Google zinc battery storage project in West Virginia takes aim at the data center power crunch

Google has signed on to its first large-scale deployment of zinc-hybrid batteries through a $350 million clean energy portfolio in West Virginia built to feed round-the-clock power to a planned data center campus in the state. The Google zinc battery storage project, named Mammoth Solar, was announced September 2 by Google, solar developer MN8 Energy and battery manufacturer Eos Energy Enterprises. It will be built on a reclaimed coal mine site in Kanawha County and will sell into the PJM wholesale market serving the region.

MN8 Energy develops, owns and operates the facility while Google buys its output, giving the project a long-term anchor buyer. The portfolio is one integrated system rather than a standalone solar farm: 86 MW of utility-scale solar backed by 70 MW / 280 MWh of lithium-ion storage and 10 MW / 100 MWh of zinc-based storage. The decision to pair two chemistries with different discharge profiles, instead of building one larger battery bank, is the detail that defines the deal.

The configuration exists because solar generation and data center loads run on different clocks. Output stops at sunset while servers draw electricity through the night, and a battery that empties after four hours leaves the early morning uncovered. Google operates with a stated goal of 24/7 carbon-free energy on every grid where it has infrastructure, so its procurement has moved toward capacity available after dark rather than megawatt-hours that mostly arrive between mid-morning and mid-afternoon.

Inside the Mammoth Solar portfolio

Each storage technology in the design has a defined job. The lithium-ion system discharges for about four hours at full nameplate, which covers the evening ramp and the early part of the night. The Eos Z3 zinc-hybrid system, rated near ten hours of discharge, carries the load through the pre-dawn hours when the lithium bank is spent and solar output has not resumed. Together the two banks let MN8 sell availability rather than a narrow band of daytime energy.

The zinc bank is deliberately modest next to the solar array: 10 MW of capacity against 86 MW of generation. Its value is duration. A 100 MWh reservoir released over ten hours is built to close the overnight gap, the stretch a four-hour lithium system cannot bridge on its own, and that role is why the project counts as long-duration storage rather than another short-cycle battery addition.

This is the first commercial-scale long-duration energy storage deployment in West Virginia and the largest installation of Eos zinc-hybrid batteries to date. The companies put the capital investment at up to $350 million and expect about 200 jobs tied to the Kanawha County site, where land that once supplied coal now stores electricity for computing infrastructure.

West Virginia sits inside PJM, the wholesale market operator that coordinates electricity across much of the mid-Atlantic and Midwest, and Mammoth Solar serves that market for Google's regional digital infrastructure, including a data center facility the company plans to build in the state. The project is one response to accelerating electricity demand from data centers, which the Department of Energy has framed as a chance to speed up clean energy deployment. In its guidance on meeting that demand, the agency lists solar, land-based wind, battery storage and efficiency among the most rapidly scalable and cost-competitive near-term options, while pressing for grid modernization and affordable rates.

Why the Google zinc battery storage project is a template for the AI power crunch

Electricity has become the binding constraint on data center growth, and the constraint bites hardest after sunset. A facility earns nothing until it is connected to power, AI-oriented campuses draw electricity at a much faster rate than conventional server rooms, and transmission interconnection timelines can stretch for years. Securing generation and storage at a known site, as this deal does, is one way to shorten that wait.

The buying power behind such deals has turned Google and Amazon into two of the most consequential purchasers of battery storage in the world, funding gigawatt-scale portfolios that keep AI loads supplied around the clock. The consistent pattern in those portfolios is diversification: variable renewables, several storage chemistries, firm generation and flexible demand each take a share, with no single technology carrying the whole load. Google's other 2026 energy moves follow the same playbook, from geothermal development in Utah to a roughly 300 MW iron-air battery folded into a 1,900 MW wind and solar package for Xcel Energy, plus an accelerator cohort of 28 startups working on grid and data center power problems.

Not every hyperscaler power deal takes this route. Some buyers pursue behind-the-meter generation located at the data center itself, sidestepping interconnection queues entirely. Mammoth Solar instead sells through PJM, a bet that a utility-scale solar site with ten-hour storage can serve a facility across the grid as reliably as an on-site plant while still letting that capacity support other demand in the region.

The Google zinc battery storage project applies the same diversification logic inside one facility. Ten hours of zinc runtime is far short of the multi-day designs emerging elsewhere in long-duration storage, but it is more than twice the four-hour reach of the lithium bank and long enough to change what MN8 can sell: firm capacity for a 24/7 load instead of time-of-day energy. That distinction between selling electricity and selling availability is the commercial heart of the deal.

For Eos Energy Enterprises, the order is a commercial reference point, and for Google it is another first: no earlier Google-backed development has been built around zinc chemistry. The Z3 platform moves from smaller installations to a customer-backed deployment at procurement scale, and investors priced the news immediately: Eos shares climbed roughly 12 percent in pre-market trading after the announcement. Because the zinc-hybrid fleet has never operated at this scale, Mammoth Solar will also test how the chemistry holds up under the cycling demands of a data center load curve.

MN8 carries the development and operating risk, which makes the project a test of the financing model as much as the technology. A renewable developer that owns a solar plus long-duration storage asset, with a hyperscaler as anchor buyer, is a structure other merchant developers can apply across PJM states facing the same demand pressures. The reclaimed mine setting adds a local dimension: West Virginia's first commercial long-duration storage installation is also an example of what replaces coal capacity in the region's grid mix.

Why this matters

Google's backing of zinc-based batteries is a step change in how hyperscalers buy power for AI. Instead of renewable certificates or daytime generation, the buyer contracts firm, round-the-clock clean capacity assembled from solar, four-hour lithium and ten-hour zinc systems, with risk spread across a developer, a manufacturer and a technology company. That structure is repeatable in any utility region confronting data center load growth, and the Google zinc battery storage project doubles as an early test of whether reclaimed coal sites in states like West Virginia can host the energy infrastructure behind the next wave of compute.

Sources

Our new partnership brings long-duration energy storage to West Virginia.

Related Articles

• Google adds $1.5B for Alabama data center AI push
• Google and Intersect Power Launch Meitner Energy Center for AI Infrastructure in Texas
• Stratos Project Controversy: Utah Data Center Faces Backlash Over 9GW Power and Water Demands

✔Human Verified

Researched and cross-referenced against primary sources by the Bytevyte editorial team. This article was generated with the assistance of artificial intelligence and reviewed by the Bytevyte editorial team.