NVIDIA Rubin architecture maximizes AI inference throughput via co-design
NVIDIA details five hardware and software strategies in its Rubin architecture that boost AI factory efficiency by optimizing memory usage for KV caches and model weights.
Kubernetes, GPUs, DevOps, observability and keeping production healthy.
NVIDIA details five hardware and software strategies in its Rubin architecture that boost AI factory efficiency by optimizing memory usage for KV caches and model weights.
OpenShell, a new CNCF Sandbox project from NVIDIA, provides fine-grained runtime controls for AI agents on Kubernetes using Agent Sandbox and supervisor-based policy enforcement.
Cloudflare reduced DNS propagation latency by 50% and memory usage by 85% by replacing full zone rebuilds with targeted builds for specific record changes.
NVIDIA releases an open-source system that correlates logs, metrics, and scheduler state to identify root causes in distributed GPU workloads.
NVIDIA releases an open-source tool that provides a common structural description for distributed AI workloads on Kubernetes, eliminating the need for framework-specific adapters.
A deep dive into how memory latency stalls CPUs during database aggregations and how radix partitioning improves cache utilization to boost performance.
A federal judge denied xAI's motion to dismiss a lawsuit alleging the company operated gas turbines at its Colossus data center without required air permits.
Nvidia is negotiating to guarantee billions in construction loans for Nscale, securing a major Anthropic contract amid growing investor skepticism about AI infrastructure costs.
Google has activated its first space data center prototype, running Gemma 4 on a TPU to test radiation effects and orbital inference capabilities.
The CNCF Sandbox project OpenChoreo integrates the OpenTelemetry Collector to provide distributed tracing and long-term storage for Kubernetes events within its Internal Developer Platform.
Oracle engineers outline a five-layer architecture for isolating autonomous AI agents on Kubernetes, balancing security and performance using runtimes like gVisor and Kata Containers.
Volcano v1.16 introduces NamespaceQueue for tenant isolation, a fairshare plugin for historical usage tracking, and performance optimizations for large batch workloads.
Kubernetes control planes face hidden failure paths where valid mechanisms like informers, schema aggregation, and liveness probes amplify local defects into system-wide outages.
NVIDIA CUDA 13.4 introduces transparent memory enforcement for GPU sharing. The new GPU Fractioning Operator and KAI Scheduler v0.18.0 bring these capabilities to Kubernetes clusters.
SpaceX has committed to a $40 billion financing deal with Apollo Global Management to lease Nvidia GPUs, marking the largest committed GPU financing to date.