Niv-AI exits stealth with the goal of optimizing the power efficiency of GPUs. The company announced $12 million in seed funding to develop technologies for measuring and managing GPU power surges.

Optimizing GPU power consumption has become an increasing priority, especially in the context of training large models and on-premise inference, where energy costs and cooling requirements represent a significant fraction of the TCO. Technologies such as those proposed by Niv-AI could help reduce these costs while improving the overall performance of the systems.

For those evaluating on-premise deployments, there are trade-offs between CapEx and OpEx, and AI-RADAR offers analytical frameworks on /llm-onpremise to evaluate these aspects.