StorageReview.com

Enterprise infrastructure since 1998. Storage, servers, networking, AI. Independent reviews, lab benchmarks, breaking news from the data center.

[StorageReview.com @storagereview]

The New Standard for Academic AI Infrastructure https://t.co/mvOT9qNVmL

[StorageReview.com @storagereview]

Synology's flagship PAS7700 at Computex 2026: active-active dual controllers, 48 U.3 NVMe bays in 4U, NVMe-oF over TCP, FC, and RoCE. Failover in about 5 seconds, up to 2M IOPS and 30GB/s. Built for VDI, virtualization, and AI storage. #Synology #NVMe #Computex2026 @Synology

[StorageReview.com @storagereview]

$100M Series C for ZutaCore, backed by Mitsubishi Electric, Carrier Ventures, and Samsung. Waterless two-phase cooling for processors past 4,000W, 75+ deployments live. Full story, including the 2MW test rig: https://t.co/zIWr9es9Uy #ZutaCore #LiquidCooling #AIInfrastructure @ZutaCore

[StorageReview.com @storagereview]

28 dual-socket, dual-node servers in one rack. MSI's 21" ORv3 liquid-cooled rack at Computex runs up to 100kW with an integrated CDU and 48V busbar, beside a 19" air-cooled rack of 16 dual-node systems. Cold plates and leak detection per node. #Computex2026 #MSI #LiquidCooling @msitweets

[StorageReview.com @storagereview]

CoolIT validated a 15kW single-phase DLC coldplate: nearly 4x its 4kW design from March 2025 and 10x more than current AI GPUs need. 1.2 L/min/kW water-glycol, 45°C warm water. How it extends the DLC roadmap past 2030: https://t.co/m1AYf7utiY #LiquidCooling #DLC #DataCenter @CoolITSystems

[StorageReview.com @storagereview]

A 1U server pricing 100.8 options/sec: HPE's liquid-cooled XD230 with Intel Xeon 6980P and Micron 8800 MT/s MRDIMMs just set the STAC-A2 record, running cold Greeks in 33ms. How memory bandwidth got it there: https://t.co/05GrtyNsrW #STACA2 #HPE #MRDIMM @HPE @MicronTech @intel

[StorageReview.com @storagereview]

MSI's COMPUTEX 2026 lineup runs liquid: ORv3 racks up to 100kW with integrated CDUs, MGX servers scaling to eight RTX PRO 6000s, and the XpertStation WS300 putting 748GB of coherent memory on a desk. Full rundown: https://t.co/0MPsIbs6rz #COMPUTEX2026 #MSI #LiquidCooling @msitweets

[StorageReview.com @storagereview]

Synology's Active Protect Manager 2.0: native backup for Proxmox, Nutanix, ESXi, and Hyper-V with cross-platform restore between any of them. Google Workspace joins M365, and anomaly detection runs during backup, not after.
#Computex2026 #Synology #DataProtection @Synology

[StorageReview.com @storagereview]

IBM and Red Hat are putting $5B and 20,000 engineers behind Project Lightwell, an AI-driven clearinghouse for open source security, days after joining Anthropic's Project Glasswing. We break down how the two initiatives fit together: https://t.co/wP3Mv6eBj9 #OpenSource #IBM @IBM

[StorageReview.com @storagereview]

MSI's ExpertStation WS300 at Computex 2026: a fully liquid-cooled NVIDIA GB300 workstation. Copper cold plates, internal pump, radiators on the side and top. Next to it: slot-loading RTX Pro 6000s with quick-disconnect liquid fittings.
#Computex2026 #MSI #GB300 @msitweets

[StorageReview.com @storagereview]

We stopped by Seasonic at Computex 2026 for a look at the PRIME ENTERPRISE line, their flagship power platform for AI. The PX series tops out at the PX-3200 (3200W, requires 230/240V at 16A), with PX-3000 and PX-2800 below it, all ATX 3.1 and 80 PLUS Platinum. The TX-1600 covers workstation duty and is showing up in GB300 partner systems. Everything is built around EDPP for the idle-to-peak load swings of training and inference, plus OptiGuard, which monitors current and temperature per pin on the 12V-2x6 connector and shuts down before anything melts. Watch:

@Seasonic #Computex2026 #AI #powersupply

[StorageReview.com @storagereview]

Full Micron 6600 ION 245TB review with FIO, GPU Direct Storage, and DLIO checkpoint results: https://t.co/dddKiYe9Ob @MicronTech

[StorageReview.com @storagereview]

https://t.co/4VtLsqJgnw

[StorageReview.com @storagereview]

JetCool brought its SmartPlate direct-to-chip liquid cooling to the Dell PowerEdge R770 and R7725. Fully sealed, server-integrated, and no facility water hookup required, so it drops into existing air-cooled data centers without re-plumbing the room. Average 13% IT power reduction. Available now, report on the website.

#JetCool #liquidcooling #Dell #PowerEdge @jetcool_usa

[StorageReview.com @storagereview]

Full review with all benchmarks, power testing, and our take on the price math against the RX 9070: https://t.co/bffKBpdK2r

[StorageReview.com @storagereview]

https://t.co/yAIRcmB3Vr

[StorageReview.com @storagereview]

First deep look at the HGX Rubin NVL8 in near-production form from Dell Tech World. Eight Rubin GPUs at 288GB HBM4 each, four NVLink 6 switches bridging them under the manifold, ConnectX-9 at 1.6 TB/s per port for scale-out, BlueField-4 for KV cache offload, and host CPU flexibility across NVIDIA Vera, AMD, and Intel. Liquid cooled end to end. Watch:

@NVIDIA #Rubin #AI #GPU @Dell

[StorageReview.com @storagereview]

At Computex 2026 NVIDIA put the same Spark and Blackwell silicon into a Windows laptop, desktop, and workstation, with RTX Spark (6,144 CUDA cores, 1 PFLOP FP4, 128GB unified) at the center, plus DGX Station and a big physical AI push. Our full breakdown: https://t.co/5HSEWR5kX4 @NVIDIAAI

[StorageReview.com @storagereview]

https://t.co/9VsKOKt2ZI

[StorageReview.com @storagereview]

Vera talk is heating up, so resharing these from GTC: NVIDIA's Vera CPU 8-way compute tray up close.

Liquid-cooled, SOCAMM LPDDR5X packed around the CPUs, with 1.2 TB/s memory bandwidth and 1.8 TB/s NVLink-C2C on the spec board.

Now the servers are getting close to landing. HPE brought the ProLiant DL394 Gen12 to COMPUTEX, and Dell is bringing its own Vera systems to market.
#NVIDIA #Vera #HPE #Dell @NVIDIAAI