Intel announced a set of data center updates at Computex 2026 in Taipei spanning compute, networking, and its AI accelerator roadmap. The headline item is the introduction of Intel Xeon 6+ processors, paired with an expanded 800 Series Ethernet lineup based on Intel Ethernet E835 controllers and adapters. Intel also provided an update on its next data center GPU, code-named Crescent Island. Collectively, the announcements reflect Intel’s positioning around agentic AI, in which the CPU increasingly serves as the control plane for orchestration, concurrency, and data movement across clusters.

Intel data center leadership framed AI scaling as a coordinated systems problem rather than a component upgrade cycle. As AI agents become more autonomous and workloads become more distributed, Intel is emphasizing tight coupling between CPU, memory, and networking to reduce bottlenecks and improve efficiency under real power and rack constraints.

Xeon 6+ Extends the Xeon 6 Family with Density, Efficiency, and Scale-Out Focus

Xeon 6+ processors extend the Xeon 6 family, emphasizing performance density, power efficiency, and operational scale for cloud-native, network-intensive, and agentic AI-driven workloads. Intel said Xeon 6+ is built on Intel 18A, marking the company’s first use of that process node in a data center CPU. The platform’s focus is sustained throughput in environments where watts per rack, throughput per core, and predictable latency are primary constraints.

Intel’s stated feature set targets scale-out infrastructure that needs to add new AI-adjacent services without requiring a disruptive data center redesign. Configuration highlights include up to 288 efficient cores, which Intel says deliver up to 2.5 times more performance than the previous generation and up to 45 percent better per-thread performance per watt versus the competition, along with 12-channel DDR5 and 96 lanes of PCIe Gen5 with CXL support to move data across heterogeneous infrastructure. Intel also claims up to 9:1 server consolidation versus 2nd Gen Xeon systems. In addition, Intel is introducing Intel Application Energy Telemetry on Xeon 6+, providing real-time, workload-level telemetry of CPU energy and activity to improve visibility into consumption and utilization.

On security, Intel highlighted silicon-level protections, including Intel SGX and Intel TDX, aimed at confidential computing and multi-tenant deployments. Intel also said Xeon 6+ platforms are already being tested in telecom network infrastructures and configured into data center systems across the ecosystem, with servers, networking, and integrated solutions from ASUS, Dell Technologies, Ericsson, GIGABYTE, HPE, Lenovo, Supermicro, and others developing on Xeon 6+ today.

Intel Ethernet E835 Targets Power-Efficient 10GbE to 200GbE for AI and Virtualized Data Centers

Intel expanded its 800 Series Ethernet portfolio with Intel Ethernet E835 controllers and network adapters, positioning networking as a key limiter for modern AI, cloud, and distributed workloads. The E835 line targets dense, virtualized deployments where bandwidth and latency consistency must be delivered within tight power envelopes.

Intel said the E835 family supports up to 200GbE and a range of port configurations, including 2x25GbE, 4x25GbE, 2x100GbE, and 1x200GbE, with additional configurations enabled through Intel’s Ethernet Port Configuration Tool. For efficiency, Intel said the E835-CQDA2 adapter delivers up to 1.9 times higher performance per watt than the comparable NVIDIA ConnectX-6 Dx and 1.4 times higher than the Broadcom BCM957508-P2100G, positioning the product line as reducing energy consumption without sacrificing throughput.

On offloads and data path efficiency, E835 supports RDMA via RoCEv2 and iWARP, along with Dynamic Device Personalization to streamline packet processing. For security and manageability, Intel highlighted a hardware root of trust, signed SPDM, DMTF-based manageability, and OS support for Linux, VMware ESXi, and Windows. Intel also called out a 10+ year lifecycle target for long-term fleet standardization. Recommended pricing varies by configuration and is listed at intel.com/ethernet.

Xeon 6300 Adds a 12-Core Entry Option for SMB Servers

Intel also announced general availability of a new 12-core option in the Xeon 6300 family for entry servers, extending the platform beyond 8 cores. The key message is a drop-in upgrade path for existing entry-level server designs, allowing SMB environments to increase compute capacity without a platform change, and with availability through major OEMs.

Crescent Island: Intel Updates Next Data Center GPU for Inference and Token-Heavy Workloads

Intel provided an update on Crescent Island, its next-generation data center GPU built on the Xe 3P architecture. Intel is positioning the part around memory capacity, bandwidth, and efficiency as differentiators for agentic AI inference, particularly as token-intensive workloads grow.

According to Intel, Crescent Island pairs LPDDR5X with up to 480GB of capacity. It targets a 350W air-cooled PCIe form factor for scale-out deployments where rack-level thermals and power delivery are limiting factors. Intel also highlighted broad data-type support, ranging from native FP4 and MXFP4 to FP64, as well as expanded support for AI operations and scalability features. On software, Intel reiterated its commitment to an open, programmable stack designed to reduce friction in heterogeneous environments, with Arc Pro cited as a development platform aligned with the same Xe foundation for forward and backward compatibility.

Related: Supermicro Launches 12 Xeon 6+ Optimized Platforms, Expands X14 and DCBBS Offerings

In related news, Super Micro Computer announced 12 new server platforms optimized for Intel Xeon 6+ processors, spanning its Hyper, SuperBlade, FlexTwin, and GrandTwin families. Supermicro is emphasizing high core density and performance-per-watt for high-density cloud, virtualization, 5G analytics, and other throughput-intensive workloads. The new systems expand the company’s X14 lineup, with Supermicro positioning its DCBBS platform, a modular infrastructure approach built from validated components and subsystems, as the integration layer across the portfolio.

Supermicro’s updated portfolio spans multiple server families tuned for different densities, cooling, and deployment models, giving operators a clearer path to standardize on a common platform while still matching hardware to workload requirements. The Hyper series targets mainstream rack deployments with single- and dual-socket 1U and 2U systems that prioritize configurability. In practice, that means flexible CPU options, high-memory configurations for virtualization and data services, and the ability to pair the platform with advanced networking where east-west traffic and storage connectivity become performance limiters.

For environments where rack efficiency is the primary constraint, Supermicro’s SuperBlade and multi-node designs push density higher while maintaining serviceability. SuperBlade packages up to 10 compute nodes into a compact 6U chassis, using shared infrastructure to improve utilization at scale, a fit for large fleet deployments that benefit from simplified power and management domains. FlexTwin and GrandTwin take a multi-node approach in rack form factors, with FlexTwin emphasizing liquid-cooled, dual-socket nodes that operate independently while sharing power and cooling resources, and GrandTwin focusing on single-socket density optimized for high-core-count, E-core-heavy workloads where throughput per rack and thermal efficiency matter.

Supermicro said the new X14 platforms can scale up to 576 efficient cores per server in dual-socket configurations, improving deployment efficiency and energy consumption for large-scale data centers.

At the rack and data center level, Supermicro positions DCBBS as the integration layer that ties these systems together into a modular AI infrastructure. Built from validated components and subsystems, DCBBS is designed to reduce deployment friction by offering a repeatable building-block approach that scales from individual servers and networking to full rack-scale solutions, with supporting software and services for operators building out larger cloud and AI footprints.

Supermicro also cited architectural gains it attributes to Xeon 6+ systems, including double the core count, up to 17 percent higher IPC, five times more last-level cache, and 25 percent faster memory support compared to previous generations.

The post Intel Launches Xeon 6+ on 18A With 288 E-Cores, E835 200GbE Ethernet, and Crescent Island GPU Details appeared first on StorageReview.com.

NetApp and Cisco have introduced an expanded set of FlexPod-validated solutions to simplify the deployment of secure, scalable AI infrastructure. The announcement builds on the long-standing FlexPod partnership, positioning the platform as a pre-validated foundation for organizations looking to address the performance, data management, and security demands of modern AI workloads.

The companies are targeting enterprises that need predictable infrastructure outcomes without the integration overhead typically associated with AI environments. FlexPod continues to serve as a converged architecture that combines compute, networking, and storage, now extended with capabilities aligned with AI training and inference pipelines.

NetApp highlighted that AI workloads are placing increasing demands on data infrastructure as IT teams are tasked with delivering reliable, consistent performance across environments. Dallas Olson, Chief Commercial Officer at NetApp, noted that the FlexPod partnership has already saved customers up to 20% of their time in infrastructure management and maintenance, and that the companies are now applying their combined expertise to accelerate AI adoption while reducing risk with built-in security. Cisco reinforced the need to build security into AI infrastructure from the start, with Jeremy Foster, GM and SVP at Cisco, pointing to AI-specific risks such as data exposure, governance gaps, and compliance challenges.

Validated Architectures for AI Workloads

The updated FlexPod solutions are delivered as pre-tested reference architectures designed to support organizations at different stages of AI adoption. These configurations integrate NetApp data services, Cisco networking, and NVIDIA AI technologies to provide a consistent and scalable foundation.

For enterprise AI deployments, the architecture supports use cases such as retrieval-augmented generation and semantic search. The design reduces integration complexity by allowing AI capabilities to run directly where the data resides, with built-in, end-to-end security. NetApp AFX, the company’s disaggregated all-flash storage system, allows independent scaling of performance and capacity, which is critical for AI pipelines with uneven resource demands.

The solution also incorporates NetApp’s AI Data Engine, which is being developed to address data discovery, preparation, and governance challenges. This integrates with the NVIDIA AI Data Platform reference design, providing a structured approach to managing large enterprise datasets for AI use. Security is implemented across the stack through Zero Trust-aligned controls, with Cisco Secure AI Factory with NVIDIA providing policy-driven protections throughout the AI lifecycle.

On the networking side, Cisco AI networking infrastructure with Nexus One transforms the network into a deterministic, high-performance fabric that maximizes XPU utilization, reduces job completion times, and delivers predictable AI outcomes at scale. NetApp and Cisco also collaborated with NVIDIA to build FlexPod solutions based on NVIDIA Enterprise Reference Architectures, enabling organizations to design, deploy, and scale high-performance AI factories using a validated full-stack approach.

Simplified AI Inferencing and Departmental Adoption

In addition to full-scale deployments, NetApp and Cisco are targeting smaller teams and departmental use cases with pre-integrated solutions for AI inferencing and RAG workflows. These configurations are designed to reduce both cost and operational complexity, enabling organizations to leverage existing datasets without requiring specialized AI infrastructure expertise.

By standardizing deployment models, the companies aim to lower barriers to entry for AI adoption while maintaining enterprise-grade data management and security controls.

Extending AI to the Edge

The FlexPod expansion also includes validated architectures for edge environments, where AI inference and data processing must occur close to data sources. These solutions combine Cisco Unified Edge platforms with NetApp storage to support containerized and virtualized workloads in distributed locations.

The approach emphasizes centralized management and automation, allowing IT teams to deploy and operate AI infrastructure consistently across multiple sites. Policy-based configuration and orchestration enable repeatable deployments, reducing the operational overhead associated with managing isolated edge stacks.

Data Foundation and Ecosystem Integration

NVIDIA’s involvement centers on aligning data infrastructure with AI processing requirements. The companies are integrating NetApp’s data management capabilities with NVIDIA’s AI Data Platform to provide a unified, AI-ready data foundation. This includes support for data preparation, governance, and secure access, which remain key challenges for enterprises scaling AI initiatives.

The combined solution is validated within the Cisco Secure AI Factory framework, enabling organizations to deploy scalable AI environments on FlexPod with integrated security and data services. The goal is to provide a consistent architecture that supports both current AI workloads and future expansion without requiring significant redesign.

Expanded Cyber Resilience Collaboration with Splunk SOAR Integration

At the same time, NetApp and Cisco announced an expansion of their collaboration focused on cyber resilience and operational visibility. The expanded collaboration introduces deeper integration between NetApp storage and Splunk analytics and orchestration, strengthening defense-in-depth strategies at the data layer.

The companies are positioning intelligent data infrastructure as a core component of enterprise security, particularly as AI-driven threats increase in speed and sophistication. By combining observability, automation, and storage-level controls, the joint solution aims to reduce response times and limit the impact of cyber incidents.

Storage-Level Security Automation with Splunk SOAR

A key component of the announcement is the introduction of a NetApp Splunk Security Orchestration, Automation, and Response playbook. This integration extends Splunk’s existing visibility into NetApp environments by enabling automated response actions directly on ONTAP storage systems.

Splunk Enterprise Security is already integrated with NetApp Ransomware Resilience to collect analytics from the data layer, enhancing incident triage and prioritization. The new SOAR playbook builds on that foundation by allowing security teams to operationalize those insights. Automated actions can now be triggered by signals from NetApp Ransomware Resilience and other solutions in the environment, including blocking a suspicious user, taking snapshots of data, and taking data volumes offline to protect against further infection.

This approach shifts part of the incident response process closer to the data itself, reducing reliance on manual intervention and enabling faster containment. NetApp emphasized that integrating storage systems into security workflows helps reduce the blast radius of ransomware attacks, improves recovery times, and lowers overall remediation costs.

Enhancing Defense-in-Depth Strategies

NetApp and Cisco are aligning this integration with broader defense-in-depth strategies by connecting storage infrastructure into the security operations ecosystem. The solution combines NetApp’s data management and ransomware resilience features with Cisco’s secure AI infrastructure and Splunk’s analytics and orchestration capabilities.

