As compute gets increasingly powerful, the fact of the matter is: most AI workloads do not require the entire capacity of a single GPU. Computing power required across the model development lifecycle looks like a normal bell curve – with some compute required for data processing and ingestion, maximum firepower for model training and fine-tuning, and stepped-down requirements for ongoing inference.

We are excited to announce the launch of our AI Infrastructure Control Plane, designed as a universal operating system for AI infrastructure. With this launch, we make it easier for IT teams and DevOps to gain ultimate control over their AI Infrastructure, manage complex environments, maximize compute utilization, and deliver an optimized self-serve experience for their AI Builders.