solidigm.
…
Dell has a partnership with Diskover Data whereby “Diskover Data can rapidly index the custom metadata already present in PowerScale, dramatically enhancing data discovery and analysis for end users. Additionally, Diskover Data can write new metadata directly into the PowerScale file system, providing a layer of external context that is permanently associated with the file.” Customers can use “metadata indexed in Diskover Data to feed the federated query engine of the Dell Data Analytics Engine” and AI. DEll says “PowerScale is more than just a storage solution—it’s the storage engine powering the modular Dell AI Data Platform and collaborating with the Nvidia AI Data Platform to shape the future of AI infrastructure.” Read a blog to find out more.
…
DDN has announced support for Nvidia’s Vera Rubin platform. It says DDN’s platform is engineered to align with Nvidia’s latest hardware and software innovations, including Spectrum-X Ethernet for Storage and DOCA-accelerated services on BlueField-4, enabling:
- Exascale data access to feed high-density Rubin GPU configurations at line rate
- Distributed KV cache tiering supporting the Nvidia Inference Context Memory Storage Platform, which expands inference context beyond GPU memory while maintaining ultra-low latency
- Network-integrated storage services that leverage BlueField-4 acceleration engines for metadata processing, telemetry, and control-plane operations
- Dynamic, telemetry-driven data placement to optimize performance as workloads shift in real time
…
Reuters reports Marvell Technology saying it will buy networking equipment provider XConn Technologies in a deal worth about $540 million, as the chipmaker doubles down on datacenter hardware amid a race to expand artificial intelligence infrastructure.
…
MSP-focussed data protection cloud SW supplier N-able has appointed Patrick Pulvermueller to its board. He is the founder and managing director of Redsalt GmbH and affiliated companies. Pulvermueller previously served as the CEO of Acronis, and held various executive roles at GoDaddy. As N-able continues to expand its cybersecurity footprint and deliver advanced business resilience solutions, N-able believes Pulvermueller’s expertise in cybersecurity leadership will provide a valuable perspective as the company shapes its future and explores new opportunities.
…
Peer Software is featured in a Gartner Voice of The Customer for Hybrid Cloud Storage report. Request a copy here.
…
Phison announced the E37T PCIe Gen5 controller at CES 2026. the E37T brings core architecture refinements over previous generations to support the latest 3D NAND with up to 4800 MT/s speeds, providing a 38% performance gain for maximum value. The new controller features DRAM-less design and 4-channels making the E37T the ideal single-sided design for notebooks, handhelds and compact systems that require M.2 2280, 2242, or 2230 form factor storage without sacrificing performance.
…
Samsung has announced preliminary Q4 2025 revenues of ₩93 trillion ($63.3 billion), up 22.7% Y/Y and with DRAM accounting for ₩28.2 trillion ($19.2 billion), and a profit of ₩20 trillion ($13.6 billion), up 207.7%. HBM memory and DRAM are currently extremely profitable as supply is constrained and (AI-driven) demand high and enduring for the next year or so.
…
Snowflake has signed a definitive agreement to acquire AI-powered observability supplier Observe. With this acquisition, Snowflake will deliver the next generation of AI-powered observability, built on open standards and designed for the scale, complexity, and economics required by modern AI-driven enterprises. Observe was built on Snowflake since its inception and the two will provide:
- Agentic AI for faster troubleshooting: The combination of Observe’s AI-powered Site Reliability Engineer (SRE) with trusted data in Snowflake enables a shift from reactive monitoring to proactive, automated troubleshooting. Observe’s AI SRE leverages a unified context graph that correlates logs, metrics, and traces, allowing teams to detect anomalies earlier, identify root causes faster, and resolve production issues up to ten times faster, improving operational resilience as systems grow more distributed, dynamic, and autonomous.
- An open-standard architecture built for scale: The acquisition also establishes a unified, open-standard observability architecture based on Apache Iceberg and OpenTelemetry, standards which Snowflake has continuously contributed to. This approach allows enterprises to manage massive telemetry volumes using economical object storage, elastic compute, and interoperable standards, an essential foundation for operating next generation AI agents and applications at scale. By treating telemetry as first-class data within the Snowflake AI Data Cloud, enterprises can apply analytics and AI consistently across observability and business data, with greater flexibility, governance, and efficiency.
- Full telemetry data retention with efficient economics: As AI-driven applications generate unprecedented volumes of logs, metrics, and traces, enterprises have increasingly been forced to rely on sampling and short retention windows to manage cost. By unifying Observe’s AI-powered observability platform with Snowflake’s scalable and trusted data foundation, organizations can eliminate these tradeoffs and retain high-fidelity telemetry data, reducing observability cost substantially while improving visibility across their entire data estate..
…
Upcoming future Supermicro systems that incorporate the Nvidia Vera Rubin platform, including the new BlueField-4 data processor, will deliver a dramatic increase in inference and agentic AI-native performance across multiple nodes. For data centers, when using the Nvidia BlueField-4 data processor, the following results can be expected:
- Lower power usage for data transfer, allowing for more GPUs per data center
- Massive KV cache capacity for long-context reasoning
- High-speed pod-wide and efficient access to data
- Maximize GPU utilization
- Accelerate agentic AI serving, reducing time-to-first-token
For more information, visit: https://www.supermicro.com/en/accelerators/nvidia/vera-rubin .
…