NetApp noted that the rapid evolution of AI-enabled cyberattacks requires faster and more automated responses. Extending SOAR workflows to ONTAP enables organizations to take direct action on enterprise data during incident response, rather than treating storage as a passive layer. This enables more effective containment and reduces the window of exposure during an attack.

Cisco highlighted the importance of end-to-end visibility across the entire technology stack, including the data layer. By integrating NetApp storage into Splunk SOAR workflows, companies enable coordinated responses spanning networking, compute, and storage. This integration is intended to improve collaboration between security and storage teams while increasing confidence in automated response actions.

Operational Impact and Enterprise Readiness

The automation introduced by the NetApp Splunk SOAR playbook is expected to improve key security metrics, such as mean time to contain incidents, while reducing the manual effort and skills required to protect data. By embedding response capabilities directly into storage systems, organizations can respond more quickly without requiring specialized storage intervention during an incident. The playbook is available now for download from SplunkBase.

The post NetApp and Cisco Expand FlexPod With Validated AI Architectures and Splunk SOAR Storage Response appeared first on StorageReview.com.

Nutanix announced that its Nutanix Unified Storage (NUS) solution is now NVIDIA-Certified at the enterprise level, validating the platform for enterprise and cloud provider deployments running large-scale production AI workloads. The certification provides a validated configuration intended to reduce integration risk and help customers scale AI infrastructure with more predictable storage behavior.

The company also disclosed plans to extend its AI-native storage roadmap with support for NVIDIA Vera BlueField-4 STX. Nutanix framed this as an effort to improve data access and storage efficiency while simplifying operations as AI environments grow.

Addressing Storage as a GPU Utilization Constraint

As organizations build out “AI factory”-style infrastructure, Nutanix is targeting a recurring problem in production environments: GPU capacity is often limited by the ability to consistently feed data to accelerators. Fragmented infrastructure, siloed datasets, and inconsistent I/O can introduce bottlenecks that slow deployments and reduce effective GPU utilization. Nutanix is positioning NUS, along with its NVIDIA certification, to deliver a more consistent data path and reduce deployment variability across the stack.

Thomas Cornely, EVP of Product Management at Nutanix, emphasized that the goal is to remove infrastructure fragmentation and data silos so AI pipelines can sustain reliable throughput at scale. Jason Hardy, VP of Storage Technology at NVIDIA, similarly highlighted storage as a gating factor for enterprise AI, noting that certification provides customers with a more interoperable platform, reducing bottlenecks and improving GPU efficiency. Both sets of comments centered on interoperability, predictable scaling, and validated configurations rather than point performance claims.

Reference Architecture

Nutanix described the NVIDIA-Certified NUS reference architecture as being built on a 10-node all-NVMe cluster. On the protocol side, it uses enhanced parallel NFS (pNFS) and GPUDirect Storage over NFS with RDMA. The objective is a low-latency, high-throughput data path between GPU hosts and storage, designed to maintain resilience and minimize downtime as environments scale.

For the network fabric, Nutanix stated the design uses NVIDIA Spectrum-X Ethernet, including Spectrum-4 switches and BlueField-3 DPUs. Nutanix also provided scaling figures, claiming linear performance growth from 10GB/s read and 5GB/s write at 32 GPUs up to 160GB/s read and 80GB/s write at 1,024 GPUs.

Workload Coverage and Supported GPU Platforms

Nutanix positioned the architecture as a foundation for a range of AI workflows, including training, fine-tuning, inference, and RAG pipelines. The company also described broad compute compatibility, citing support for x86-based systems and multiple NVIDIA GPU configurations, including NVIDIA RTX 6000 PRO Blackwell, NVIDIA H200 NVL, NVIDIA HGX platforms with B200, H200, or H100 GPUs, and NVIDIA GH200 Grace Hopper Superchip configurations.

Availability

Nutanix said the NVIDIA-Certified Nutanix Unified Storage reference architecture is available now. Planned support for NVIDIA BlueField-4 STX is expected in the second half of 2026.

The post Nutanix Unified Storage Earns Enterprise-Level NVIDIA Certification for Production AI Workloads appeared first on StorageReview.com.

Although Dell’s PowerEdge R4715 and R5715 are two separate products, they should be viewed as a configurable matrix. The matrix includes two chassis, four AMD EPYC 9005 Series CPU options, a wide range of storage configurations, and the full Dell management and support ecosystem. The combination is purpose-built for SMB organizations that need to match infrastructure investment to actual workload requirements, and for the channel partners who help them get there.

Both platforms launched in March 2026. We covered each one individually in our R4715 and R5715 reviews. This piece is different. Rather than evaluating either server in isolation, we examine how the two platforms and the four CPU options perform across the workloads SMB buyers actually run and where configuration choices make the most difference.

Hypervisor flexibility is a key reason these platforms make sense right now. The virtualization market is evolving, and organizations of all sizes are reevaluating the assumptions underlying their infrastructure. Some are sticking with their established stack; others are migrating, and many are running two or more hypervisors in parallel for the foreseeable future. The R4715 and R5715 support the full range of common options, including VMware ESXi, Microsoft Hyper-V, Proxmox VE, and the major Linux KVM distributions, with consistent management and provisioning regardless of choice. For SMB customers who do not have the luxury of standardizing on a single platform, that flexibility is part of the value and one of the reasons we tested across multiple hypervisors for this piece.

The Dell Ecosystem Advantage for SMB and the Channel

The conversation around server platforms often centers on silicon, which makes sense at the spec sheet level. But for SMB buyers and the VARs and system integrators who support them, the operational experience with the silicon is often the deciding factor. Dell’s PowerEdge ecosystem is mature and well understood, and it manifests in ways that disproportionately benefit organizations with lean IT teams.

iDRAC10 and OpenManage Enterprise are the most visible components. The same management plane spans the entire 17th Generation PowerEdge family, so an SMB that buys an R4715 today can later add an R7725 or any other PowerEdge model without needing to learn a new toolset. For VARs and SIs supporting scores of customers, that consistency is even more valuable. A technician who knows iDRAC understands every PowerEdge customer’s infrastructure. The platform supports remote console access, firmware management, hardware health monitoring, and full Redfish API access for automation. For customers without dedicated infrastructure staff, that capability often means the difference between a phone call to a partner and a site visit.

Beneath the management layer, Dell brings a security and supply chain story that is hard to match. Silicon root of trust, cryptographically signed firmware, secured component verification, and TPM 2.0 with FIPS certification are all standard. ProSupport and ProDeploy services are available with global coverage, which matters for distributed SMBs and for partners selling into multiple geographies. Dell’s supply chain is one of the few in the industry that can deliver predictable lead times at scale. For VARs trying to close deals against backorder uncertainty, that is a competitive asset in itself.

For an SMB IT team of two or three people, or for a channel partner supporting dozens of customers with limited bench depth, the Dell ecosystem significantly reduces the operational surface area. The R4715 and R5715 inherit all of that.

R4715 and R5715 at a Glance

Here’s a brief recap for readers who have not seen our individual reviews. The R4715 is a 1U single-socket server optimized for compute density. It supports up to 24 DDR5 RDIMMs, three PCIe Gen5 slots, and a range of storage options, including 2.5-inch and 3.5-inch SAS/SATA configurations and an 8-bay 2.5-inch U.2 NVMe configuration. This is the right choice when rack density and per-rack-unit compute matter more than drive count.

The R5715 is a 2U single-socket server optimized for storage capacity and I/O expandability. It supports the same 24 DDR5 RDIMMs and the four CPU options. However, the R5715 adds a fourth PCIe Gen5 slot and steps up to either 12 bays of 3.5-inch SAS/SATA drives or 16 bays of 2.5-inch SAS/SATA drives. The 3.5-inch configuration can reach 288TB of raw capacity in a single node, which is the configuration we built our R5715 around for this article.

Both platforms are air-cooled, ship with iDRAC10, and support 800W and 1100W power supplies in Platinum or Titanium efficiency grades. These PowerEdge servers do not support GPUs, DPUs, or Fibre Channel, which aligns with Dell’s positioning of these servers as right-sized platforms with a specific target rather than maximum-flexibility platforms.

It is worth understanding where these two servers sit within Dell’s broader AMD-based PowerEdge lineup. The R4715 and R5715 are the value-optimized entry points, purpose-built for the SMB workloads covered in this article. Customers who need accelerators, higher core counts, or greater expansion have a clear path up the stack to the R6715 and R7715, which add GPU and DPU support, processors scaling well beyond 32 cores, and additional PCIe capacity for accelerator-driven and performance-intensive workloads. This tiering is an advantage for the channel; a VAR can place an R4715 or R5715 with a customer today and scale that customer up to the R6715 or R7715 as requirements grow, all within the same management plane, deployment workflow, and support model.

Platform Specifications

Specification	Dell PowerEdge R4715	Dell PowerEdge R5715
Processor
Processor	One 5th Generation AMD EPYC 9005 Series processor, up to 32 cores
Form Factor	1U rack server	2U rack server
Memory
DIMM Slots	24 DDR5 DIMM slots
Maximum Memory	1.5 TB (up to 64 GB per DIMM)
Memory Speed	Up to 5200 MT/s
Memory Type	Registered ECC DDR5 RDIMMs only
Storage
Internal Controllers (RAID)	PERC H365i, H965i
Internal Boot	BOSS-N1 DC-MHS
External HBAs	N/A
Front Drive Bays	4x 3.5-inch SAS 8x 2.5-inch SAS/SATA 8x U.2 NVMe Gen4	12x 3.5-inch SAS/SATA 16x 2.5-inch SAS/SATA
Power
Power Supplies	Platinum 800W, 1100W Titanium 800W, 1100W FTR supported
Cooling & Fans
Cooling Options	Air cooling
Fans	Up to four sets (dual fan module) hot-plug fans	Up to six hot-plug fans
Dimensions
Height	42.8 mm (1.68 inches)	86.8 mm (3.41 inches)
Width	482.0 mm (18.97 inches)
Depth (with bezel)	816.921 mm (32.16 inches)	802.4 mm (31.59 inches)
Depth (without bezel)	815.141 mm (32.09 inches)	801.51 mm (31.55 inches)
Bezel	Optional metal bezel

Four AMD EPYC 9005 Series CPU Options

Dell offers four specific CPU SKUs across both servers. The selection is deliberate, covering the range of SMB workloads without overlap or unnecessary complexity. Each CPU is built on AMD’s Zen 5 microarchitecture and shares the same platform-level memory and PCIe characteristics.

CPU	Cores	Default TDP	cTDP Range	Base Clock	Max Boost	L3 Cache
EPYC 9335	32	210W	200-240W	3.0 GHz	4.4 GHz	128 MB
EPYC 9255	24	200W	200-240W	3.2 GHz	4.3 GHz	128 MB
EPYC 9135	16	200W	200-240W	3.65 GHz	4.3 GHz	64 MB
EPYC 9015	8	125W	120-155W	3.6 GHz	4.1 GHz	64 MB

 

The 32-core 9335 is the top of the stack and the most flexible for compute-bound workloads. The 24-core 9255 is the closest match to the price-performance sweet spot we observed in our testing, particularly for database workloads, where marginal gains in core count slow after 24 cores. The 16-core 9135 and 8-core 9015 are where the value story is strongest. Because much of the software SMBs run is licensed per core, including Windows Server, many relational databases, and some hypervisor and backup platforms, the core count chosen at purchase carries forward as a recurring cost for the life of the deployment.

Selecting an 8-core or 16-core SKU that comfortably covers the workload, rather than over-provisioning cores that sit idle, reduces both acquisition cost and ongoing licensing exposure. The 16-core 9135 is especially relevant here, as it aligns with the minimum core count for Windows Server licensing, making it a natural starting point for Windows-centric environments that want to right-size the CPU to the license floor without leaving performance on the table. The 8-core 9015 is the lowest-power and lowest-cost option, ideal for storage-forward workloads or roles where the CPU is not the constraint. Each CPU runs at the same memory speeds and supports the same PCIe Gen5 lane count, so configuration decisions higher up the stack are not constrained by SKU choice.

Performance Testing

Testing Configurations	Dell PowerEdge R4715	Dell PowerEdge R5715
Tested CPU’s	AMD EPYC 9335, 9255, 9135, 9015	AMD EPYC 9015
Memory	384GB DDR5	384GB DDR5
Boot Storage	BOSS RAID1	BOSS RAID1
Front Storage Configuration	8x Samsung PM9D3a RI U.2 Gen5 NVMe SSDs (1.92TB) Raid 10 x 6	12x 20TB HDDs in RAID6

Database Performance: HammerDB MariaDB TPC-C

The headline workload for this evaluation is HammerDB running TPC-C against MariaDB 12.3.1. TPC-C is a long-established OLTP benchmark that produces measurable, comparable results across CPU and storage configurations and represents the kind of transactional database workload at the core of most SMB application stacks. We tested two distinct profiles: a CPU-intensive profile that stresses transaction processing and an I/O-intensive profile that places greater load on the storage subsystem. Both profiles were run across all four CPU options on the R4715 flash configuration to produce a clean CPU scaling curve, and then on the R5715 HDD configuration to show what changes when the storage substrate shifts.

