Credo to Supply Optical Interconnects for TensorWave’s AMD-Based AI Clusters
Credo Technology Group will integrate its multi-rate co-packaged optical interconnect modules into TensorWave’s next-generation AMD-based AI clusters to boost data throughput and cut latency for machine-learning workloads. The collaboration targets rack-scale deployments using AMD EPYC processors and Instinct GPUs in hyperscale data centers.
1. Partnership Overview
Credo Technology Group and TensorWave have entered a strategic collaboration to integrate Credo’s optical connectivity solutions into TensorWave’s upcoming AI cluster platforms. The agreement formalizes joint engineering efforts to deliver fully optimized hardware stacks for enterprise and hyperscale AI deployments.
2. Technology Integration
Credo will supply its multi-rate, co-packaged optical interconnect modules supporting 200Gbps and 400Gbps data rates, designed to interface directly with AMD EPYC CPUs and AMD Instinct GPUs. TensorWave will embed these modules at the rack level to achieve higher link density, lower power consumption and reduced end-to-end latency.
3. Market Impact
The collaboration positions both companies to capitalize on the surging demand for high-performance AI infrastructure in cloud and on-premise data centers. By combining Credo’s connectivity expertise with TensorWave’s system design, the partnership aims to accelerate customer deployments and drive incremental revenue growth for Credo.