AMD Data Center

AMD Instinct software ecosystem spotlight 💡

@CausalFlow runs its inference optimization stack on AMD Instinct GPUs, helping enterprises deploy large open models in production with the throughput and cost efficiency that serving real traffic demands.

Together, CausalFlow and AMD are making it practical to move demanding inference workloads onto an open platform.

Explore CausalFlow and the growing AMD Instinct software ecosystem: amd.com/en/ecosystem/i…

As AI clusters scale, networking becomes just as critical as compute.

AMD Pensando Pollara 400 and Vulcano 800 AI NICs are designed to meet growing demands for open, programmable and efficient AI networking: bit.ly/4ijjBcN

The next frontier for AI video isn't just faster generation. It's creating experiences people can shape in real time.

See Nunchux AI's work with AMD Instinct MI355X GPUs and AMD ROCm software: bit.ly/4yRTJd5

At Advancing AI 2026, @rapt_ai demonstrated six different models sharing a single AMD Instinct MI300X GPU — automatically, using fine-grained compute-unit and memory slices, with no manual configuration.

The same GPU went from 278 to 650 tokens per second, a 2.4× increase, at full utilization with no failed workloads, helping teams serve far more work from the hardware they already have.

Watch the demo ⏬

The AMD Instinct software ecosystem continues to expand.

ClearML brings its AI Development Center and GPU orchestration platform to AMD Instinct GPUs, helping teams partition, schedule and share accelerators and deploy AMD Inference Microservices so more of every GPU hour goes to useful work.

Together, ClearML and AMD are giving AI teams tighter control over how compute is allocated across training and inference.

Explore @clearmlapp and the growing AMD Instinct software ecosystem: amd.com/en/ecosystem/i…