Niv-AI Targets AI Data-Centre Power Waste
Niv-AI, a Tel Aviv startup founded last year, emerged from stealth with $12 million in seed funding to reduce electricity waste in AI data centres. The company is deploying millisecond-level rack sensors and plans to train an AI model to predict and synchronize GPU power loads, aiming to mitigate millisecond demand spikes that force up to 30% GPU throttling or costly temporary storage. Niv-AI expects operational systems in a handful of US data centres within six to eight months.
Key Points
- 1Deploys millisecond-level rack sensors to measure GPU power spikes across thousands of GPUs.
- 2Addresses costly grid-management and throttling that force up to 30% GPU derating or paid storage.
- 3Enables data-centres to increase GPU utilization and present predictable, grid-friendly power profiles.
Scoring Rationale
Credible VC-backed startup addresses a clear industry pain, but solution remains early-stage and unproven at scale.
Sources
Public references used for this report.
Practice interview problems based on real data
1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems
