NVIDIA Supply Agreements Target Major AI Packaging and Optics Expansion
NVIDIA announced capacity agreements across substrates, advanced packaging, testing and optics, including a US$1.1 billion advance supporting planned U.S. KYEC facilities.
INFRADAY COVERAGE
Data Centers & Digital Infrastructure
Ideas and analysis across the systems that connect our economy.
Google Cloud Plans Vera Rubin A5X Scale-Up Across Multiple Sites
Google Cloud plans general availability of NVIDIA Vera Rubin A5X instances in November and expansion to as many as 960,000 Rubin GPUs across multiple sites.
NVIDIA DSX Links AI Factory Power, Cooling and Grid Operations
NVIDIA outlined DSX integrations spanning substations, facility power, cooling and compute as developers confront power and interconnection bottlenecks.
NVIDIA Expands Cosmos Data Factory for Robotics and Physical AI
NVIDIA expanded Cosmos with five cloud-native data blueprints and new tools for training and testing world models used in robotics and industrial AI.
NVIDIA Unveils Containerized DSX POD AI Factories for 2027 Deployment
NVIDIA introduced a containerized Vera Rubin AI-factory reference design that partners plan to deliver worldwide in early 2027.
NVIDIA and Cisco Propose 400G-per-Lane Pluggable Optics for AI Fabrics
The proposed USFP form factor targets 400-gigabit-per-lane signals and up to 12.8 terabits per second in an OSFP-sized footprint.
AMD Frames Physical AI Around End-to-End Control Timing
AMD’s October 19 physical-AI paper argues that robotics infrastructure should be measured from sensor input to physical action.
Meta Details Data Center Cooling and Utility-Cost Practices
Meta’s October explainer describes closed-loop cooling, utility planning and its approach to paying for data-center infrastructure.
Cisco Extends Open Ethernet Systems for Distributed AI Factories
Cisco’s OCP 2026 release adds 1.6-terabit silicon, production systems and optical work for AI scale-out and scale-across networks.
AMD and Intel Align PyTorch CPU Optimization Around oneDNN
AMD and Intel will converge x86 CPU optimizations for PyTorch in oneDNN, with AMD integrating ZenDNN work.