SK hynix says it shipped samples of its 12-stack HBM4E (High Bandwidth Memory) to major customers, marking the next step in the company’s AI memory roadmap. The company said the samples were delivered on schedule and that it will work closely with partners toward timely mass production. HBM4E builds on the company’s HBM portfolio with

SK hynix has announced plans to invest KRW 100 trillion (about US$65 billion at the current exchange rate) in Cheongju, South Korea, to expand its NAND flash manufacturing and advanced packaging capacity to meet rising AI-driven memory demand. The announcement was made by President and CEO Kwak Noh-Jung during the National Briefing on the Strategy

AMD silicon arguably got here first: Strix Halo mini PCs and laptops were shipping with 128GB of unified memory well before NVIDIA entered the picture. But the local-AI desktop as a category is one NVIDIA effectively created when it put a Grace Blackwell superchip in a one-liter box and called it DGX Spark. The pitch

NVIDIA has dropped the subscription requirement for Omniverse. The platform is now free for development, production, and redistribution, with no NVIDIA AI Enterprise subscription required. For a product that carried a $4,500-per-GPU-per-year list price under NVIDIA AI Enterprise, and whose Nucleus server pricing prompted a developer forum thread titled “Pricing: $25,000 for the nucleus server!?”,

Sandisk and Kioxia have both begun sampling BiCS10 1Tb TLC 3D NAND, the 10th generation of the BiCS flash memory the two companies jointly develop and manufacture. The announcements land in parallel because the technology is shared: Sandisk is framing it around higher density, faster interfaces, and better power efficiency for data-intensive workloads, while Kioxia

Couchbase has announced the general availability of its AI Data Plane, positioning it as a unified data infrastructure layer for enterprise AI agents. The release is aimed at organizations trying to move agentic AI projects beyond pilots by combining persistent memory, real-time context retrieval, and consistent data access across cloud, edge, and lakehouse environments within

Intel’s Core Ultra 7 265K sits in the middle of the Arrow Lake-S desktop stack, pairing 20 cores with a much larger architectural change than you might expect. The layout consists of 8 Lion Cove P-cores and 12 Skymont E-cores, for 20 cores and 20 threads, with Hyper-Threading removed in this generation. The P-cores have

Scality and OVHcloud have expanded their technology partnership with a joint storage platform aimed at organizations requiring sovereign infrastructure for AI and mission-critical data. The combined solution is designed to keep sensitive data under customer control while supporting high-performance AI workloads without relying on U.S. hyperscale cloud providers. The offering is available as either a

AMD has introduced the Versal Premium Gen 2 Memory-on-Package (MoP) adaptive SoCs, adding integrated LPDDR5X memory directly into the device package to increase bandwidth while reducing board complexity and footprint. The new devices target applications that require high memory throughput within constrained power and space budgets, including networking, physical AI, aerospace and defense, test and