Dell PowerEdge R4715

The HammerDB results show a clear scaling trend as core counts increase across the R4715 platform. Starting with the 8-core EPYC 9015, the system reached 480,818 NOPM in the CPU-intensive profile and 296,105 NOPM in the I/O-intensive profile before leveling off as the number of virtual users increased. Moving to the 16-core EPYC 9135 brought a substantial jump in throughput, pushing CPU-intensive performance to 737,445 NOPM and I/O-intensive performance to 493,093 NOPM, while also allowing the system to sustain higher virtual user counts before saturation.

The jump to the 24-core EPYC 9255 pushed the platform past the one million NOPM mark in the CPU-intensive profile, peaking at 1,017,429 NOPM, while the I/O-intensive profile climbed to 740,574 NOPM. At this point, the additional cores continued to translate directly into usable transactional throughput, while the NVMe storage subsystem kept pace with the growing database load.

At the top end, the 32-core EPYC 9335 delivered the highest results across both profiles, reaching 1,133,714 NOPM in the CPU-intensive workload and 910,321 NOPM in the I/O-intensive workload. The scaling curve remained relatively smooth even at high virtual user counts, indicating that the R4715 flash configuration effectively utilized the larger CPU configurations without storage bottlenecks prematurely limiting performance.

Dell PowerEdge R5715

Then we tested the Dell PowerEdge R5715, configured with 12×20TB HDDs in RAID 6, alongside the 8-core AMD EPYC 9015. In the CPU-intensive profile, the platform reached a peak of 484,715 NOPM at 16 virtual users, with throughput scaling cleanly as additional users were introduced before leveling off near saturation.

The I/O-intensive profile peaked at 308,012 NOPM with 24 virtual users, indicating solid transactional performance of the high-capacity HDD array under moderate concurrency. As the workload ramped, the scaling curve flattened as the spinning-disk subsystem approached its practical performance ceiling under sustained concurrent database activity.

Windows Server Shared Storage

The second practical use case targets a different SMB workload pattern: Windows-based shared storage. For organizations running file shares, departmental applications, or general-purpose Windows Server roles, these platforms can meet modern SMB performance expectations. We ran FIO on Windows Server to characterize sequential and random performance across two storage configurations: a RAID 6 HDD array on the R5715 as a baseline, and an 8-drive SSD JBOD array on the R4715 representing a high-performance storage configuration. The comparison illustrates the operational gap between the two storage tiers in this platform class.

The gap between the two storage configurations becomes immediately apparent in the FIO results. While the RAID6 HDD array in the R5715 delivered respectable sequential throughput for a high-capacity spinning-disk platform, the SSD-equipped R4715 operated in a completely different performance class, particularly in random workloads, where SMB environments tend to feel storage latency most.

Sequential performance on the HDD array reached up to 3.7GB/s writes and 2.2GB/s reads in the 4-thread tests, which is more than adequate for traditional file serving, backups, and bulk storage tasks. However, the SSD configuration pushed sequential throughput into the tens of gigabytes per second, exceeding 56GB/s reads and 26GB/s writes while maintaining dramatically lower latency.

The separation widened even further in 4K random workloads. The HDD array peaked at under 1,300 IOPS in random-write testing, with latency climbing above 100ms, whereas the SSD configuration delivered over 4 million IOPS with sub-millisecond latency. In practical terms, this translates directly into application responsiveness, multi-user file share performance, VM storage behavior, and the ability to sustain concurrent SMB workloads without the storage layer becoming a bottleneck.

FIO Workload	R5715 RAID6 HDD 1T	R5715 RAID6 HDD 4T	R4715 8x SSD 1T	R4715 8x SSD 4T
Sequential Read (128K)
Bandwidth	1,475.89 MB/s	2,198.89 MB/s	56,861.09 MB/s	56,866.76 MB/s
IOPS	11,807	17,589	454,885	454,918
Latency	2.70ms	7.28ms	0.56ms	2.25ms
Sequential Write (128K)
Bandwidth	2,665.31 MB/s	3,726.63 MB/s	26,739.52 MB/s	26,753.48 MB/s
IOPS	21,322	29,811	213,912	214,011
Latency	1.49ms	4.39ms	1.20ms	4.78ms
Random Read (4K)
Bandwidth	1.00 MB/s	3.60 MB/s	7,268.78 MB/s	16,143.19 MB/s
IOPS	256	919	1,860,803	4,132,645
Latency	125.11ms	139.00ms	0.13ms	0.17ms
Random Write (4K)
Bandwidth	4.88 MB/s	4.67 MB/s	7,555.53 MB/s	16,010.24 MB/s
IOPS	1,248	1,195	1,934,214	4,098,613
Latency	25.63ms	106.98ms	0.07ms	0.13ms

Proxmox Backup Server

Beyond raw benchmarks, the R5715 with HDD storage is exactly the kind of platform that suits a virtualized backup target workload. To validate that fit, we deployed Proxmox Backup Server on the R5715 configured with the 8-core EPYC 9015 and the same 12-bay 3.5-inch HDD array. Proxmox is a representative example of the broader open-source hypervisor and infrastructure ecosystem that has gained significant traction in SMB environments, and Proxmox Backup Server, in particular, is well-suited to this server’s storage profile.

We deployed Proxmox Backup Server 4.2.0 and used it to back up the virtual machines that power our Proxmox community Discord server environment.

In our specific configuration, backup and restore operations were somewhat limited by the system’s 1GbE networking connection, which became the primary bottleneck during larger transfers. However, the platform supports straightforward networking upgrades via OCP expansion cards, making it easy to migrate to 10GbE or even 25GbE connectivity. With faster networking in place, the R5715 would be able to handle significantly higher backup throughput and restore performance, especially in environments with larger VM datasets or more demanding backup windows.

Conclusion

The Dell PowerEdge R4715 and R5715 succeed by getting the configuration matrix right. Two chassis with clearly differentiated form factors, four CPU options that cover the practical range of SMB workloads without overlap, and a storage menu broad enough to support everything from low-cost bulk capacity to all-flash performance. The configuration flexibility is not theoretical. Across the workloads we tested, the right answer changed with each one. The 24-core 9255 hit the sweet spot for transactional database performance on flash. The 8-core 9015 provided exactly enough CPU for a Proxmox Backup Server deployment with bulk HDDs. The R4715 with flash storage was the appropriate choice for Windows shared storage, while the R5715 with HDDs was the appropriate choice for capacity-focused workloads where peak I/O is not the constraint.

For SMB customers and the partners who serve them, the value of these platforms lies in the ability to match the build to the workload, rather than buying more capacity than the workload requires. The Dell ecosystem, including iDRAC10 management, ProSupport, the security stack, and supply chain predictability, amplifies that value at the operational level. Lean IT teams and channel partners both benefit when the platform behind the customer’s workload is a known quantity.

Dell’s positioning of these servers as a way to consolidate legacy infrastructure and reduce per-socket and per-core licensing exposure holds up against the data we collected. With four CPU options spanning 8 to 32 cores at the same platform level, customers and partners have the flexibility to right-size acquisition, licensing, and operational costs to match actual requirements. That is the value proposition, and the R4715 and R5715 deliver against it.

Dell PowerEdge R5715 Product Page

Dell PowerEdge R4715 Product Page

The post From Database and Virtualized Workloads to Backup: Dell PowerEdge R4715 and R5715 for SMB Realities appeared first on StorageReview.com.

I just can’t wrap my head around why KIOXIA hasn’t brought its consumer SSDs to North America. And, in the same vein, I still don’t fully understand why Micron retired the much-loved Crucial brand. That said, I’m convinced KIOXIA-branded SSDs would do extremely well here, potentially rivaling those top-tier bestsellers we all know. In the …

The post KIOXIA Shows Off Exceria Pro G2 Gen5 SSD Along with Enterprise Options | Computex 2026 Update appeared first on The SSD Review.

ZutaCore has secured $100 million in Series C funding, with participation from Mitsubishi Electric, Carrier Ventures, Samsung Electronics, and others through its corporate venture arm, Samsung Ventures. The investment is aimed at accelerating global commercialization, expanding deployments, and advancing research and development as demand for high-density AI and HPC infrastructure continues to rise.

The funding comes as data center operators face increasing thermal challenges driven by next-generation processors that are pushing well beyond traditional power envelopes. Liquid cooling adoption has accelerated across hyperscale and enterprise environments, a trend StorageReview has covered extensively as operators shift from air to direct liquid cooling to manage higher rack densities and improve efficiency.

Two-Phase Cooling Targets Next-Gen Power Levels

ZutaCore’s platform focuses on waterless two-phase direct-to-chip cooling, designed to support processors exceeding 4,000W. This approach uses phase-change heat transfer at the chip level to remove heat more efficiently than traditional air or single-phase liquid cooling.

The company is positioning its technology to integrate alongside existing air and single-phase liquid systems, enabling incremental deployment within current data center designs. This hybrid compatibility is increasingly important as operators adopt liquid cooling in stages rather than full facility retrofits.

ZutaCore reports more than 75 deployments across the Americas, Europe, and Asia, reflecting growing production adoption of two-phase cooling in AI and HPC environments.

Investment Supports Scale and Product Development

The Series C funding will be used to expand global operations and address increasing customer demand. It will also support ongoing R&D focused on in-package thermal management and system-level integration for megawatt-scale deployments.

As AI clusters scale into multi-megawatt configurations, cooling infrastructure must evolve to maintain performance and reliability. ZutaCore is targeting these requirements with thermal management designs that extend from the chip package itself to full megawatt-class system deployments.

The company also highlighted continued collaboration with ecosystem partners to align cooling solutions with emerging chip roadmaps and accelerate deployment timelines.

Validation at Megawatt Scale

To support scaling efforts, ZutaCore has expanded its executive team with four key hires: Yaniv Reinhold as Chief Financial Officer, Sharon Shafran as Chief Operating Officer, Yoni Nir as Chief Research and Development Officer, and Sarah Warshavsky Oberman as Chief People Officer. The additions bring experience in global finance, semiconductor technologies, and large-scale system deployment, aligning with the company’s focus on hyperscalers, neoclouds, and demanding enterprise compute environments.

This type of pre-deployment validation is becoming more critical as liquid-cooled AI infrastructure increases in complexity and cost.

Expanding Product Portfolio

ZutaCore continues to extend its product portfolio, including the OmniTherm cold plate designed for NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs. The design enables waterless two-phase cooling within a single-slot PCIe form factor, supporting full-power operation in standard enterprise and AI server configurations.

This reflects a broader industry shift toward component-level liquid cooling solutions that can be deployed in conventional server architectures while still delivering the thermal performance required for modern accelerators.

Leadership Expansion to Support Growth

To support scaling efforts, ZutaCore has expanded its executive team with hires across finance, operations, R&D, and human resources. The additions bring experience in global operations, semiconductor technologies, and large-scale system deployment, aligning with the company’s focus on expanding into hyperscale, neocloud, and enterprise markets.

Industry Context

The funding round underscores growing momentum behind liquid cooling technologies as AI workloads reshape data center design. StorageReview has observed increasing adoption of both single-phase and two-phase cooling approaches, with vendors aligning solutions to support higher power densities, warm-water operation, and improved energy efficiency.

ZutaCore’s latest funding and deployment activity reflect the next phase of this transition, where cooling is no longer a supporting function but a primary enabler of compute scalability.

The post ZutaCore Raises $100M Series C to Scale Waterless Two-Phase Cooling for AI Data Centers appeared first on StorageReview.com.

CoolIT Systems has announced the development of what it describes as the first 15kW direct liquid cooling (DLC) coldplate design, positioning single-phase liquid cooling as a viable path for AI infrastructure well beyond 2030. The company reports that the design delivers nearly four times the performance of earlier single-phase coldplates, indicating that the architecture can scale alongside increasing GPU and accelerator power levels.

Single-phase DLC is already widely deployed across AI data centers, particularly in hyperscale environments. This latest development focuses on extending that model to support significantly higher thermal design power targets without requiring a transition to more complex cooling approaches.

Scaling Thermal Capacity for AI Accelerators

The 15kW coldplate represents a substantial increase in cooling capacity compared to prior designs. CoolIT states that the new coldplate delivers nearly 4x the capacity of the 4kW design it announced in March 2025 and more than 10x the cooling required for current-generation AI GPUs.

This level of thermal headroom is increasingly relevant as next-generation AI accelerators push beyond traditional power envelopes. Higher per-device wattage combined with dense system packaging is driving the need for more efficient heat removal at the component level.

Microchannel Design and Warm-Water Operation

The coldplate is based on CoolIT’s Split-Flow microchannel architecture, designed to optimize heat transfer across high-power silicon. Validation was performed using a standard water-glycol coolant at a flow rate of 1.2 L/min/kW.

The system is designed to operate in 45°C warm-water environments, aligning with broader industry trends toward higher coolant temperatures to improve overall data center efficiency. Warm-water cooling reduces reliance on mechanical chillers and enables more efficient heat reuse strategies.

