AI data centers are consuming electricity at a pace that is reshaping the power market, and many of the largest facilities are still turning to natural gas rather than clean energy. A startup called Firma Power argues that the problem is not cost or reliability, but the difficulty of assembling enough clean power from a fragmented landscape of wind, solar, and battery projects. The company has built software that it says can change how these massive facilities secure electricity, potentially opening a faster path to clean energy for the AI boom.

The scale of the challenge is unusual. A typical data center once needed around 50 megawatts of power. AI-focused facilities now demand 500 megawatts or more, roughly the amount of electricity a city like San Francisco uses. That jump has pushed utilities, grid operators, and technology companies into a scramble for new generation, and gas plants have often filled the gap because they can be planned, permitted, and financed through familiar channels.

The power problem behind AI’s gas habit

Firma Power’s central insight is that clean energy is not scarce. The United States is about to add a record 86 gigawatts of new wind, solar, and battery capacity this year, according to the company’s framing of the market. That pipeline is large enough to support a significant share of new data center demand. Yet many individual projects remain stalled because developers cannot find buyers, or because they get tangled in grid interconnection and permitting issues.

That mismatch is the heart of the problem. A data center needs a reliable, large-scale supply of power. Clean energy developers need creditworthy buyers who can support their projects. In theory, the two sides should meet easily. In practice, the deals are fragmented across multiple projects, contracts, and timelines. A single wind farm or solar array may be too small or too intermittent to serve a hyperscale facility on its own. Batteries can help, but they add another layer of contracting and coordination.

Firma cofounder Jesse Jenkins says almost everything sitting in the current pipeline is wind, solar, or storage, yet gas still wins by default. The reason is not that clean energy is too expensive or unreliable, in the company’s view. It is that developers and data center operators struggle to put together a portfolio that is large enough, coherent enough, and bankable enough to move forward.

From 50 megawatts to city-scale demand

The change in data center size explains why the old procurement model is under strain. A 50-megawatt facility could be served through a relatively straightforward utility arrangement or a single large power purchase agreement. A 500-megawatt AI campus is a different kind of customer. It needs power at a scale that can require multiple generation projects, new transmission capacity, and a carefully managed schedule for coming online.

That comparison matters because AI training and inference loads are not small. They run continuously, and they can ramp quickly as new hardware is installed. Data center operators want certainty that power will be available when servers arrive. Utilities want to avoid building expensive infrastructure that may become stranded. Clean energy developers want long-term contracts that make their projects financeable. Firma’s pitch is that software can reconcile these needs faster than the traditional project-by-project approach.

Why clean energy loses by default

The clean energy pipeline is broad, but it is also messy. Wind and solar projects are proposed in different locations, with different interconnection queues, different permitting timelines, and different output profiles. Batteries can smooth some of that variability, but they are not a generation source on their own. To serve a large data center, a buyer may need to combine several projects into a single supply strategy.

That is where developers often get stuck. A project may have land, equipment, and a promising grid connection, but no buyer. Another may have a buyer but face delays in interconnection. A third may be economically attractive but too small to matter to a hyperscaler. Firma Power’s software is designed to look across proposed projects, calculate the capacity a data center needs, and identify combinations that could work.

The company says its platform, FirmaOS, can search for proposed projects and match them to a data center’s needs in minutes instead of months. That speed is not just a convenience. In a market where utilities and developers plan years ahead, a faster matching process can determine whether a clean energy portfolio comes together before a gas alternative locks in.

How FirmaOS assembles clean power portfolios

Firma Power’s answer is to treat clean power procurement as a portfolio problem rather than a single-project purchase. The software calculates the capacity needed and identifies clean-energy projects that could provide it. It then helps the company negotiate with developers and hand the customer one contract instead of several messy ones.

Matching projects in minutes, not months

A data center operator does not want to manage a dozen separate negotiations with wind, solar, and battery developers. FirmaOS aims to do the matching work up front, finding projects that fit the facility’s load profile, location, and timeline. The company says this can bring power online faster and at a competitive cost compared with gas plants.

One contract across several projects

The bundled contract is a key part of the model. Instead of forcing a customer to stitch together multiple agreements, Firma Power negotiates with developers and presents a single package. That can reduce legal and administrative friction, make financing easier, and give the data center a clearer path to clean power. The process wraps up in days, according to the company, while waiting on a new gas plant or a utility’s planning cycle can take years.

The approach also helps connect clean-energy developers with the buyers they need to support their projects. Many proposed wind, solar, and battery installations are economically viable only if they can secure long-term offtake. By aggregating demand from large data centers, Firma Power could give developers a more reliable route to market.

Founders, funding, and early traction

Firma Power was founded by energy experts Jesse Jenkins, Scott Burger, and Brent Alderfer. The startup has raised about $1.4 million in February 2026. It is now working on roughly a dozen projects with hyperscalers and AI cloud companies, according to the company.

That early traction matters because hyperscalers are among the most demanding power customers in the world. They need enormous volumes of electricity, and they often have public clean energy goals. At the same time, they face pressure to bring AI capacity online quickly. If Firma Power can show that clean energy portfolios can be assembled at the speed and scale those customers need, its model could spread beyond a handful of pilot projects.

What the startup says it can unlock

Firma Power estimates that just three grids alone could add 200 gigawatts of clean energy by 2030, which is more than what the entire country needs. That estimate is a signal of how much proposed clean energy capacity is waiting for buyers and grid access. The company’s bet is that better matching software can unlock some of that capacity for AI data centers.

The broader context is a power system under pressure. AI demand is rising quickly, and utilities are revising load forecasts upward. Gas plants are being proposed to meet the surge, even as clean energy costs have fallen. Firma Power’s argument is that the clean energy is already in the pipeline; the missing piece is a faster, more standardized way to connect it to the customers who need it.

The company says its platform can reduce the time and complexity of clean power procurement, giving data centers a competitive alternative to gas. If that works at scale, the next wave of AI infrastructure may not have to default to fossil fuels simply because clean power is harder to assemble. Firma Power is now working on roughly a dozen projects with hyperscalers and AI cloud companies, and it estimates that three grids alone could add 200 gigawatts of clean energy by 2030.

Key facts at a glance

• AI data centers can require 500 megawatts or more, compared with about 50 megawatts for a typical older data center.
• The United States is set to add a record 86 gigawatts of new wind, solar, and battery capacity this year.
• Firma Power says clean energy is not scarce, but clean projects are hard to assemble into large, reliable portfolios.
• The startup’s FirmaOS platform matches proposed clean energy projects to data center needs in minutes instead of months.
• Firma Power negotiates with developers and offers customers one contract instead of multiple agreements.
• The company was founded by Jesse Jenkins, Scott Burger, and Brent Alderfer.
• It raised about $1.4 million in February 2026 and is working on roughly a dozen projects with hyperscalers and AI cloud companies.
• Firma estimates that three grids alone could add 200 gigawatts of clean energy by 2030.

The startup’s model depends on turning a fragmented clean energy pipeline into a set of bankable portfolios that large power buyers can accept. Its software calculates capacity needs, identifies proposed wind, solar, and battery projects, and bundles them into a single contract. The company says that process can be completed in days, while a new gas plant or utility planning cycle can take years.

Source:Digital Trends News