Niv-AI raises $12 million to train AI on data center power loads
Niv-AI, a startup founded by Tomer Timor and Edward Kizis, has secured $12 million in seed funding from investors including Glilot Capital, Grove Ventures, Arc VC, Encoded VC, Leap Forward and Aurora Capital Partners.
Its big goal? Fix the huge energy waste in data centers by using smart sensors to track exactly how much power GPUs are using, so performance doesn't get slowed down when electricity demand spikes.
Niv-AI is already deploying rack-level sensors
Niv-AI is already deploying rack-level sensors and expects an operational system in a handful of US data centers in six to eight months.
The idea is to collect detailed power data and train AI models that can predict and balance power loads, kind of like giving the grid a brain.
This is timely too: NVIDIA's CEO recently warned that growing electricity needs could hurt GPU performance, so Niv-AI's tech could help keep things running smoothly while cutting down on wasted energy.