Nvidia reveals next-gen DPU to help offload gigascale AI infrastructure
BlueField-4 and new ConnectX-9 SuperNIC set to launch as part of the 2026 Vera Rubin slate
BlueField-4 and new ConnectX-9 SuperNIC set to launch as part of the 2026 Vera Rubin slate
Frontier successor ditches uniform rack requirements to enable mixed configurations in single racks
New AMD-powered 6U system promises 95% cable reduction and 70% space savings over traditional servers
The Open Chiplet Atlas ecosystem brings together 50 organizations to create royalty-free standards for multivendor chiplet integration
Offering includes Palo Alto Networks PQC software
CTran and NCCLX releases look to challenge Nvidia's NCCL dominance
New G4 VMs leverage RTX Pro 6000s to support AI workloads up to 100 billion parameters
Hyperscaler couldn’t fit its custom network cards in Nvidia’s racks. So it built a sidecar
Also, Oracle first to deploy AMD's MI355X GPUs in OCI Supercluster
Tech giant reveals networking backbone for Prometheus, its multi-gigawatt AI clusters
New network interface card doubles Thor performance and adds congestion control for clusters scaling beyond 100,000 XPUs
Trial with Silicon Quantum Computing showed quantum system capable of predicting network issues as accurately as deep learning, but far faster
Highly anticipated DGX Spark aims to bring petascale AI computing directly to office desks with some 200 Gb/s of network bandwidth
ESUN working group brings together AMD, Nvidia, Cisco, and others to develop open standards for connecting GPUs within AI clusters
ChatGPT developer will deploy Broadcom's networking solutions in its racks alongside jointly developed GPUs
Partnership with Cornell and Stanford pushes beyond paradigm where photonic devices perform only single, fixed tasks
Working groups will examine power, cooling, and physical design specifications to enable interoperable AI facilities and avoid costly retrofits
Chip giant may take a leaf from FlexRAN positioning
Chip giant's networking technologies help provide bidirectional bandwidth needed to outpace AMD’s MI355Xs
Guardium Cryptography Manager automatically detects gaps in unsecure enterprise data
Chip giant's third-generation co-packaged optics aims to power hyperscale AI infra to the next level
Firms warns of data threat to passive optical networks
Startup spun-out of Caltech in 2023
Tech giant to integrate its Omnia software stack with the consortium's tools to improve provisioning and management of high-performance computing systems