Alignment with Industry Direction

The announcement reflects broader momentum behind single-phase DLC as a standard approach for AI infrastructure. NVIDIA has indicated support for single-phase liquid cooling operating at elevated supply temperatures in its platform roadmap, reinforcing the relevance of warm-water-compatible coldplate designs.

By demonstrating performance at 15kW, CoolIT is positioning single-phase DLC as capable of supporting both current deployments and future accelerator generations without requiring architectural changes to cooling systems.

Expanding Cooling Beyond the GPU

In parallel with the coldplate development, CoolIT is working to extend liquid-cooling coverage to additional server components. This includes targeting peripheral devices and addressing localized hot spots within advanced AI processors.

These efforts aim to increase total heat capture at the system level and improve thermal consistency across increasingly complex AI server designs. As power densities rise, comprehensive cooling strategies that go beyond the primary compute die are becoming necessary to maintain performance and reliability.

The post CoolIT Systems Demonstrates 15kW Coldplate, Extending Single-Phase DLC Beyond 2030 appeared first on StorageReview.com.

If Alphabet's record-breaking $85 billion stock sale signals investor appetite for AI-related offerings, we can see that investors are ready to chow.

The company is reducing its workforce as it exits 22 countries, reduces management layers, and invests in its infrastructure to scale its platform.

Financial services infrastructure continues to be defined by the need to process larger risk models within fixed power and space constraints. As Monte Carlo-based analytics scale, system bottlenecks increasingly shift from compute to memory bandwidth, where data movement dominates runtime. Monte Carlo methods use repeated random sampling to calculate the probability of different outcomes in complex systems, quantifying risk across thousands of scenarios under high uncertainty.

A newly audited STAC-A2 benchmark result demonstrates the impact of pairing high-core-count processors with next-generation memory. An HPE ProLiant Compute XD230 1U server, configured with Intel Xeon 6980P processors and Micron 8800 MT/s DDR5 MRDIMMs, achieved the highest recorded performance for STAC-A2 cold runs at the baseline problem size, with tests run in HPE’s labs and independently audited by STAC on May 5, 2026. Results were generated using an Intel-optimized STAC-A2 implementation.

STAC-A2 as a Measure of Real-World Risk Workloads

STAC-A2 is widely used across capital markets to evaluate infrastructure performance for Monte Carlo simulation of option Greeks. These calculations underpin derivatives pricing, hedging strategies, and regulatory capital requirements. The benchmark reflects production-like workloads, emphasizing parallelism, memory bandwidth utilization, and latency sensitivity.

For trading firms and banks, incremental improvements in throughput or latency directly reduce end-of-day batch windows and enable more complex scenario analysis within existing operational timelines.

Record Performance and Generational Gains

In testing, the XD230 configuration delivered leading results across throughput, energy efficiency, and space efficiency among audited STAC-A2 submissions. Compared to a prior-generation platform using Intel Xeon Platinum 8592+ processors, the system demonstrated significant gains.

The STAC-A2 tests conducted by HPE in its labs achieved the highest throughput, energy efficiency, and space efficiency among all audited and tested STAC-A2 servers.

Benchmark	Result	What It Measures
Portfolio Throughput	100.8 options/sec	Options priced per second across a portfolio
Energy Efficiency	231,271 options/kWh	Options priced per kilowatt-hour consumed
Space Efficiency	133.8 options/hr/in cubed	Options priced per hour per cubic inch of server
Baseline Greeks (cold)	0.033 seconds	Time to compute all Greeks, baseline problem size
Max Assets	160	Max assets completed in 10 min (25K paths, 252 timesteps)
Max Paths	1,000,000	Max paths completed in 10 min (5 assets, 252 timesteps)

 

Portfolio throughput improved by 2.38x, while cold and warm runs of the baseline problem size were 10.42x and 1.62x faster, respectively. Large-problem-size runs saw improvements of 2.04x and 2.07x. Efficiency also improved, with 1.58x higher energy efficiency and 3.26x higher space efficiency.

These results indicate a substantial generational improvement, enabling higher compute density and faster time-to-insight without expanding infrastructure.

MRDIMMs Drive Memory Bandwidth Scaling

The system configuration included 24 x 64GB Micron DDR5 MRDIMMs operating at 8800 MT/s, providing up to 1.5TB of memory across 12 channels per socket. This memory subsystem is designed to meet the bandwidth demands of 256 total cores.

Monte Carlo workloads continuously move large datasets through memory during path generation, correlation, and regression steps. In this context, memory bandwidth directly affects both throughput and latency.

Compared to a similar Xeon 6980P configuration using RDIMMs, the MRDIMM-based system showed measurable gains. Portfolio throughput increased modestly from 93.2 to 100.8 options per second, while Greeks calculations improved by up to 23% on large problem sizes. Energy efficiency reached 231,271 options per kWh, up from 178,172, and space efficiency improved to 133.8 options per hour per cubic inch, up from 80.7.

These results reinforce that higher memory bandwidth can translate into tangible performance and efficiency improvements for data-intensive financial workloads.

Intel Xeon 6980P and Software Optimization

The Intel Xeon 6980P processor provides up to 128 performance cores per socket, along with 504MB of L3 cache and AVX-512 support for vectorized workloads. These features are aligned with the compute characteristics of Monte Carlo simulations.

The benchmark used the STAC-A2 Pack for oneAPI (Rev R) with the Intel oneAPI Base Toolkit and HPC Toolkit, version 2025.3. Intel has iterated on STAC-A2 implementations for over a decade, and this result reflects continued optimization at both the hardware and software levels.

Liquid Cooling Enables 1U Density

Achieving sustained performance from dual 128-core processors in a 1U chassis required liquid cooling. In this configuration, CPUs were connected to a coolant distribution unit (CDU) tied into the facility water loop, while other components remained air-cooled.

This hybrid approach enabled high compute density without thermal throttling. Compared to an air-cooled Xeon 6980P configuration, the liquid-cooled system delivered 1.23x better energy efficiency. It also achieved the highest reported efficiency for an Intel Xeon 6 system at 231,271 options per kWh, along with 65.8% better space efficiency than the next-best Xeon 6980P result.

Implications for Financial Data Centers

The benchmark highlights three areas of impact for financial institutions. Improved energy efficiency allows more calculations within fixed power budgets. Higher throughput reduces the time required for end-of-day and intraday risk runs. Increased density enables scaling within existing rack footprints.

These factors are particularly relevant for organizations operating under colocation constraints, power caps, or sustainability targets. The combination of MRDIMMs, high-core-count CPUs, and liquid-cooled 1U systems offers a path to expand analytics capacity without facility expansion.

Ecosystem Collaboration

The result reflects coordinated engineering across vendors. Intel provided the optimized STAC-A2 implementation, Micron supplied high-speed MRDIMMs to address bandwidth constraints, and HPE delivered the dense, liquid-cooled platform. STAC independently audited all results.

This submission builds on prior STAC-A2 work, where Micron memory helped deliver 35.2-millisecond results in an earlier collaboration. The current results extend those gains into throughput, efficiency, and density.

Full audited results are available from STAC under SUT ID INTC260430.

The post HPE XD230 STAC-A2 Record: Intel Xeon 6980P and Micron MRDIMMs Lead Financial Risk Benchmarks appeared first on StorageReview.com.

At COMPUTEX 2026, MSI showcased a broad range of AI and data center infrastructure platforms, with a clear emphasis on liquid cooling as rack power density continues to climb. The company’s exhibit featured OCP ORv3 rack-scale architectures, NVIDIA MGX-based GPU servers for training and inference, an NVIDIA DGX Station-derived desktop system for local AI work, and DC-MHS multi-node and enterprise platforms for modular cloud and enterprise deployments.

MSI Enterprise Platform Solutions General Manager Danny Hsu framed the portfolio around a practical challenge many operators face. “Scaling AI infrastructure now requires a balance between compute performance, thermal efficiency, and deployment flexibility,” Hsu said. MSI positioned its lineup to span everything from high-density rack-scale rollouts to deskside AI development environments.

Liquid-cooled ORv3 and air-cooled EIA rack architectures

MSI expanded its rack infrastructure options to include both OCP ORv3 liquid-cooled designs and standard 19-inch EIA air-cooled racks. The ORv3 approach targets high-density AI and cloud deployments. At the same time, the EIA option is intended to fit within existing enterprise data center standards without requiring a shift to new rack dimensions.

The 21-inch 44OU ORv3 liquid-cooled rack architecture is designed for deployments up to 100kW and includes an integrated liquid-to-liquid coolant distribution unit (L2L CDU). MSI demonstrated it configured with 28 of its 1OU2N Open Compute multi-node systems. MSI also emphasized the ORv3 rack’s use of 48V busbar power distribution as part of an overall design aimed at higher compute density with liquid cooling.

For environments that remain standardized on traditional racks, MSI also showcased a 19-inch, 48RU EIA air-cooled rack architecture. In this configuration, the rack supports 16 2U2N multi-node systems and offers both AMD EPYC 9005 and Intel Xeon 6 platform options, aligning the design with mixed enterprise and cloud deployment needs.

NVIDIA MGX GPU servers and a DGX Station-class desktop platform

On the GPU side, MSI’s portfolio centers on NVIDIA MGX, which it is using as the basis for a family of 2U, 4U, and 6U servers targeting AI training, inference, HPC, and data-intensive workloads across air-cooled and liquid-cooled configurations. MSI stated support for NVIDIA H200 NVL, NVIDIA RTX PRO 6000, and NVIDIA RTX PRO 4500 Blackwell Server Edition GPUs, and indicated ongoing work within the MGX ecosystem toward next-generation NVIDIA Vera Rubin rack-scale platforms.

MSI highlighted several specific MGX server models. The CG681-S6093 is a liquid-cooled 6U dual-socket AMD EPYC system designed to scale to eight NVIDIA RTX PRO 6000 Blackwell Server Liquid Cooled Edition GPUs, with 32 DDR5 DIMMs and NVIDIA ConnectX-8 SuperNIC connectivity supporting up to 8x 400Gbps networking. The CG480-S5063 is a 4U dual-socket Intel Xeon 6 platform that supports up to eight double-wide GPUs, 32 DDR5 DIMMs, up to 20 E1.S NVMe drives, and five additional PCIe 5.0 expansion slots for storage-intensive AI and HPC configurations.

MSI also detailed two 4U dual-socket AMD EPYC 9005 options, the CG481-S6053 and CG480-S6053, each designed to support up to eight double-wide GPUs, 24 DDR5 DIMMs, and eight U.2 NVMe drives. MSI differentiated the two by highlighting high-bandwidth networking on the CG481-S6053, including up to 8x400G QSFP112 via NVIDIA ConnectX-8 SuperNICs, while positioning the CG480-S6053 with five additional PCIe 5.0 expansion slots for added flexibility.

Rounding out the MGX server highlights, MSI described the CG290-S3063 as a 2U, single-socket Intel Xeon 6 system that supports up to four double-wide GPUs, 16 DDR5 DIMMs, and rear U.2 NVMe storage, targeting inference, edge AI, and space-constrained data center deployments.

For local AI development and fine-tuning, MSI showcased its XpertStation WS300, a desktop platform based on the NVIDIA DGX Station architecture. MSI said the system is powered by the NVIDIA GB300 Grace Blackwell Ultra Desktop Superchip and supports up to 748GB of coherent memory and 7.1TB/s of HBM3e bandwidth, targeting high-throughput CPU-GPU communication and larger model workflows on-prem. MSI also specified dual 400GbE networking via NVIDIA ConnectX-8 SuperNICs and a liquid-cooled thermal design to sustain higher performance in a deskside form factor, with Windows support positioned for local AI application and agent development.

DC-MHS multi-node platforms for Open Compute and 19-inch environments

MSI also embraced modularity with 21-inch Open Compute and 19-inch Core Compute multi-node platforms based on the DC-MHS architecture, targeting hyperscale and cloud operators seeking simpler platform transitions and consistent building blocks across deployments.

Within the 21-inch Open Compute lineup, MSI described air- and liquid-cooled 1OU2N, 2OU2N, and 2OU4N designs optimized for density and 48Vdc busbar power distribution. Two examples were highlighted. The CD281-S4051-X4 is a liquid-cooled 2OU 4-node platform, with each node supporting a single AMD EPYC 9005 processor, 12 DDR5 DIMMs, and four E1.S NVMe bays, aimed at inference and cloud-native infrastructure. The CD281-S4051-X2 is a 2OU 2-node platform, also based on AMD EPYC 9005, with each node configured with 12 DDR5 DIMMs, 12 E3.S NVMe bays, and dual full-height, half-length PCIe 5.0 slots, positioning it for storage-rich scale-out designs.

