IBM Launches Sovereign Core on Intel Xeon 6 for Trusted AI Workloads
IBM introduces a new AI solution designed for sensitive workloads, leveraging Intel Xeon 6 processors to enhance security and trust in data processing.
343 articles
The chips and systems powering modern AI — GPUs, accelerators, data-center infrastructure, and the hardware roadmaps from Nvidia, AMD, and custom silicon makers driving training and inference.
IBM introduces a new AI solution designed for sensitive workloads, leveraging Intel Xeon 6 processors to enhance security and trust in data processing.
AMD details how to adapt an FP16 GEMM kernel to FP8 on RDNA 4 GPUs, enabling larger tiles and deeper K slices due to smaller element sizes.
Google announced a new service that allocates dedicated hardware-isolated micro virtual machines to each AI agent, enhancing efficiency and scalability for agentic systems.
As AI inference demand surges, data storage costs are becoming a critical infrastructure challenge, with back-end data footprints growing significantly.
Nvidia's DSX POD could reduce AI data center deployment from six months to as little as 14 days by preassembling components in containers.
NVIDIA introduces Confidential Bare Metal, adding to its Confidential Computing portfolio to help secure model and data privacy for AI workloads.
NVIDIA announced updates to its local AI hardware lineup, including new models, software tools, and support for multiple operating systems.
NVIDIA's DSX POD reduces data center deployment time from six months to 14 days, enabling faster AI infrastructure setup.
Amulya Vishwanath, ex-Nvidia turned VC, says physical AI is in its LEGO era, with foundational robotics building blocks being rethought as the industry moves toward scalable innovation.
Samsung is building five semiconductor fabrication plants in Pyeongtaek, South Korea, with three already operational and a fourth coming online to meet rising AI demand.
Tenstorrent, an AI chip startup, secured $1.5 billion at a $6.2 billion valuation, more than doubling its previous valuation from $2.6 billion.
Deepgram improved its voice AI processing efficiency by 11–70 percent using AWS Trainium2, reducing costs per audio token processed.