Google's $1B Iron-Air Battery Powers AI Data Centre

Google Bets $1 Billion on Form Energy's Iron-Air Battery to Power Its AI Data Centre in Minnesota

The race to power artificial intelligence infrastructure with clean, reliable energy has just hit a landmark moment. Google has officially committed to a historic $1 billion deal with Form Energy, a long-duration battery storage startup, to procure what is being described as the world's largest battery system ever announced by energy capacity. The agreement, tied to a new data centre in Pine Island, Minnesota, is not just a story about real estate or renewable energy — it is a defining chapter in AI funding news and the broader conversation about how the tech industry plans to sustainably power its exponentially growing computational demands. As AI systems scale from simple chatbots to fully autonomous reasoning engines, the pressure on data centre infrastructure has never been greater, and this deal signals that tech giants are now willing to make billion-dollar bets on entirely new categories of energy technology.

The project is being executed in partnership with Xcel Energy, one of the country's largest electricity providers, under a sweeping Electric Service Agreement that combines wind, solar, and long-duration iron-air battery storage into a single, vertically integrated clean energy solution. The goal is straightforward but ambitious — to keep Google's core digital services, including Search, YouTube, Google Maps, and Workspace, running continuously on carbon-free electricity while simultaneously ensuring that existing residential and commercial customers on Xcel's grid do not face higher bills as a result of Google's massive additional energy demand. This balance between corporate sustainability goals and public utility responsibilities makes the deal structurally unique, and it sets a template that other AI companies hunting for stable green power may soon look to replicate.

The $1 Billion Battery Order That Changed Everything

At the heart of this landmark announcement is Form Energy's iron-air battery system, a 300-megawatt installation capable of discharging electricity continuously for up to 100 hours — which translates to a total energy capacity of 30 gigawatt-hours. To put that figure in perspective, most large-scale battery deployments today measure in the range of a few hundred megawatt-hours at best. A 30 gigawatt-hour system is not just a step forward — it is an entirely different league of energy storage, and it has never been deployed at this scale anywhere in the world before this announcement. The fact that Google is the customer financing this historic first deployment through a roughly $1 billion purchase order from Form Energy places this squarely in the conversation about the most consequential AI funding decisions of the decade.

Form Energy's iron-air battery technology operates on a surprisingly elegant electrochemical principle. When the battery is discharging, oxygen from the surrounding air is drawn into the system, where it reacts with iron particles, causing them to rust. This rusting process releases electrons, which flow through the circuit as usable electrical current. When the battery needs to recharge, the process is simply reversed — electricity is applied to the iron oxide, stripping out the oxygen and restoring the iron to its original metallic state. This cycle can repeat thousands of times, and because iron is one of the most abundant and inexpensive materials on the planet, the cost economics of iron-air batteries are potentially far more attractive than lithium-ion alternatives at multi-day durations. The technology does not require rare earth minerals, does not pose the thermal runaway risks associated with lithium batteries, and is designed from the ground up to operate over days rather than hours — which is precisely what AI data centres need when renewable generation dips due to weather events or grid stress.

Form Energy has been developing this technology for several years and has already built a dedicated manufacturing facility in Weirton, West Virginia, specifically designed to produce these batteries at scale. However, before this Google deal, the company had not secured a single commercial customer at anywhere near this level of deployment. The $1 billion commitment from Google is therefore transformational, not just financially but also as a validation signal to investors, regulators, and the broader energy market that iron-air technology is ready for real-world deployment at utility scale.

Form Energy's Funding Journey and What Comes Next

The latest AI funding news surrounding Form Energy extends well beyond the Google battery purchase itself. Following the announcement of the Xcel Energy and Google agreement, Form Energy's CEO Mateo Jaramillo moved quickly to capitalise on the momentum, reportedly raising a new $500 million funding round to accelerate manufacturing and operational expansion. This fresh capital raise comes on top of approximately $1.4 billion that the company has already secured from investors over its lifetime, according to data from PitchBook. With the combined weight of venture capital backing, the $1 billion Google order, and a now-proven commercial pathway, Form Energy is positioning itself for an initial public offering expected to take place next year.

The timing of this AI funding trajectory is no accident. As AI infrastructure investment continues to surge globally, energy storage has emerged as one of the most critical bottlenecks slowing down the deployment of new data centres. Hyperscale operators like Google, Microsoft, and Amazon are under growing pressure from shareholders, regulators, and the public to demonstrate that their AI ambitions are not coming at the cost of carbon targets. Long-duration energy storage — the kind that can smooth out renewable generation over periods of days rather than minutes — is the missing piece that makes 24/7 carbon-free electricity actually achievable rather than just a marketing promise. Form Energy's ability to secure this AI funding at such a massive scale tells us that investors now believe the technology is past the experimental stage and entering a phase of commercial scaling.

For the AI World community tracking developments at the intersection of artificial intelligence, energy, and venture capital, this story represents a crucial data point. The companies that win the AI infrastructure race will not just be those with the best models or the fastest chips — they will be the ones that can reliably access affordable, sustainable power at the scale that modern AI workloads demand.