For 19-inch rack deployments, MSI’s Core Compute portfolio includes 2U2N and 2U4N systems spanning Intel Xeon 6 and AMD EPYC 9005 options. MSI cited the CD270-S3071-X4 and CD270-S3071-X2 as Intel Xeon 6 or 6+ platforms with 12 DDR5 DIMMs per node, supporting Xeon 6+ configurations up to 288 E-cores. The X4 variant carries three U.2 NVMe bays per node for compute-focused deployments, while the X2 doubles that to six bays per node for virtualization and data-centric services. MSI also listed the CD270-S3061-X4 as a 2U 4-node Intel Xeon 6 platform with 16 DDR5 DIMMs and three U.2 NVMe bays per node for mainstream scale-out and containerized use. On the AMD side, MSI described the CD270-S4051-X4 and CD270-S4051-X2 as EPYC 9005-based multi-node systems with 12 DDR5 DIMMs per node, with the X2 again shifting toward mixed compute and storage at six U.2 NVMe bays per node versus three on the X4.

Enterprise servers, modular HPMs, and motherboards

Beyond multi-node, MSI detailed a set of DC-MHS enterprise servers and modular HPMs intended to bridge modular cloud infrastructure and traditional enterprise requirements, including GPU-ready configurations.

On DC-MHS enterprise servers, MSI listed dual-socket Intel Xeon 6 options, including the CX270-S5062 (2U) and CX170-S5062 (1U), each supporting 32 DDR5 DIMMs. The 2U system pairs eight U.2 NVMe drives with support for up to two double-wide 600W GPUs, while the 1U model steps up to 12 U.2 NVMe drives for high-density cloud infrastructure. MSI also outlined single-socket AMD EPYC 9005 systems, the CX271-S4056 (2U) and CX171-S4056 (1U), with up to 24 DDR5 DIMMs, and a separate single-socket Intel Xeon 6 pair, the CX271-S3066 (2U) and CX171-S3066 (1U), supporting up to 16 DDR5 DIMMs. Across both single-socket families, the pattern holds: the 2U models support up to two double-wide 600W GPUs and eight U.2 drives, while the 1U models carry 12 U.2 drives for scale-out deployments.

On the modular side, MSI referenced DC-MHS HPM support across the M-DNO Type-2, M-DNO Type-4, and M-FLW form factors, listing specific modules for Intel Xeon 6 and AMD EPYC 9005 platforms. MSI also cited standard enterprise motherboards for Intel Xeon 6 (D3060) and AMD EPYC 9005 (D4050), indicating continued coverage for mainstream enterprise and workstation server builds outside DC-MHS deployments.

The post MSI Brings Liquid-Cooled ORv3 Racks, NVIDIA MGX Servers, and a DGX Station Desktop to COMPUTEX 2026 appeared first on StorageReview.com.

IBM and Red Hat have expanded their enterprise security strategy with two announcements in quick succession: joining Anthropic’s Project Glasswing on May 19 and introducing Project Lightwell on May 28, aligning large-scale engineering investment with industry collaboration to address growing risks in open-source and AI-driven environments.

IBM and Red Hat Launch Project Lightwell

Project Lightwell represents a $5 billion commitment focused on securing open source software across its full lifecycle, from upstream development through enterprise production. The initiative combines advanced AI capabilities with a global workforce of more than 20,000 engineers to create a coordinated model for identifying, validating, and remediating vulnerabilities at scale.

At the center of the effort is a trusted security clearinghouse that serves as an intermediary between enterprises and the open-source ecosystem. The clearinghouse is designed to ingest vulnerability data from real-world deployments, apply AI-assisted validation and testing, and deliver production-ready patches through commercial subscription services. These patches are intended to integrate directly into enterprise software supply chains with lifecycle management and enterprise-grade assurance.

IBM and Red Hat are extending their established enterprise open-source model beyond curated platform components to include independent libraries, language toolchains, AI frameworks, and data streaming platforms. IBM already uses more than 62,000 open source packages and maintains deep expertise in over 10,000, spanning technologies such as Linux, Java, Kubernetes, Kafka, Ansible, and Terraform. The expansion reflects the operational reality that enterprises rely on a broad mix of community-driven software outside vendor-managed distributions.

The clearinghouse enables organizations to report vulnerabilities discovered in active environments within a controlled framework, receive validated patches optimized for production use, and coordinate responsible disclosure upstream to maintainers. This approach is designed to reduce fragmentation in vulnerability handling while reinforcing long-term ecosystem stability through upstream contributions.

The initiative comes as open-source software continues to underpin enterprise infrastructure, with more than 90 percent of Fortune 500 companies relying on OSS, and as advances in AI accelerate both vulnerability discovery and exploitation. Anthropic recently reported that its Mythos Preview model identified nearly 3,900 high- or critical-severity vulnerabilities in open source software alone. IBM and Red Hat are positioning Project Lightwell as a response to this shift, applying AI and engineering scale to compress remediation timelines and improve consistency across complex software supply chains.

AI-Assisted Engineering Model

A core component of Project Lightwell is the deployment of a large-scale engineering organization augmented by AI-driven tooling. IBM and Red Hat are emphasizing technical capacity as a strategic asset, with engineers operating across both upstream communities and enterprise environments.

The engineering teams will contribute to upstream maintenance alongside project maintainers and handle enterprise-specific requirements, such as vulnerability triage, prioritization, and validation. AI is used to support high-volume analysis, secure patch development, dependency hardening, and release engineering.

This model addresses a key challenge for enterprises managing diverse open-source dependencies, where vulnerability management is often fragmented and resource-intensive. By combining AI-assisted workflows with dedicated engineering resources, IBM and Red Hat aim to standardize and scale remediation processes.

Early adopters are concentrated in financial services: Bank of America, BNY, Citi, Goldman Sachs, JPMorganChase, Mastercard, Morgan Stanley, Royal Bank of Canada, State Street, Visa, and Wells Fargo are already collaborating with IBM and Red Hat on Project Lightwell. IBM says insights from these initial deployments will shape how vulnerabilities are identified, validated, and remediated at scale in large, regulated environments.

Integration with IBM’s AI Security Portfolio

Project Lightwell aligns with IBM’s broader security portfolio for the AI era, outlined in the company’s May 19 announcement. IBM Concert aggregates application, infrastructure, and network signals into a unified operational view, with the goal of moving organizations from passive monitoring to coordinated response. Its security capabilities extend into the developer’s IDE through IBM Concert Secure Coder, which detects and prioritizes risks by business impact and generates automatic remediations as code is written, preventing vulnerabilities from reaching production.

IBM Consulting is also advancing Autonomous Security services, which use multi-agent systems to coordinate detection, decision-making, and response at machine speed. These services are intended to help enterprises adapt vulnerability management and open source governance to increasingly compressed timelines driven by AI-enabled threats.

IBM Joins Project Glasswing

Shortly before the Lightwell announcement, IBM joined Project Glasswing, the Anthropic-led industry initiative to defend critical software infrastructure. The coalition brings together security and technology leaders to identify and remediate vulnerabilities in widely used software and share findings across industries. The Lightwell release notes that the new initiative incorporates lessons from Glasswing and OpenAI’s Trust Access for Cyber program.

As part of Glasswing, IBM has been identifying and remediating vulnerabilities in widely used software and sharing those findings with the broader community. Rob Thomas, SVP Software and Chief Commercial Officer at IBM, said the company has been “hardening our own products and contributing fixes back to the open-source community,” adding that “the collaboration makes the entire ecosystem stronger.”

A central component of this work is coordinated disclosure. IBM shares findings with affected vendors and maintainers in accordance with established disclosure practices, enabling patches to be developed and validated before public release. The company also contributes fixes directly to upstream projects, ensuring that remediations are incorporated into future releases and supported branches, reducing divergence between enterprise deployments and community codebases.

IBM is also applying its Glasswing work to its own portfolio. The company says that by contributing fixes proactively and maintaining enterprise-grade versions of widely used open-source components, IBM and Red Hat can move quickly when issues arise, pairing the flexibility of open source with reliable, rapid support.

The Glasswing collaboration model emphasizes shared learning across participants. IBM contributes findings through coordinated disclosure, upstream open-source patches, and best practices shared with fellow participants, reflecting the company’s position that openness and scrutiny are prerequisites for security at scale. This cross-industry visibility is increasingly important as AI lowers the barrier to discovering and exploiting vulnerabilities.

Positioning for AI-Driven Threat Environments

Together, Project Lightwell and IBM’s participation in Project Glasswing reflect a shift toward more coordinated, AI-assisted security models. IBM is combining internal engineering scale, ecosystem collaboration, and AI-driven tooling to address the growing complexity and speed of modern threat environments.

The approach focuses on securing open source software at its source while improving how vulnerabilities are identified, validated, and remediated across enterprise supply chains. By linking upstream engagement with production-grade validation and industry-wide intelligence sharing, IBM and Red Hat are positioning these initiatives as foundational elements of enterprise security in the AI era.

The post IBM and Red Hat Launch $5B Project Lightwell, Join Anthropic’s Project Glasswing appeared first on StorageReview.com.

OpenAI released a set of six plug-ins aimed at specific jobs: data analytics, creative production, sales, product design, equity investing, and investment banking. Available from within the Codex app, each of the new tools bundles integrations, instructions, and context to allow Codex to approximate a specific job.

Micron’s 6600 ION NVMe SSD has reached 245.76TB, pushing the company’s capacity-focused PCIe Gen5 QLC line into quarter-petabyte territory. Micron began shipping the 245TB model on May 5, 2026, and is positioning it as the highest-capacity commercially available SSD. The drive sits at the top of the 6600 ION family, uses Micron’s ninth-generation G9 QLC NAND, and is available at this capacity in E3.L 9.5mm and U.2 15mm form factors. The 30.72TB, 61.44TB, and 122.88TB models also support E3.S.

Micron’s G9 QLC uses a six-plane architecture and pushes NAND I/O to 3.6 GB/s, which the company says makes it the fastest QLC currently shipping in a data center SSD. That gives Micron a useful platform for a drive that is clearly aimed at hyperscale object storage, AI data lakes, analytics, content repositories, and other environments where capacity per rack and watts per terabyte matter more than small-block write intensity.

Under Micron’s rack assumptions, a 720-drive E3.L configuration reaches 176.9PB of raw capacity with 245.76TB 6600 ION drives, compared with 31.7PB for the same number of bays populated with 44TB HDDs. Power density moves in the same direction. At Micron’s 30W peak rating, the 245TB 6600 ION delivers roughly 8.2TB per watt. The HDD Micron uses for comparison is 44TB at 10W, or about 4.4TB per watt. That 10W HDD number is an estimate because Seagate has not publicly disclosed full power specifications for its 44TB Mozaic 4+ drives, but it is consistent with the broader power envelope of current high-capacity enterprise HDDs.

That comparison is relevant because the macro picture of data centers is shifting. The IEA estimates global data center electricity consumption at roughly 415 TWh in 2024 and projects it will more than double to about 945 TWh by 2030, with AI identified as the largest driver of the increase. In the U.S., data centers are projected to account for nearly half of electricity demand growth through 2030. For large operators, grid access, floor space, cooling capacity, and watts per terabyte are no longer secondary planning details; they are constraints that shape what can actually be deployed.

There are trade-offs, and they are important to consider before getting into the test data. A 245TB QLC SSD is built for density and read-heavy access, not for replacing every high-write TLC workload. The 245.76TB 6600 ION is rated for up to 13.7 GB/s sequential read and 1.78M random read IOPS, which puts it near the top of the QLC field, on paper. Writes are more restrained, with sequential write rated at 3.0 GB/s and random write at 42,000 IOPS for both 4K and 16K transfers. The top-capacity model also uses a 16K indirection unit rather than the 4K IU used in the 30.72TB version, which matters for workloads that issue sub-16K random writes. That shows up in endurance as well, with 4K RDWPD dropping to 0.075 while 16K RDWPD holds at 0.3.

On the enterprise feature side, the 6600 ION checks the expected boxes. The drive supports OCP 2.6, NVMe 2.0d, NVMe-MI 1.2d, SPDM 1.2, CNSA 2.0 firmware verification with dual-signed updates, and SED options. It is TAA-compliant and FIPS 140-3 L2 certifiable. Micron also rates the drive at 2.5 million hours MTTF at 50°C per OCP 2.5 REL-1, which is important given the role these drives are expected to play in dense, always-on storage infrastructure.

