Electricity serves as a fundamental resource for artificial intelligence, yet contemporary processing techniques are increasingly overwhelming data center operators' capacity to effectively manage their power grid interactions. This often compels them to reduce operational capacity by as much as 30%.
Nvidia CEO Jensen Huang highlighted this inefficiency at the company's annual GTC customer conference, stating, “There is so much power squandered in these AI factories.” The company further emphasized the financial impact, proclaiming during the presentation, “Every unused watt is revenue lost.”
Addressing this critical challenge, the startup Niv-AI has officially launched from stealth, securing $12 million in seed funding. Its mission is to resolve power inefficiencies by deploying innovative sensors for precise GPU power measurement and developing advanced tools for more efficient energy management.
Founded last year in Tel Aviv by CEO Tomer Timor and CTO Edward Kizis, the company has attracted significant backing from investors including Glilot Capital, Grove Ventures, Arc VC, Encoded VC, Leap Forward, and Aurora Capital Partners. Niv-AI chose not to disclose its current valuation.
Modern frontier laboratories, utilizing thousands of GPUs in parallel for training and running complex AI models, experience frequent, millisecond-scale power demand surges. These spikes occur as processors rapidly switch between intensive computation tasks and inter-GPU communication.
These rapid power fluctuations pose a significant challenge for data centers in effectively managing their electricity draw from the grid. To prevent power shortfalls, operators resort to either investing in temporary energy storage solutions or deliberately reducing their GPU utilization. Both approaches invariably diminish the return on investment from their costly high-performance chips.
Lior Handlesman, a partner at Grove Ventures and a board member at Niv-AI, underscored the urgency of the situation, stating, “We just can’t continue building data centers the way we build them now.”
Niv-AI's initial strategic step focuses on gaining a comprehensive understanding of these power dynamics. The company is currently deploying rack-level sensors capable of detecting GPU power usage at millisecond precision, both on its own hardware and in collaboration with design partners. This initiative aims to map the precise power profiles associated with various deep learning tasks, ultimately enabling the development of mitigation strategies that can unlock greater utilization from existing data center infrastructure.
Building upon this data, Niv-AI’s engineers naturally intend to develop an AI model. This model will be trained to predict and synchronize power loads throughout the data center, serving as an intelligent “copilot” for data center engineers.
Niv-AI anticipates having an operational system deployed in several US data centers within the next six to eight months. This solution holds particular appeal as hyperscale operators increasingly encounter challenges with land acquisition and supply chain disruptions when attempting to construct new data centers. The founders envision their ultimate product as a crucial, previously absent “intelligence layer” bridging the gap between data centers and the electrical grid.
“The grid is actually afraid of the data center consuming too much power at a specific time,” Timor explained to TechCrunch. He elaborated on the dual nature of the challenge: “The problem we’re looking at is a problem with two sides of the rope. One is to try to help the data centers utilize more GPUs, and hopefully make more of the power that they’re already paying for. On the other hand, you can also create much more responsible power profiles in between the data centers and the grid.”
The Editorial Staff at AIChief is a team of professional content writers with extensive experience in AI and marketing. Founded in 2025, AIChief has quickly grown into the largest free AI resource hub in the industry.