Gigascale Capital Raises $250 Million Climate Fund as AI Power Demand Reshapes Energy Investment

Gigascale Capital, the venture firm founded by former Meta CTO Mike Schroepfer, closed its first institutional fund at $250 million, TechCrunch reported. The fund targets early-stage companies in energy, grid infrastructure, and critical minerals, an investment thesis built partly on the premise that AI compute demand is making climate technology a market-performance play rather than a virtue signal.

The Thesis

Schroepfer’s argument is that clean energy wins when it outperforms incumbents on cost and speed, not when it appeals to conscience. “Solar went from 40 gigawatts a year to 600 in a decade because it got cheaper, not because it got more virtuous,” Schroepfer said in a statement published by Axios.

TechCrunch noted that climate tech investment has become increasingly focused on energy and infrastructure, a shift driven largely by AI demand. Data centers and AI inference workloads consume growing amounts of electricity, creating an opening for startups that can supply power cheaper or more flexibly. Schroepfer told the Inevitable podcast that bring-your-own power “is going to be a competitive advantage over time” in energy-intensive industries.

Portfolio and Scale

Gigascale has already invested in more than 25 companies through Schroepfer’s family office, according to BusinessWire. The portfolio includes Commonwealth Fusion Systems (nuclear fusion, which raised $863 million with backing from Nvidia, Google, and Bill Gates), Heron Power ($140 million raise for grid technology), Mill (food waste), and Form Energy (100-hour iron-air batteries used in Google’s 1.9 GW clean energy deal), as reported by TechCrunch.

The $250 million vehicle is the first with institutional LPs. Axios reported that while Gigascale works with early-stage startups, it plans to “opportunistically” complement this with later-stage support, enabling founders from first check through scaled deployment.

The AI Infrastructure Connection

The fund’s relevance to the agent ecosystem is indirect but structural. AI agents running continuously require persistent compute, and persistent compute requires power. Natural gas turbines have waitlists stretching into the early 2030s, per TechCrunch. That bottleneck creates demand for alternative energy sources, which is exactly where Gigascale’s portfolio sits: nuclear (Commonwealth Fusion), grid tech (Heron Power), and long-duration storage (Form Energy).

For teams building agent infrastructure that scales, the constraint is increasingly not software. It is power, cooling, and physical plant. Gigascale is betting that the companies solving those constraints will win on economics, and that AI’s appetite for electricity will accelerate the timeline.