Micron 6600 ION 245TB Specifications

Specification	Micron 6600 ION 245.76TB
Platform Overview
Capacity	245.76TB
Form Factors	U.2 (15mm) E3.L
Interface	PCIe Gen5 x4 NVMe (v2.0b)
NAND	Micron G9 QLC NAND
Performance
Sequential Read	13,700 MB/s
Sequential Write	3,000 MB/s
Random Read	1,780,000 IOPS
Random Write (4K)	42,000 IOPS
Random Write (16K)	42,000 IOPS
Read Latency	100µs (QD1, Typical)
Write Latency	20µs (QD1, Typical)
Power and Endurance
Maximum Power Consumption	≤30W
Idle Power	≤5W
Endurance	1.0 SDWPD (128KB sequential write) 0.3 RDWPD (16KB random write) 0.075 RDWPD (4KB random write)
MTTF	2.5 million device hours
UBER	<1 sector per 1017 bits read
Features and Security
Compliance	OCP 2.6 NVMe 2.0d NVMe-MI 1.2d TAA-compliant FIPS 140-3 L2 certifiable
Security Features	CNSA 2.0 SPDM 1.2 Micron SEE SED options
Additional Features	SGLs SRIS PCIe lane reversals

Micron 6600 ION 245TB Performance

Drive Testing Platform

We use a Dell PowerEdge R760 running Ubuntu 22.04.2 LTS as our test platform for all workloads in this review. Equipped with a Serial Cables Gen5 JBOF, it offers wide compatibility with U.2, E1.S, E3.S, and M.2 SSDs. Our system configuration is outlined below:

• 2 x Intel Xeon Gold 6430 (32-Core, 2.1GHz)
• 16 x 64GB DDR5-4400
• 480GB Dell BOSS SSD
• Serial Cables Gen5 JBOF
• NVIDIA L4

Drives Compared

• Solidigm P5336 122.88TB (Gen4 | 2.5″ | U.2)
• Solidigm P5336 61.44TB (Gen4 | 2.5″ | U.2)
• Micron 6550 ION 61.44TB (Gen5 | E3.S)
• DapuStor J5060 61.44 TB (Gen4 | 2.4″ | U.2)
• DapuStor R6060 122.88TB (Gen5 | E3.L)

FIO Performance Benchmark

To measure the storage performance of each SSD across common industry metrics, we leverage FIO. Prior to this review, each drive undergoes the same testing process, which includes a preconditioning step of two full drive fills with a sequential write workload, followed by steady-state performance measurement. With the Micron 6600 ION pushing to nearly 1/4 petabyte, we leveraged the sprandom preconditioning process for the random workloads on this SSD. This was to speed up the preconditioning process, which would otherwise have taken days or weeks per drive fill. As each workload type being measured changes, we run another preconditioning fill of that new transfer size.

In this section, we focus on the following FIO benchmarks:

• 128K Sequential
• 64K Random
• 16K Random
• 4K Random

128K Sequential Write (IODepth 16 / NumJobs 1)

For 128K Sequential Write, the Micron 6600 ION delivered 2,838.0MB/s. While that placed it behind the Micron 6550 ION at 10,456.4MB/s and the DapuStor R6060 at 3,920.6MB/s, it remained competitive with the rest of the comparison group. The Solidigm P5336 122.88TB reached 3,152.5MB/s, while the DapuStor J5060 and Solidigm P5336 61.44TB posted 2,883.1MB/s and 2,503.5MB/s, respectively. The result highlights the 6600 ION’s read-centric design, where write performance remains adequate but is not the primary focus.

128K Sequential Write Latency (IODepth 16 / NumJobs 1)

For 128K Sequential Write latency, the Micron 6600 ION measured 704.0µs. The Micron 6550 ION led the workload with just 191.0µs, while the DapuStor R6060 followed at 510.0µs. The Solidigm P5336 122.88TB delivered 634.0µs, narrowly outperforming the 6600 ION, while the Solidigm P5336 61.44TB recorded the highest latency at 798.0µs. Despite landing in the middle of the pack, the 6600 ION remained comfortably below the 1ms threshold.

128K Sequential Read (IODepth 64 / NumJobs 1)

For 128K Sequential Read, the Micron 6600 ION 245TB reached 12,729.8MB/s, finishing second overall behind the Micron 6550 ION 61.44TB, which led the group at 13,979.7MB/s. The DapuStor R6060 122.88TB was close behind at 11,554.0MB/s, while the remainder of the comparison group trailed significantly, with both Solidigm P5336 drives and the DapuStor J5060 clustered around the 7.1GB/s mark. The 6600 ION clearly separated itself from most of the field, with only the 6550 maintaining a measurable advantage.

128K Sequential Read Latency (IODepth 64 / NumJobs 1)

For 128K Sequential Read latency, the Micron 6600 ION recorded 628.0µs, giving it the second-best result in the comparison. The Micron 6550 ION posted the lowest latency at 571.9µs, while the DapuStor R6060 followed at 692.1µs. The remaining drives all exceeded 1ms of latency, with the Solidigm P5336 122.88TB posting the highest figure at 1123.0µs. The 6600 ION maintained a substantial latency advantage over the majority of the field, trailing only its smaller Micron sibling.

64K Random Write

For 64K Random Write, the Micron 6600 ION achieved 2,999.6MB/s. The Micron 6550 ION dominated this workload at 10,516.5MB/s, while the DapuStor R6060 followed at 3,916.9MB/s. The Solidigm P5336 122.88TB, DapuStor J5060, and Solidigm P5336 61.44TB delivered 3,182.7MB/s, 2,883.7MB/s, and 2,721.4MB/s, respectively. The 6600 ION landed in the middle of the comparison group, ahead of several competing QLC offerings.

64K Random Write Latency

For 64K Random Write latency, the Micron 6600 ION recorded 83.0µs, one of the strongest results in the group. Only the DapuStor R6060 posted a lower figure at 63.0µs. The Micron 6550 ION followed at 95.0µs, while the DapuStor J5060 recorded the highest latency at 1386.0µs. This workload showcased the 6600 ION’s ability to maintain extremely responsive write behavior despite its massive capacity.

64K Random Read

For 64K Random Read, the Micron 6600 ION reached 11,946.5MB/s, placing third overall. Both the DapuStor R6060 and Micron 6550 ION edged ahead at 13,274.8MB/s and 13,204.6MB/s, respectively. The rest of the field remained well behind, with the Solidigm P5336 models and DapuStor J5060 all hovering near 7.1GB/s. While not the outright leader, the 6600 ION remained firmly within the top tier of performers.

64K Random Read Latency

For 64K Random Read latency, the Micron 6600 ION posted 669.0µs. The Solidigm P5336 122.88TB surprisingly delivered the lowest latency at 563.0µs, followed by the Micron 6550 ION at 607.0µs. The DapuStor R6060 recorded the highest latency among the leading performers at 1285.0µs. The 6600 ION landed near the front of the group and maintained a healthy balance between throughput and response time.

16K Random Read

The Micron 6600 ION reached approximately 808K IOPS in the 16K Random Read workload, placing it among the top performers in the comparison group. The Micron 6550 ION led the field at roughly 858K IOPS, while the DapuStor R6060 followed closely behind at approximately 818K IOPS. The rest of the comparison group trailed significantly, with the Solidigm P5336, Solidigm P5336, and DapuStor J5060 all landing near the 450K IOPS mark. The result demonstrates the 6600 ION’s ability to deliver strong random read performance despite its massive capacity.

16K Random Read Latency

The Micron 6600 ION maintained one of the more consistent latency curves in the 16K Random Read workload, staying below 150µs through most of the test before finishing around 640µs at peak load. The Micron 6550 ION ended with the lowest latency at roughly 300µs, while the DapuStor R6060 finished near 650µs. The DapuStor J5060 showed the most volatility, with several large latency spikes and a peak exceeding 1ms.

16K Random Write

The Micron 6600 ION delivered approximately 192K IOPS in the 16K Random Write workload. The Micron 6550 ION remained the clear leader at roughly 661K IOPS, with the DapuStor R6060 reaching about 224K IOPS. The Solidigm P5336 61.44TB and 122.88TB followed closely at approximately 201K IOPS and 195K IOPS, while the DapuStor J5060 landed just below the 6600 ION. Although the 6600 ION was not designed as a write-focused drive, it remained competitive against the other high-capacity QLC SSDs in the comparison.

16K Random Write Latency

In the 16K Random Write workload, the Micron 6600 ION maintained low latency through much of the test before rising sharply at the highest queue depths, finishing around 11.6ms. The Micron 6550 ION delivered the most stable behavior, remaining under 500µs throughout, while the DapuStor R6060 peaked near 18.7ms and the Solidigm P5336 ended around 9.3ms. The 6600 ION remained competitive until saturation, where latency increased rapidly under maximum load.

4K Random Read

The Micron 6600 ION 245TB reached approximately 1.75 million IOPS in the 4K Random Read workload, delivering one of the strongest results in the comparison group. Only the Micron 6550 ION and DapuStor R6060 finished ahead, while the remaining drives trailed by a substantial margin. The result reinforces the 6600 ION’s ability to handle highly transactional read-heavy workloads despite being optimized for extreme storage density.

4K Random Read Latency

The Micron 6600 ION recorded 289.0µs average latency in the 4K Random Read test. This placed it among the better-performing drives in the comparison, balancing high IOPS output with responsive access times. While several competitors posted lower latency figures, the 6600 ION remained close within the leading tier of drives tested.

GPU Direct Storage

One of the tests we conducted on this testbench was the Magnum IO GPU Direct Storage (GDS) test. GDS is a feature developed by NVIDIA that allows GPUs to bypass the CPU when accessing data stored on NVMe drives or other high-speed storage devices. Instead of routing data through the CPU and system memory, GDS enables direct communication between the GPU and the storage device, significantly reducing latency and improving data throughput.

How GPU Direct Storage Works

Traditionally, when a GPU processes data stored on an NVMe drive, the data must first travel through the CPU and system memory before reaching the GPU. This process introduces bottlenecks, as the CPU acts as an intermediary, adding latency and consuming valuable system resources. GPU Direct Storage eliminates this inefficiency by enabling the GPU to access data directly from the storage device via the PCIe bus. This direct path reduces data-movement overhead, enabling faster, more efficient data transfers.

AI workloads, especially those involving deep learning, are highly data-intensive. Training large neural networks requires processing terabytes of data, and any delay in data transfer can lead to underutilized GPUs and longer training times. GPU Direct Storage addresses this challenge by ensuring that data is delivered to the GPU as quickly as possible, minimizing idle time and maximizing computational efficiency.

In addition, GDS is particularly beneficial for workloads that involve streaming large datasets, such as video processing, natural language processing, or real-time inference. By reducing the reliance on the CPU, GDS accelerates data movement and frees up CPU resources for other tasks, further enhancing overall system performance.

GDSIO Sequential Read Throughput

In our GDSIO sequential read throughput test, the Micron 6600 ION delivered the strongest small-block showing of the group and stayed competitive all the way through the larger transfer sizes. At 16K, it opened at 409.6MiB/s with a single thread, climbed steadily to 532.5MiB/s at four threads, 614.4MiB/s at eight, and 921.6MiB/s at 16, then pushed to 1.4GiB/s at 32 threads, 2.0GiB/s at 64, and settled at 1.9GiB/s by 128 threads. That gave it a clear lead in the 16K section over the DapuStor J5060, DapuStor J6060, Micron 6550 ION, and Solidigm D5-P5336, with the gap widening considerably as thread counts moved into the upper half of the test.

At 128K, the 6600 ION continued to look strong. It posted 921.6MiB/s at one thread, jumped to 1.9GiB/s at four, 3.0GiB/s at eight, and 3.8GiB/s at 16, then climbed to 4.5GiB/s at 32 threads, 5.0GiB/s at 64, and finished at 5.2GiB/s at 128. That put it at the top of the comparison group across nearly every thread count in the 128K runs, with only the DapuStor J6060 closing the gap at the very heaviest concurrency. The drive’s ability to scale smoothly from low to high thread counts at this block size was one of its standout traits.

The 1M results were a bit more uneven at the low end but finished on a high note. The 6600 ION opened at 2.9GiB/s with a single thread, dipped to 2.2GiB/s at four threads, then recovered to 3.2GiB/s at eight, 3.7GiB/s at 16, and 4.2GiB/s at 32. From there, it climbed sharply to 5.0GiB/s at 64 threads and topped out at 5.9GiB/s at 128, matching the DapuStor J6060 for the highest 1M result in the group. So while the 6600 ION had a small soft spot in the mid-thread range at 1M, its overall sequential read profile was excellent, particularly in the 16K and 128K sections where it consistently led the comparables.

GDSIO Sequential Read IOPS

In the GDSIO sequential read IOPS test, the Micron 6600 ION delivered the strongest small-block performance in the group by a wide margin. At 16K, it opened at 24.9K IOPS with a single thread, climbed to 34.5K at four threads, 41.3K at eight, and 61.5K at 16, then surged to 92.4K at 32 threads and peaked at 130.4K at 64 before settling slightly to 126.7K at 128. That peak was well ahead of the rest of the comparables, with the Solidigm D5-P5336 122.88TB the next closest at roughly 93K and the DapuStor J5060, DapuStor J6060, and Micron 6550 ION 61.44TB all topping out in the 55K to 63K range. The 6600 ION’s separation from the field grew sharply from 16 threads onward, making it the clear leader through the entire small-block portion of the test.

At 128K, the 6600 ION continued to lead the group across nearly every thread count. It posted 7.6K IOPS at one thread, jumped to 15.6K at four, 24.5K at eight, and 30.9K at 16, then climbed to 36.6K at 32 threads, 41.2K at 64, and finished at 43.0K at 128. That kept it ahead of the DapuStor J6060, which trailed by a few thousand IOPS at the heaviest thread counts, and well ahead of the DapuStor J5060, Solidigm D5-P5336, and Micron 6550 ION. The drive’s scaling from low to high concurrency at this block size was smooth and consistent, with no dips or plateaus.

