Data Centers Are Becoming Power Plants: NJ Battery Project Proves It | Taha Abbasi

Taha Abbasi explains why the convergence of data centers and energy storage is creating an entirely new category of infrastructure — and why it matters for the future of the grid.

Data centers are no longer just consumers of electricity — they are becoming producers and storage systems in their own right. Calibrant Energy has signed a definitive agreement with Iron Mountain to build, own, and operate a 23-megawatt-hour (MWh) battery at Iron Mountain’s New Jersey data center. This project, announced on February 19, 2026, represents a paradigm shift in how we think about large-scale computing facilities.

The Data Center Energy Problem

The explosion of artificial intelligence workloads has created an unprecedented demand for data center capacity — and for the electricity to power it. AI training runs for large language models like GPT-5, Grok, and Claude can consume megawatts of power continuously for months. Data center operators are building facilities faster than utilities can provide grid connections, creating a bottleneck that threatens to slow AI development.

This energy crunch has driven data center operators to think differently about their relationship with the electrical grid. Rather than simply drawing power from utilities, companies like Iron Mountain are installing battery storage systems that can absorb excess grid power during low-demand periods and discharge it during peak hours — or even sell it back to the grid.

As Taha Abbasi explains, this transforms the data center from a pure electricity consumer into a grid asset. A 23 MWh battery system is not just a backup power source — it is a virtual power plant that can provide grid services like frequency regulation, peak shaving, and demand response.

How Battery Storage Changes the Economics

The economics of co-locating batteries with data centers are compelling. Data centers already have the electrical infrastructure — transformers, switchgear, grid connections — needed to handle large power flows. Adding battery storage leverages that existing infrastructure rather than duplicating it. The batteries can also reduce the data center’s peak demand charges, which often represent 30-50% of their total electricity costs.

Taha Abbasi draws a parallel to Tesla’s Megapack business, which has been deploying grid-scale battery systems at an accelerating pace. The technology is the same — lithium-ion battery cells arranged in containerized systems — but the application is different. Tesla’s Megapacks are typically deployed by utilities or independent power producers. Projects like Calibrant’s put the same capability directly at the point of highest demand.

The AI-Energy Nexus

The irony is not lost on industry observers: AI companies that are building systems to optimize energy usage are themselves among the largest energy consumers on the planet. xAI’s Memphis supercluster, Google’s data centers, and Microsoft’s Azure facilities collectively consume more electricity than many small countries.

But this convergence also creates opportunity. AI can optimize battery charging and discharging cycles to maximize economic value. Machine learning algorithms can predict grid demand patterns with far greater accuracy than traditional methods. And as AI workloads become more flexible — training runs can be paused and resumed, inference can be load-balanced across facilities — data centers can become more responsive to grid conditions.

As Taha Abbasi sees it, the data center of the future is not just a building full of servers. It is an integrated energy system that generates, stores, and intelligently distributes power while simultaneously running the world’s most demanding computational workloads. The Iron Mountain project in New Jersey is an early example of this convergence — but it will not be the last.

Implications for the Broader Energy Transition

If data centers become significant grid storage assets, it could accelerate the adoption of renewable energy. Solar and wind power’s biggest limitation is intermittency — they produce power when the sun shines and wind blows, not necessarily when demand peaks. Distributed battery storage at data centers could absorb renewable overproduction and discharge it when needed, effectively smoothing out the renewable energy supply curve.

Taha Abbasi believes this model could scale rapidly. With hundreds of hyperscale data centers planned or under construction globally, each with its own battery system, the aggregate storage capacity could rival purpose-built grid storage projects. The AI industry’s insatiable demand for power might just be the catalyst that solves one of the energy transition’s most persistent challenges.

About the Author: Taha Abbasi is a technology executive, CTO, and applied frontier tech builder. Read more on Grokpedia | YouTube: The Brown Cowboy | tahaabbasi.com