At Google Cloud Next, Google announced its next-generation AI accelerators: the TPU v8t “Sunfish” for training and the TPU v8i “Zebrafish” for inference, along with its new Virgo data center fabric. From Google’s blog posts, it is clear these chips are optimized for “the agentic era”: training frontier mixture-of-experts models at the hundreds-of-thousands-of-chips scale, and
QNAP Systems has introduced the QAI-h1290FX, an edge AI storage server designed for organizations that want to run large language models, retrieval-augmented generation search, and other generative AI workloads on their own infrastructure. The system is designed for enterprises that balance AI adoption with requirements for data privacy, low latency, governance, and operational control, enabling
HPE has announced an expansion of its ProLiant edge compute portfolio, introducing new hardware designed to support AI inferencing and mission-critical applications in distributed or harsh environments. The update includes the new HPE ProLiant Compute EL2000 chassis, which serves as the foundation for two Gen12 servers, and an enhanced version of the HPE ProLiant DL145
Proxmox has released Proxmox Backup Server 4.2, a new update built on Debian 13.4 “Trixie” with refreshed packages, improved hardware support, and security updates from the newer base system. The release also moves to Linux kernel 7.0 as the stable default and includes ZFS 2.4 for storage environments that rely on ZFS-backed infrastructure. The update
Backblaze’s Q1 2026 Network Stats report shows a slower winter period for neocloud and hyperscaler traffic, with activity beginning to rise again in March. The report examines network-level infrastructure data across Backblaze’s environment, providing context on how data movement is changing as AI-related workloads continue to influence cloud and storage usage patterns. Since the launch
KIOXIA has introduced the BG8 Series, a new client SSD line aimed at PC OEMs that brings the PCIe Gen5 interface into more mainstream systems. The lineup is designed for a broad range of everyday computing hardware, including slim laptops, consumer and commercial notebooks, and desktop PCs. KIOXIA BG8 Features and Performance KIOXIA’s BG8 Series
At Google CloudNext, IBM and Google Cloud announced an expanded collaboration to address a recurring challenge for enterprise customers: modernizing core systems and operationalizing AI across hybrid and multi-cloud environments without increasing complexity. The joint effort focuses on improving interoperability across platforms, data, and tooling while maintaining operational consistency and security. The partnership combines Google
The DGX Spark platform is familiar territory for us at this point. We’ve reviewed the Dell, ASUS, Acer, and Gigabyte takes on NVIDIA’s GB10 Grace Blackwell reference design, and the core ingredients are consistent across all of them: 1,000 TOPS of FP4 compute, 128GB of unified LPDDR5x memory, and dual 200GbE networking in a 150mm
NVIDIA and Google Cloud used Google Cloud Next in Las Vegas to outline a new phase of their long-standing engineering partnership, introducing updates to the Google Cloud AI Hypercomputer platform to scale agentic and physical AI for production environments. The companies continue to co-design infrastructure spanning silicon, systems, networking, and software to support increasingly complex
LaCie has been a fixture in our lab for well over a decade. From the 8big Rack Thunderbolt 2 we covered in 2014 through the many generations of 5big, 6big, 8big, and Rugged devices that have followed, the formula has been consistent: premium Neil Poulton-designed enclosures, Seagate drives inside, Mac-centric polish, a solid warranty, and a
Amazon