Niv-AI Emerges to Solve AI's Hidden Power Crisis With $12M

TEL AVIV, Israel – March 18, 2026 – As the artificial intelligence revolution accelerates, a silent crisis is brewing in the engine rooms of the digital world: data centers are running out of power. Today, Israeli deep-tech startup Niv-AI emerged from stealth with $12 million in funding to tackle this critical bottleneck, promising to unlock vast amounts of stranded energy and supercharge the world’s AI infrastructure.

Backed by a consortium of investors including Glilot Capital, Grove Ventures, and Arc VC, the Tel Aviv-based company is introducing its Power-Compute AI Stack. The platform is designed to solve the "Instantaneous Power Capacity" crisis, a fundamental challenge where the violent, millisecond-level power demands of AI processors clash with the physical limits of electrical grids, forcing operators to sacrifice performance for safety.

The Silent Crisis Strangling AI

The insatiable appetite of AI for computational power has a direct and increasingly problematic consequence: an equally insatiable demand for electricity. Modern Graphics Processing Units (GPUs), the workhorses of AI, consume energy not in a smooth, predictable flow, but in violent, erratic spikes. When switching between computational tasks, a rack of GPUs can surge its power draw by 50% or more in a fraction of a second.

This behavior is invisible to standard facility meters, leaving data center operators blind to the extreme stress placed on their infrastructure. To prevent catastrophic equipment damage or grid failures, they are forced to implement a brute-force solution: maintaining a massive power buffer. Industry experts and Niv-AI estimate this practice strands up to 30% of a data center's contracted electrical capacity. This means operators are paying for power they cannot safely use, a costly inefficiency that can amount to hundreds of millions of dollars per facility annually and acts as a direct brake on AI scaling.

The problem is systemic and escalating. Industry analysts at Gartner predict AI will drive a 160% increase in data center energy consumption over the next two years. The International Energy Agency projects that by 2026, electricity demand from data centers could double from 2022 levels. As next-generation GPUs push rack power densities from tens of kilowatts toward the megawatt range, this power bottleneck threatens to become the single greatest limiting factor for AI's growth.

An 'Electrical Fingerprint' to Tame the Spikes

Niv-AI proposes a more elegant solution than simply building more power plants or throttling expensive GPUs. The company's Power-Compute AI Stack acts as an intelligence layer, creating what its CEO calls a "new dialect between energy and computation."

The system begins by deploying high-frequency sensors that capture the unique "electrical fingerprint" of AI workloads in real-time. This provides an unprecedented, high-resolution view of power consumption, revealing the millisecond-level transients that were previously invisible. This granular data is then fed into the platform's AI models, which learn to predict the timing and magnitude of incoming power spikes.

Armed with this predictive insight, the system acts as a high-speed traffic controller for computational tasks. It actively orchestrates workloads, introducing micro-second staggers to smooth out the collective power demand across the facility before the dangerous spikes can occur. This active management allows data centers to safely operate much closer to their true capacity limits, unlocking the stranded power and maximizing the utilization of their existing GPU fleets.

"The AI industry has hit a physical wall, and the industry's current answers, batteries, capacitors, or simply curtailing GPU activity, are not enough," said Tomer Timor, CEO of Niv-AI. "We built observability and orchestration layers that the grid and datacenter must have, and we use them to give data centers back the performance they are already paying for."

A New Control Plane for Power

Investors see Niv-AI's software-driven approach as a game-changer for the economics of AI infrastructure. Rather than requiring costly and slow-to-deploy physical upgrades, the platform promises to extract significant value from assets that are already in place.

"We see Niv-AI as the foundational control plane for data center power," stated Arik Kleinstein, Founding Partner at Glilot Capital, one of the lead investors. "While an incredibly high-resolution understanding of the AI factory's heartbeat is required, the true value lies in acting on it. Niv-AI doesn't just observe the data; they actively orchestrate the workloads to solve the industry's biggest bottleneck of Instantaneous capacity."

This sentiment is echoed by other key backers who recognize the urgency of the problem. The ability to increase GPU density and compute output without triggering massive new capital expenditures on power infrastructure represents a powerful value proposition for data center operators and the hyperscale cloud providers they serve.

Lior Handelsman, Managing Partner at Grove Ventures and a co-founder of solar energy giant SolarEdge, drew a parallel to a previous energy revolution. "Niv-AI brings an unmatched level of visibility and control to an infrastructure that is desperate for it," he noted. "The same combination of physics, deep-tech, and smart software that revolutionized solar energy at SolarEdge is now being unleashed on the data centers of the AI era."

From Elite Tech Units to the Data Center Floor

The company's ability to tackle such a complex, physics-based problem is rooted in its founding team. Comprising veterans of Israel's elite intelligence and technological special forces, the team, led by CEO Tomer Timor and CTO Edward Kizis, brings a rare blend of expertise in applied physics, electrical engineering, and the architecture of complex, mission-critical systems. This background provides the deep-tech rigor required to build a reliable and effective solution for the unforgiving environment of a modern data center.

Having secured its $12 million in funding in under eight months since its founding in 2025, Niv-AI is moving aggressively to scale its operations. The company is already collaborating with design partners to integrate its platform into real-world GPU environments. The new capital will be used to rapidly expand its engineering team and build out its deep-tech research lab in Tel Aviv, which is designed to simulate AI workload fingerprints under a wide range of real-world electrical conditions.