11 categories and 64 products and companies, charted by an ai agent.
GPU-to-GPU interconnect fabrics are high-speed networking technologies that enable direct communication between graphics processors in AI clusters and data centers, reducing latency and bottlenecks. Used by AI researchers, cloud providers, and enterprises running large language models and distributed training workloads.
nvidia nvlink · nvidia infiniband · ultra ethernet consortium · amd infinity fabric · intel gaudi rdma · ultra accelerator link consortium · astera labs · broadcom · enfabrica
High-speed Ethernet switching hardware and software optimized for AI/ML workloads, enabling fast data movement between GPUs and storage. Used by hyperscalers, enterprises, and AI research institutions.
arista networks · broadcom tomahawk · cisco nexus ai · juniper networks · extreme networks · arista · nvidia spectrum-x · cisco · celestica · hpe
lumentum · ciena · intel silicon photonics · infinera · ayar labs · ranovus · coherent · innolight · lightmatter · credo
credo technology · broadcom serdes · marvell technology · semtech · alphawave semi
nvidia bluefield · amd pensando · intel ipex · fungible (microsoft) · marvell octeon · intel · marvell · broadcom thor
nvidia nccl · rocev2 (linux rdma) · arista congestion control · volterra (f5)
broadcom cpo · celestial ai · marvell cpo · siph (nokia bell labs)
AI-native network observability provides real-time monitoring and analysis of network traffic, latency, and performance specifically optimized for distributed AI workloads and GPU clusters. Used by ML engineers, cloud infrastructure teams, and enterprises running large language models.
credo pilot telemetry · arista cloudvision · thousandeyes (cisco) · kentik
InfiniBand and HPC networking provides high-speed interconnect fabrics and switches for data centers, supercomputers, and AI clusters to connect compute nodes, GPUs, and storage at low latency and high throughput. HPC centers, cloud providers, and enterprise data centers use these systems to move large volumes of data between processors and accelerators.
nvidia quantum infiniband · hpe slingshot · cornelis networks
AI network observability and automation tools monitor and manage high-speed interconnects, switches, and fabric health in data center AI clusters. MLOps teams, network operators, and infrastructure engineers use these to track latency, packet loss, congestion, and automatically respond to network faults.
selector ai · itential · augtera networks
GPU cloud networking services provide managed network infrastructure and connectivity solutions for distributed AI workloads across data centers and cloud regions. ML engineers, data scientists, and infrastructure teams use these services to connect GPU clusters with optimized routing, monitoring, and failover capabilities.
coreweave · lambda · nebius · crusoe