Programmable data processing for network & security offload. AMD Pensando data processing units offload networking, security, and storage functions from the host CPU, freeing compute for application and AI workloads while improving network programmability.
Moves networking, encryption, and storage processing off the host CPU, freeing cycles for application workloads.
A fully programmable pipeline adapts to new protocols and policies without new hardware.
Encryption and security policy enforcement happen in the DPU itself, without a software tax on the host.
Talk to our infrastructure architects about how AMD Pensando DPUs fits into your data center or AI networking project — or see our full AMD partnership.
Talk to an Expert