AI Hardware News
AI hardware news covering GPUs, chips, datacenters, edge devices, compute infrastructure, and the systems powering modern artificial intelligence.
SK Group and NVIDIA have formed a $500 billion partnership to build AI factories and advance next-generation memory technology for global compute demand.
Alphabet shares rise as Google develops a new 'Frozen v2' server chip, aiming for 6-10x greater efficiency for its Gemini AI models by 2028.
Japan Partners With NVIDIA to Accelerate National Physical AI Infrastructure Development
Japan partners with NVIDIA to build a national physical AI infrastructure, leveraging Isaac, Metropolis, and Jetson to advance robotics and manufacturing.
Google's LiteRT.js brings native on-device inference to the browser, supporting WebGPU and WebNN for up to 60x faster performance than CPU-based execution.
Meta is investing over $50 billion to build the 5GW Hyperion data center supercluster in Richland Parish, Louisiana, to support its growing AI initiatives.
Apple has filed a lawsuit against OpenAI, alleging that senior leadership orchestrated a campaign to steal proprietary hardware trade secrets.
NVIDIA Research introduces HORIZON, an agentic framework that automates hardware design by iteratively evolving git worktrees to hit 100% benchmark success.
Meta's latest AI smart glasses are reigniting privacy fears as critics raise concerns about the potential for recording bystanders without consent.
Prague-based EquiLibre Technologies, founded by former DeepMind researchers, reaches a $500M valuation by applying poker AI to stock market trading.
MIT researchers developed Gleanmer, an ultra-efficient chip that uses 3D Gaussian mapping to help tiny, low-power robots navigate complex environments.