104 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.
1X, a Norwegian-American robotics company, revealed its soft robot companion Neo with hands offering 25 degrees of freedom, slightly less than human hands, in a bid to create a safe, helpful domestic assistant.
Nvidia's stock price has fallen 15% since its May peak, while Micron's value nearly tripled as memory demand outpaces supply.
Meta plans to start producing its latest AI-specific chips in September, aiming to reduce GPU costs amid component shortages.
AMD details AI training network traffic trends, showing congestion peaks and harmonic-like patterns in large-scale clusters.
Meta is testing AI glasses that continuously record audio and photos, capturing users' entire day for potential AI training.
US manufacturers are facing rising electricity costs as AI data centers strain the power grid, with some bills increasing from $1,600 to $12,000 per month.
AMD outlines how occupancy on its MI355X GPU is determined by four resource limiters, with matrix cores maintaining near peak utilization even at lower occupancy levels.
Chinese startup Deepseek is building its own AI chip, designed for inference workloads, as it seeks to reduce reliance on Nvidia and Huawei.
Solos released two new smart glasses, including a $299 camera-enabled model with a $79 privacy shield accessory for controlling camera access.
Forterra's more than 100 autonomous ATVs have been deployed in Ukraine for nine months, marking the largest US defense tech company deployment of autonomous ground vehicles in combat.
SK Hynix plans to raise $28 billion via a U.S. IPO, offering ADRs to U.S. investors as AI demand outpaces supply.
TinyML models are enabling real-time health monitoring in regions with limited internet access, as seen in Brazil’s Patient Simulator Lab.