AI Developer Turns to On-Site Biomass to Secure Power in California

Posted

A newly announced acquisition in California is offering a different lens on one of the year’s most persistent operational challenges: how to secure reliable power for AI infrastructure without relying on already-constrained grids.

NewYork GreenCloud has acquired the Buena Vista Biomass Power plant in Ione, California, outlining plans to redevelop the site into what it describes as its first large-scale “carbon-negative AI Factory.” The project pairs on-site biomass generation with behind-the-meter, liquid-cooled AI computing infrastructure, allowing high-performance workloads to run independently of regional transmission capacity.

The existing 18-megawatt biomass plant is slated for expansion and conversion into a 41-megawatt facility, designed to supply continuous renewable power directly to an on-site AI data center capable of supporting next-generation GPU clusters.

Designing Around the Grid, Not Into It

Rather than seeking new interconnections or competing for limited capacity, the project is structured to bypass grid congestion altogether. Power generated at the site is intended to feed the AI factory directly, avoiding transmission bottlenecks and long interconnection queues that are slowing data center and industrial development across multiple states.

NYGC Chief Executive Officer Joe Church described the facility as the first step in what the company sees as a national platform of biomass-powered AI sites, emphasizing performance delivery alongside reduced carbon impact.

The redevelopment is being led jointly by NYGC and BucSha Energy, which specializes in biomass-to-pyrolysis systems. According to BucSha founder Dave Shaffer, the project’s design centers on using regional biomass resources to create a stable, dispatchable energy supply capable of supporting modern AI infrastructure.

So, Why Does This Matter?

For facilities, energy, and infrastructure leaders, the significance of the project is less about its “carbon-negative” branding and more about how it addresses power availability as an execution constraint.

As AI workloads drive rapid load growth, many operators are finding that even well-capitalized projects stall due to grid access delays, transmission upgrades, or regional congestion. Behind-the-meter generation—especially when paired with baseload resources—offers one way to decouple compute growth from grid timelines.

The Buena Vista project illustrates how legacy energy assets can be repositioned to serve new digital demand, particularly in regions where permitting, interconnection, or reliability challenges make traditional data center development more complex.

Financing and Replicability Questions

The acquisition was advised by Impact Capital Partners, which described the transaction as a potential blueprint for scalable, sustainable compute. Whether that model can be replicated at scale will depend on access to suitable biomass resources, permitting pathways, and local community acceptance—factors that vary widely by region.

What is clear is that AI infrastructure developers are increasingly exploring site-specific energy solutions rather than assuming grid expansion will keep pace with demand.

A Tier 2 Signal, Not a Template—Yet

The NYGC acquisition does not resolve the broader challenge of powering AI growth nationwide. Biomass availability, land use considerations, and regulatory scrutiny limit where such projects can be deployed. But it does reinforce a central theme emerging across sectors: growth is increasingly shaped by where infrastructure can be secured, not where demand exists.

Environment + Energy Leader