Infrastructure and hardware comprise the physical compute systems, specialized chips, and data center components that power artificial intelligence workloads. Truncated covers chip architectures, supply chains, energy requirements, and the hardware deployments that enable large scale computing.
AI-generated summary Baseten released an updated delta weight sync design for models like GLM-5.3 that reduces synchronization time to 15 seconds. Baseten says the new design enables throughput improvements by as much as 3.2x through optimized GPU memory utilization.
AI-generated summary AMD is releasing an early access update that brings validated and optimized inference support for five Medical Open Network for AI Model Zoo bundles on AMD Instinct GPUs. ROCm Blogs says the five bundles clear upstream Sørensen-Dice overlap score thresholds with no changes to model weights and no changes to inference logic.
AI-generated summary Project Suncatcher has launched a prototype satellite and successfully sent a command to orbit to boot up the TPU and receive a response. @sundarpichai states that this milestone involved running a prompt and receiving a complete haiku back on earth.
AI-generated summary Canadian renewable firm PowerBank plans to develop a 2MW modular data center in Ontario, Canada. The company states that the site could generate around CA$2 million a year in lease revenue if the project becomes operational.
AI-generated summary Netflix makes available its GPU reliability platform, which uses layered health checks, automated remediation, and fault-tolerant training to keep jobs running despite hardware failures. Netflix reports more than 99.95% GPU-node availability and an increase in job success from approximately 98% to more than 99.5%.
AI-generated summary AI infrastructure company Runware announced plans to scale production at its Romania-based factory to 30 megawatts of data center capacity per day by 2028. Runware says it will manufacture the modular data centers on an assembly line and aims to bring the pods online within 48 hours.
AI-generated summary Alibaba Cloud launched its first public cloud region in the Netherlands with two new data centers in Amsterdam. Alibaba Cloud says the new infrastructure will help deliver resilient low-latency services and support local data processing needs.
AI-generated summary Wärtsilä released a new landmark analysis examining data center power constraints and the speed to powerless risk across the Americas. Wärtsilä says that power systems cannot keep pace with data center construction speed, creating a hidden risk to artificial intelligence infrastructure growth.
AI-generated summary Siemens Energy announced a $1 billion investment with the goal of increasing its market capacity for grid and gas turbine equipment. The company states that the initiative will scale up production, create more than 1,500 jobs, and target site expansions to meet market demands.
AI-generated summary Murata announced an investment of $6.4 billion to double output and meet artificial intelligence demand. The investment by Murata aims to increase production capacity in response to the growing market demand.
AI-generated summary Hyundai Motor Group announced it plans to establish two special purpose companies next month to develop artificial intelligence data centers as part of its infrastructure expansion. Hyundai Motor Group says one special purpose company will oversee data center construction and infrastructure development while the other will handle graphics processing unit services.
AI-generated summary OpenRouter launched a new provider routing algorithm designed to reward providers for delivering fast, high-throughput, and reliable inference. OpenRouter says the algorithm has unlocked improvements in time to first token and throughput, along with a reduction in capacity errors.
The top AI news in your email delivered each morning. Choose categories