Xcel Energy's Role and the Clean Energy Architecture

Xcel Energy's involvement in this project is more than that of a passive electricity supplier — the utility is actively co-designing the clean energy portfolio that will serve the Pine Island, Minnesota data centre. The full energy mix agreed upon under the Electric Service Agreement includes 1,400 megawatts of wind power, 200 megawatts of solar generation, and the 300-megawatt Form Energy iron-air battery system for long-duration storage. This combination is deliberately engineered to address the different temporal characteristics of each renewable source. Wind generation tends to be stronger at night and during certain seasons, while solar generation peaks during daylight hours. Neither source is inherently reliable around the clock, which is where the 100-hour battery becomes indispensable — it acts as an enormous reservoir that can absorb surplus renewable generation when it is abundant and release it steadily during periods of low wind or sunlight, including multi-day weather events that would otherwise create serious reliability problems.

Beyond the generation and storage components, Google has also committed $50 million to Xcel Energy's Capacity*Connect Program, a grid reliability initiative that benefits all customers on the network. This investment is part of a broader framework in which Google agrees to cover all infrastructure and capacity costs associated with its own power demand, effectively ring-fencing the financial impact of the data centre build-out so that it does not flow through to household or business electricity bills. This is a significant structural concession on Google's part, and it reflects the political and regulatory realities of operating large data centres in communities where utility customers are already sensitised to energy price increases. By insulating existing ratepayers from any cost impact, Google and Xcel are attempting to build social licence for what is, by any measure, an extraordinary concentration of new electricity demand in a single location.

The Electric Service Agreement has been submitted to the Minnesota Public Utilities Commission for regulatory approval, which is a standard requirement for utility-scale agreements of this nature in the state. The outcome of that review will determine the final timeline for construction and commissioning, though both parties have expressed confidence in the agreement's alignment with Minnesota's clean energy policy objectives.

Why AI Data Centres Need Long-Duration Storage

The specific energy challenge that AI data centres face is fundamentally different from what other large industrial electricity consumers have historically dealt with, and understanding that difference helps explain why AI funding news in the energy storage sector has been accelerating so dramatically. A conventional factory or commercial building has an electricity load profile that varies somewhat by time of day and season, but is broadly predictable and interruptible — it can scale down operations during grid stress events without catastrophic consequences. An AI data centre running large-scale inference workloads for millions of simultaneous users has essentially no tolerance for interruption. If the power goes out, the training run is lost, the inference pipeline fails, and users experience service degradation. This means that AI data centre operators need a level of power continuity that goes far beyond what standard backup generation or even short-duration lithium-ion batteries can provide.

The Google-Form Energy partnership addresses this challenge head-on. With a battery capable of discharging at 300 megawatts for 100 consecutive hours, the Pine Island data centre will be able to maintain full operations through almost any foreseeable grid disruption or renewable energy shortfall. Amanda Peterson Corio, who leads data centre energy strategy at Google, has been explicit about the company's philosophy here: integrating carbon-free energy with long-duration storage is not just a sustainability exercise — it is fundamentally about building a more resilient system that benefits both the company's operational requirements and the local communities that share the grid. This dual mandate of reliability and sustainability is becoming the standard expectation for new AI infrastructure deployments, and companies that can deliver on both simultaneously will have a significant competitive advantage in the years ahead.

The broader AI funding news landscape reinforces this direction. Investment into grid-scale storage, renewable energy procurement, and clean infrastructure for AI has been growing year over year, and the Google-Form Energy deal may well catalyse a new wave of commitments from other hyperscale operators who are watching closely. For startups working on next-generation battery chemistry, grid software, or energy management systems, the signal from this $1 billion deal is unmistakably clear: the market for clean, long-duration energy storage at AI scale is real, it is large, and it is willing to pay.

What This Means for the Future of AI Infrastructure Investment

Stepping back from the technical details, the Google-Form Energy-Xcel Energy partnership carries implications that extend well beyond a single data centre in Minnesota. This deal marks the first time a hyperscale AI operator has made a direct, billion-dollar investment in a novel battery technology that has never been commercially deployed at scale before. It is, in essence, a massive real-world pilot programme funded by one of the world's most financially powerful companies, and its success or failure will be watched extremely closely by every major player in the AI infrastructure ecosystem.

If the iron-air battery system performs as designed — delivering reliable, multi-day storage at the cost economics that Form Energy has projected — it could unlock a pathway for far more aggressive clean energy procurement by AI companies in locations where renewable generation is abundant but grid reliability has historically been a limiting factor. Regions like the American Midwest, parts of India, and large sections of Europe with strong wind resources but underdeveloped storage infrastructure could become far more viable as AI data centre destinations if long-duration battery technology proves out at this scale. This geographic unlocking of AI infrastructure potential is itself a story worth tracking closely in AI funding news over the coming months and years.

At the same time, the funding trajectory of Form Energy — from a well-backed but pre-commercial startup to a company with a $1 billion customer order, a $500 million funding round in progress, and an IPO on the horizon — is a powerful reminder that the AI era is creating enormous opportunities not just in software and model development but across the entire physical infrastructure stack that makes AI possible. From chip fabrication to cooling systems, from power generation to long-duration storage, the AI funding wave is reshaping the energy and industrial sectors in ways that are only beginning to become visible. The AI World Organisation continues to track these developments as they unfold, providing analysis and context for the communities of investors, technologists, and policymakers navigating this rapidly evolving landscape.