In the 1M runs, the IOPS spread between drives narrowed considerably, as expected with larger transfer sizes. The 6600 ION opened at 3.0K IOPS at one thread, dipped to 2.3K at four threads, then recovered to 3.2K at eight, 3.7K at 16, and 4.3K at 32. From there, it climbed to 5.2K at 64 threads and topped out at 6.0K at 128, finishing right alongside the DapuStor J6060 at the top of the group. So while the 1M section was a closer race, the 6600 ION’s 16K and 128K results were dominant, giving it the best overall sequential read IOPS profile in the comparison.

GDSIO Sequential Read Latency

Finally, in our GDSIO sequential read latency test, the Micron 6600 ION posted some of the lowest average latencies in the group, particularly at the smaller block sizes and across the higher thread counts, where the rest of the comparables started to fall off. At 16K, it opened at 39µs with a single thread, climbed to 115µs at four threads, 192µs at eight, and 259µs at 16, then moved to 344µs at 32 threads, 487µs at 64, and 1.0ms at 128. Those results kept it ahead of the DapuStor J5060, DapuStor J6060, Micron 6550 ION, and Solidigm D5-P5336 across most of the 16K runs, with the gap widening at higher thread counts, where the 6600 ION held its latency much tighter than the rest of the field.

At 128K, the 6600 ION continued to look strong. It posted 131µs at one thread, 256µs at four, 325µs at eight, and 516µs at 16, then climbed to 873µs at 32 threads, 1.6ms at 64, and 3.0ms at 128. That kept it near the bottom of the latency stack across the entire 128K section, trailing only the DapuStor J6060 by a narrow margin at the heaviest thread counts and pulling well ahead of the Micron 6550 ION and Solidigm D5-P5336, which began to climb sharply once concurrency moved past 32 threads.

The 1M runs showed the steepest latency growth, as expected with the largest transfer size. The 6600 ION opened at 332µs with a single thread, then climbed to 1.8ms at four threads, 2.5ms at eight, 4.3ms at 16, 7.4ms at 32, 12.4ms at 64, and topped out at 21.3ms at 128. Even with that increase, the 6600 ION still held the second-lowest 1M latencies in the group, trailing only the DapuStor J6060, while the Micron 6550 ION and Solidigm D5-P5336 climbed to 47.5ms and 28.9ms, respectively, at 128 threads.

GDSIO Sequential Write Throughput

In our GDSIO sequential write throughput test, the Micron 6600 ION 245.76TB delivered a more modest showing than on the read side, particularly at larger transfer sizes, where the rest of the comparables pulled ahead. At 16K, it opened at 512.0MiB/s with a single thread, climbed to 1.1GiB/s at four threads, 1.3GiB/s at eight, and 1.4GiB/s at 16, then peaked at 1.5GiB/s at both 32 and 64 threads before easing back to 1.2GiB/s at 128. That placed it in a tight cluster with the DapuStor J5060, DapuStor J6060, Micron 6550 ION, and Solidigm D5-P5336 across the 16K runs, with no drive meaningfully separating itself at this block size.

At 128K, the 6600 ION lagged the rest of the field. It posted 2.2GiB/s at one thread, climbed to 2.9GiB/s at four threads, and held that level through eight, then settled into a slow decline at 2.8GiB/s at 16, 2.7GiB/s at 32, 2.8GiB/s at 64, and 2.5GiB/s at 128. That kept it at the bottom of the group for most of the 128K runs, with the DapuStor J6060 leading at roughly 3.7 to 3.8GiB/s, the Micron 6550 ION in second through the mid-thread counts, and the Solidigm D5-P5336 and DapuStor J5060 both sitting comfortably above the 6600 ION.

The 1M runs followed a similar pattern. The 6600 ION started at 2.9GiB/s with a single thread, held 2.8GiB/s at four, returned to 2.9GiB/s at eight, then dropped to 2.6GiB/s at 16, 2.4GiB/s at 32, and finished at 2.3GiB/s at both 64 and 128 threads. That negative scaling left it at or near the bottom of the comparison group across the entire 1M section, with the DapuStor J6060 leading at the lower thread counts and the Micron 6550 ION peaking at 3.9GiB/s at 32 threads. So while the 6600 ION had a competitive 16K showing, its 128K and 1M sequential write results trailed the rest of the comparables, especially as thread counts climbed.

GDSIO Sequential Write IOPS

In our GDSIO sequential write IOPS test, the Micron 6600 ION delivered a competitive 16K showing but trailed the rest of the field once block sizes increased. At 16K, it opened at 31.8K IOPS with a single thread, climbed to 70.2K at four threads, 85.7K at eight, and 92.9K at 16, then peaked at 98.0K at 32 threads before easing back to 95.5K at 64 and 81.9K at 128. That kept it in a tight pack with the DapuStor J6060, Micron 6550 ION, and Solidigm D5-P5336 across the 16K runs, with the Micron 6550 ION nudging slightly ahead at the 32-thread peak and the DapuStor J5060 sitting a bit lower through the mid-thread counts.

At 128K, the 6600 ION fell to the back of the comparison group. It posted 18.4K IOPS at one thread, climbed to 24.0K at four threads, and held that through eight, then settled at 23.1K at 16, 22.2K at 32, 22.8K at 64, and finished at 20.6K at 128. That left it trailing the DapuStor J6060 and Micron 6550 ION by a wide margin at the lower and middle thread counts, and only narrowly ahead of the DapuStor J5060 across most of the 128K section. The drive’s negative scaling pattern beyond four threads stood out, since the leaders maintained a flatter profile.

The 1M runs followed the same trend. The 6600 ION opened at 3.0K IOPS at one thread, dipped to 2.9K at both four and eight threads, then continued declining to 2.7K at 16, 2.5K at 32, and 2.4K at both 64 and 128 threads. That placed it at or near the bottom of the comparison group throughout the 1M section, well behind the DapuStor J6060, which led at 3.9K IOPS at the low-thread counts, and the Micron 6550 ION, which peaked at 4.0K at 32 threads. So while the 6600 ION held its own at 16K, its 128K and 1M sequential write IOPS results were the weakest of the comparables, particularly as concurrency increased.

GDSIO Sequential Write Latency

In our GDSIO sequential write latency test, the Micron 6600 ION delivered competitive numbers at the smaller block sizes but climbed to the highest latencies in the group as transfer size and thread count scaled up. At 16K, it opened at 31µs with a single thread, moved to 56µs at four threads, 92µs at eight, and 170µs at 16, then continued to 324µs at 32 threads, 666µs at 64, and 1.6ms at 128. Those results placed it in a tight band with the DapuStor J5060, DapuStor J6060, Micron 6550 ION, and Solidigm D5-P5336 across the 16K runs, with no drive meaningfully separating itself at this block size.

At 128K, the 6600 ION started to fall behind the leaders. It posted 53µs at one thread, climbed to 165µs at four, 331µs at eight, and 690µs at 16, then moved to 1.4ms at 32 threads, 2.8ms at 64, and 6.2ms at 128. That put it near the upper end of the latency stack across the 128K runs, with the DapuStor J6060 delivering the lowest latency and the Micron 6550 ION close behind, while the 6600 ION and Solidigm D5-P5336 trailed at the highest thread counts.

The 1M runs were where the 6600 ION fell furthest behind. It opened at 330µs with a single thread, then climbed sharply to 1.4ms at four threads, 2.7ms at eight, 6.0ms at 16, 12.8ms at 32, 26.8ms at 64, and topped out at 53.7ms at 128. That 128-thread result was the highest 1M write latency in the comparison group, ahead of the DapuStor J5060 at roughly 45ms, the DapuStor J6060 at 43ms, the Solidigm D5-P5336 at 41ms, and the Micron 6550 ION at 39ms. So while the 6600 ION held its own at 16K, its sequential write latencies climbed faster than the comparables as block size and concurrency increased, leaving it at the bottom of the group on the heaviest 1M workloads.

DLIO Checkpointing Benchmark

To evaluate SSD real-world performance in AI training environments, we utilized the Data and Learning Input/Output (DLIO) benchmark tool. Developed by Argonne National Laboratory, DLIO is specifically designed to test I/O patterns in deep learning workloads. It provides insights into how storage systems handle challenges such as checkpointing, data ingestion, and model training. The test is designed so that each drive is fully populated with full checkpoints; larger SSDs fit more checkpoints. The chart below illustrates how both drives handle the process across 99 checkpoints (198 for the 122TB). When training machine learning models, checkpoints are essential for periodically saving the model’s state, preventing loss of progress during interruptions or power failures. This storage demand requires robust performance, especially under sustained or intensive workloads. We used DLIO benchmark version 2.0 from the August 13, 2024, release.

To ensure our benchmarking reflected real-world scenarios, we based our testing on the LLAMA 3.1 405B model architecture. We implemented checkpointing using torch.save() to capture model parameters, optimizer states, and layer states. Our setup simulated an eight-GPU system, implementing a hybrid parallelism strategy with 4-way tensor parallelism and 2-way pipeline parallelism distributed across the eight GPUs. This configuration resulted in checkpoint sizes of 1,636GB, representative of modern large language model training requirements.

In our DLIO checkpoint benchmark, which measures the (Pass Average) time required to complete checkpoint operations across three sequential passes, the Micron 6600 ION 245.76TB started in the middle of the group and slowed considerably as additional passes accumulated. On pass one, it finished in 580.98 seconds, placing it third in the comparison behind the DapuStor R6060 122TB at 465.33 seconds and the Micron P5336 ION 61.44TB at 484.30 seconds, and ahead of the Solidigm P5336 122.88TB at 530.75 seconds and the Solidigm P5336 61.44TB at 662.68 seconds.

On pass two, the 6600 ION climbed sharply to 926.70 seconds, a jump of roughly 60 percent from its first pass. That moved it to the back of the group alongside the DapuStor R6060, which posted 934.50 seconds. The Solidigm P5336 122.88TB climbed more moderately to 746.14 seconds, while the Solidigm P5336 61.44TB and Micron P5336 ION 61.44TB stayed comparatively flat at 640.71 and 570.23 seconds, respectively. That contrast was the key takeaway from the second pass: the high-capacity QLC drives saw a much steeper rise in checkpoint time once the workload was no longer running against a fresh state.

Pass three showed the same pattern continuing. The 6600 ION finished at 968.18 seconds, a small step up from pass two and the slowest result in the group, just behind the DapuStor R6060 at 965.27 seconds. The Solidigm P5336 122.88TB landed at 757.31 seconds, the Solidigm P5336 61.44TB at 639.63 seconds, and the Micron P5336 ION 61.44TB at 585.03 seconds.

Looking at the full per-checkpoint trace rather than the three-pass averages, the Micron 6600 ION 245.76TB held a steady band in the mid-560-second range for the first 24 checkpoints, then briefly spiked to around 990 seconds for checkpoints 26 through 30 before dropping back to roughly 560 seconds and holding that level out to checkpoint 130. From checkpoint 131 onward, it stepped up to the 750 to 850-second range, then climbed again around checkpoint 200 and settled into a sustained band between 880 and 1,030 seconds for the remainder of the run, finishing 390 checkpoints in total, more than any other drive in the comparison. The DapuStor R6060 showed a similar two-stage step pattern, while the Solidigm and Micron 6550 drives held tighter, lower bands, but did not run as many checkpoints overall.

Conclusion

The Micron 6600 ION 245.76TB is a capacity play first, and it executes that brief without apology. A single drive that puts nearly a quarter petabyte into a single drive slot changes the rack math in a way that is hard to argue with, and our testing supports the read-centric positioning Micron built the drive around. Sequential and random read throughput landed at or near the top of the comparison group; GDS read performance scaled cleanly across high thread counts at 16K and 128K; and 4K random read approached 1.75M IOPS. For object stores, AI data lakes, analytics, and content repositories where the access pattern is read-dominated and the constraint is capacity per watt, this drive does what it is supposed to do.

The trade-offs are equally clear, and buyers should size for them rather than around them. Sequential write throughput sat in the middle of the field, GDS write performance and latency fell behind the comparables as block size and concurrency climbed, and the DLIO checkpoint runs showed the 6600 ION slowing meaningfully once the workload moved off a fresh state, finishing at the back of the group on passes two and three. The 16K indirection unit and the 0.075 4K RDWPD rating are the relevant fine print for anyone considering sub-16K random write traffic. None of this is a surprise for a QLC drive built for density, but it does define where the 6600 ION belongs and where it does not.

The broader case rests on power and footprint, not peak performance. At 8.2TB per watt against roughly 4.4TB per watt for the high-capacity HDDs Micron compares to, and with a 720-drive rack reaching 176.9PB versus 31.7PB for an equivalent HDD count, the efficiency argument is the product’s point. For operators managing storage growth inside fixed power and floor-space budgets, the 6600 ION 245TB is a credible answer, provided the workload is matched to its read-heavy design.

Micron 6600 ION Product Page

The post Micron 6600 ION 245TB SSD Review: A Quarter Petabyte Per Drive Bay appeared first on StorageReview.com.

JetCool has launched its SmartPlate System for Dell PowerEdge R770 and R7725 servers, expanding its direct-to-chip liquid cooling portfolio to align with next-generation server platforms based on 5th Gen AMD EPYC and Intel Xeon 6 processors. The company originally announced the Dell PowerEdge support at Dell Technologies World in May, positioning the solution for enterprise data centers and distributed edge deployments where power, heat, and rack space are increasingly limiting factors. The move extends an existing relationship: JetCool, a Flex company, is a Dell OEM partner whose prior-generation SmartPlate already shipped on the PowerEdge R670 and R760, and the R770 and R7725 bring the sealed cooling approach to Dell’s next-gen Intel and AMD platforms.

At a high level, SmartPlate is designed as a server-integrated, direct-to-chip cooling approach that can be deployed without tying into the facility water system. JetCool describes the system as fully sealed and ready to install, aiming to reduce deployment friction in environments that were architected around traditional air cooling. This matters for operators looking to introduce liquid cooling selectively, such as rolling upgrades in existing rooms or edge sites where retrofits and plumbing changes are impractical.

Improved Thermal Performance

JetCool cites testing showing an average 13% reduction in IT power on these SmartPlate-equipped systems, attributing the gains to improved thermal performance and more efficient chip-level cooling. (JetCool’s broader SmartPlate marketing cites an average 15%, so 13% is the figure tied specifically to these next-gen R770 and R7725 integrations.) The company positions that reduction as a lever to increase compute density per rack, particularly in constrained facilities where adding power delivery and cooling capacity is difficult. JetCool founder Dr. Bernie Malouin said the integrations are intended to give enterprises and edge operators a practical path to higher rack density and better efficiency without changes to existing infrastructure. We have a podcast on this if you’d like to explore further.

John Sasser, CTO of JetCool partner Sabey Data Centers, described the system, based on firsthand deployment, as a way to bring direct-to-chip cooling to traditionally air-cooled facilities through a closed-loop, server-integrated design that supports phased rollouts rather than disruptive facility-level modifications.

JetCool founder Dr. Bernie Malouin said the Dell PowerEdge R770 and R7725 integrations are intended to provide a practical path to higher rack density and performance improvements without requiring changes to existing infrastructure. Sabey Data Centers CTO John Sasser also described the system as a way to bring direct-to-chip cooling into traditionally air-cooled facilities using a closed-loop design integrated directly with servers, enabling phased rollouts rather than disruptive facility-level modifications. Both statements emphasized deployment practicality and measurable efficiency improvements rather than a wholesale data center redesign.

Availability

JetCool says SmartPlate Systems for Dell PowerEdge R770 and R7725 are available immediately. The company is positioning the platform for organizations looking to increase on-prem capacity and reduce energy consumption while avoiding facility infrastructure rework.

The post JetCool Brings SmartPlate Liquid Cooling to Dell PowerEdge R770 and R7725 appeared first on StorageReview.com.

At COMPUTEX 2026, HPE announced the ProLiant Compute DL394 Gen12, a next-generation 2U server built around the NVIDIA Vera CPU. The platform is designed to support emerging agentic AI and data-intensive workloads that require high memory bandwidth, low latency, and deterministic performance. The system integrates HPE’s enterprise management and security stack, including Integrated Lights-Out and Compute Ops Management, targeting organizations deploying large-scale AI and real-time data processing environments.

The launch is tied to a new collaboration between HPE, NVIDIA, and data streaming company Redpanda, which the New York Stock Exchange is exploring for its own agentic AI infrastructure. The work centers on technology optimized for agentic AI across data storage and processing, monitoring, management, and security, with the DL394 Gen12 as the foundational compute platform. NYSE is the early proof point, which fits the financial services workload profile HPE is targeting.

HPE leadership framed the launch around the shift from generative AI to agentic AI, in which systems perform real-time reasoning and make autonomous decisions, raising the bar for latency consistency and memory throughput. NVIDIA positioned Vera as purpose-built to orchestrate AI factories, claiming 2x the efficiency and faster task completion than x86. The DL394 Gen12 is intended to put those capabilities to work in enterprise and financial services deployments.

Architecture Focused on Memory Bandwidth and Latency

The DL394 Gen12 is centered on the NVIDIA Vera CPU, which uses a monolithic architecture rather than a chiplet-based design. This approach avoids the non-uniform memory access characteristics common in high-core-count processors, where memory latency can vary depending on data locality. By eliminating NUMA-related variability, the platform is engineered to deliver more predictable performance for distributed AI workloads.

The system leverages LPDDR5X memory, enabling an aggregate bandwidth of up to 1.2TB/s and approximately 14GB/s per core. This level of throughput is intended to support high-speed data ingestion and processing, particularly for workloads that require continuous streaming and real-time inference. In this configuration, the Vera CPU functions as an orchestration layer, balancing compute and memory resources across workloads to reduce inefficiencies and improve utilization.

Integrated Security and Lifecycle Protection

Security is embedded at the hardware and firmware levels through HPE’s Silicon Root of Trust. The DL394 Gen12 also incorporates iLO 7 with a secure enclave, protecting the server lifecycle from manufacturing through decommissioning. These capabilities are designed to mitigate firmware-level attacks and ensure system integrity in regulated environments.

HPE indicated that this generation of ProLiant systems is the first to meet NIST requirements for quantum-resistant cryptography. This positions the platform for long-term deployment in environments where data protection standards are expected to evolve alongside emerging threats.

Unified Management with AI-Driven Operations

The DL394 Gen12 integrates with HPE Compute Ops Management, providing a centralized platform for managing distributed infrastructure. The software layer delivers AI-driven insights into system health, performance, and capacity, reducing operational overhead and minimizing downtime.

By consolidating monitoring and automation into a single interface, HPE aims to simplify infrastructure management for organizations operating at scale. This is particularly relevant for AI deployments, where dynamic workloads and resource demands require continuous optimization.

Availability

The HPE ProLiant Compute DL394 Gen12 is expected to be available in fall 2026 as part of the NVIDIA AI Computing by HPE portfolio.

The post HPE ProLiant Compute DL394 Gen12 Brings NVIDIA Vera CPU to Agentic AI appeared first on StorageReview.com.

QNAP has released QuTS hero h6.0 Beta, the latest version of its ZFS-based NAS operating system. This update brings several additions for enterprise NAS users, including dual-NAS high availability, immutable snapshots, centralized encryption key management, storage tiering, access controls, and AI-assisted administration.

The update extends HA support to additional QNAP models, and QNAP says more than 90 percent of the operating system’s services are now HA-ready. A handful remain unsupported in this beta, including Real-time SnapSync, Q’center, third-party apps, and VJBOD.

Dual-NAS High Availability Expands to More Systems

QuTS hero h6.0 provides expanded support for Dual-NAS High Availability. Using High Availability Manager, two NAS systems can be configured as an Active-Passive cluster designed to maintain service continuity if one system becomes unavailable.

The update extends HA support to additional QNAP models and allows HA clusters to connect to JBOD expansion enclosures. This provides additional flexibility when scaling storage capacity while maintaining redundancy within the cluster. Nearly all QNAP NAS applications are supported in the HA environment, except for third-party and legacy applications.

Immutable Snapshots and Centralized Key Management Strengthen Security

Data protection also receives several additions in h6.0, including Immutable Snapshots. Available across all QuTS hero models, the feature locks snapshot data for a defined protection period, preventing modification or deletion during that time. It serves as a safeguard against ransomware attacks and a mechanism to preserve data integrity during recovery.

The release also introduces KMIP key management integration. As a KMIP client, QNAP NAS systems can connect to centralized enterprise key management servers, enabling remote key management and automatic key application. This architecture is designed to align with FIPS 140-3 security requirements and enterprise security practices.

Additional Security Controls Added Across the Platform

Several new security features have also been integrated into QuTS hero h6.0, including FIDO2 passkeys, which provide password-free authentication. Moreover, Secure Boot verifies firmware integrity during startup to prevent unverified code from loading.

Ransomware Guard, which QNAP lists as coming soon, expands protection by adding behavioral monitoring and threat isolation to Malware Remover, enabling it to detect suspicious activity, log anomalies, and respond to threats in real time. Secure IP Access, also coming soon, adds granular IP-based controls to reduce network exposure and enforce context-based access policies.

Storage Tiering and SMB Performance Enhancements

QuTS hero h6.0 also adds support for Qtier hero, bringing QNAP’s storage tiering technology to QuTS hero NAS systems. Administrators can manually place data on SSD or HDD storage tiers depending on workload requirements, allowing performance-sensitive data and capacity-oriented storage to be managed separately. QNAP highlights workloads such as file servers, virtual machines, and video production environments as potential use cases.

Another storage-related addition is FileTiers, which is scheduled for a future update. The feature automatically moves data between hot, warm, and cold storage tiers across NAS systems based on access frequency or administrator-defined policies. Support for High Availability and HBS backup is also planned.

The update further introduces a kernel-mode SMB daemon with encryption support. QNAP says that moving SMB services into kernel mode improves throughput while maintaining encrypted file transfers.

Expanded Management and Access Controls

Several management-focused additions are included in h6.0 as well, including QNAP ID SSO, which enables single sign-on across NAS and cloud services. Additionally, Fibre Channel NPIV allows multiple virtual WWPNs to operate through a single Fibre Channel port.

ACL 2.0 introduces a redesigned permission-handling engine intended to improve performance and administration for large directory structures. Administrators also gain access to AMIZcloud Monitoring, which provides centralized visibility into HA groups, including cluster health, latency information, and alert status.

AI Features Added For Search And Administration

AI-related functionality is also expanded in the newest update via additions to both search and system management. Qsirch now supports RAG-based search using locally deployed open-source large language models, including DeepSeek, Gemma, Phi, and Mistral, running on GPU-capable NAS systems. This provides document summarization and semantic search while keeping data on local infrastructure.

The new MCP Assistant allows administrators to interact with NAS systems via natural-language commands on platforms such as Claude Desktop, VS Code, Telegram, and n8n.

Availability

QuTS hero h6.0 Beta is available now through the QNAP Download Center.

QNAP QuTS hero h6.0

The post QNAP QuTS hero h6.0 Beta Adds Dual-NAS HA, Immutable Snapshots, and On-Prem AI appeared first on StorageReview.com.

At the Mistral AI NOW Summit, VAST Data outlined its partnership with Mistral AI and Mistral Compute, detailing a joint effort to deploy NVIDIA-accelerated AI factories in Europe. The collaboration aligns three core layers required for production AI: NVIDIA accelerated compute, Mistral’s frontier models, and the VAST AI Operating System as the unified data platform.

At the center of the deployment is a large-scale implementation of NVIDIA GB300 NVL72 systems, representing one of the highest-density concentrations of this architecture in Europe. Mistral Compute will operate the infrastructure as part of its AI cloud platform, while VAST provides the data layer that manages data access, movement, governance, and performance across the environment.

Data Platform for Production AI Workloads

VAST positions its AI Operating System as the foundational data layer for AI factories, designed to support end-to-end workflows without the fragmentation typically seen in multi-system pipelines. Within the Mistral Compute environment, the platform provides a shared data architecture spanning training, inference, retrieval, and enterprise deployment. This eliminates the need for duplicated datasets and reduces operational overhead associated with siloed storage and data movement.

The platform is built to handle the full range of data types and runtime states required by modern AI workloads. This includes structured and unstructured data such as files and tables, as well as vectors, key-value cache, event streams, and persistent agent memory. These capabilities are delivered across distributed environments, enabling consistent performance and access regardless of location.

Extending Across Research, Cloud, and Enterprise

The partnership extends beyond a single deployment and reflects a broader integration across Mistral’s AI lifecycle. VAST is already deployed in AI cloud environments used by Mistral AI for model training and research, supporting the development and operation of its own models, including Voxtral, Ministral, and Codestral. Through VAST DataSpace, Mistral teams can operate across multiple cloud environments using a unified namespace, avoiding the need to redesign data pipelines when shifting workloads.

Mistral Compute, now an NVIDIA Cloud Partner, has adopted the VAST AI Operating System as its core data platform for its managed AI cloud services. The system is currently in production, supporting both internal Mistral workloads and customer-facing deployments. The addition of GB300 NVL72 infrastructure extends this environment to support higher-scale AI factory operations.

Supporting Enterprise AI and Data Control

As Mistral brings its models into enterprise environments, the requirement shifts toward integrating models with enterprise data while maintaining performance and governance. VAST provides a common operating layer that connects models to enterprise datasets, enabling consistent data access and control across training and inference workflows.

The architecture is designed to support requirements around data locality, governance, and isolation, which are increasingly critical for European enterprises and public-sector organizations. By maintaining a unified data foundation across research, cloud, and enterprise deployments, the platform enables organizations to retain control over how data is stored, accessed, and used in AI workflows.

Positioning for European AI Sovereignty

The deployment reflects a broader push toward regional AI infrastructure and data sovereignty. With compute, models, and data platform integrated within a European-operated environment, the solution provides a framework for building and running advanced AI systems with localized control.

In this model, the data layer becomes the primary control point for AI operations, governing how data flows between training, inference, and production systems. VAST’s role as the unified data platform positions it as a key component in enabling Mistral Compute to deliver both high-performance AI infrastructure and the governance capabilities required for regulated and enterprise use cases.

The post VAST Data Powers Mistral Compute AI Factories on NVIDIA GB300 NVL72 appeared first on StorageReview.com.