National Grid invests $1.75 billion in Joulent to expand AI-focused power infrastructure.
UK-based National Grid on Tuesday announced a $1.75 billion strategic minority investment in power developer Joulent through National Grid Ventures. The investment backs a developer focused on building dedicated generation for hyperscale AI data centers on faster timelines than traditional utility interconnections allow—a capability that arrives less than a week after Engine No. 1 publicly launched Joulent as a standalone energy company.
National Grid said the investment will expand Joulent's project pipeline and draw on its expertise in high-voltage infrastructure, project execution, and electrical equipment supply chains. The agreement also increases National Grid's exposure to one of the fastest-growing sources of electricity demand as AI developers seek alternatives to increasingly constrained interconnection queues.
"This investment from National Grid Ventures strengthens our ability to deliver reliable, large-scale power on the timelines AI infrastructure and advanced industry now requires," said Joulent Founder and CEO Chris James.
The investment supports Joulent's first project, Project Kilby—a 2.67 GW natural gas-powered generation campus in Reeves County, Texas. Joulent is developing the project with Chevron subsidiary Energy Forge and GE Vernova to supply a Microsoft-operated data center campus under a 20-year power purchase agreement. Joulent has formalized a 50% ownership stake in the project.
Kilby leverages Joulent's "Across-the-Meter" model, which places dedicated generation adjacent to large AI campuses to accelerate deployment while limiting immediate impacts on the transmission system. Projects are designed for eventual grid interconnection, with future developments potentially incorporating renewables beginning with solar.
The move reflects a broader shift in hyperscale planning as developers pursue dedicated power plants, behind-the-meter generation, and other alternatives to lengthy interconnection timelines. Rather than waiting years for transmission upgrades, operators are pairing new generation directly with AI campuses to secure schedule certainty while pursuing eventual grid integration.
Neil Osnato, founder of Persistence Analytics Group, characterized the investment as "less about one company and more about a structural shift in how hyperscale infrastructure is being financed and delivered." He added that "the biggest takeaway is that power availability is becoming the primary determinant of data center development. Historically, developers selected a site and then worked with the utility to secure service. Increasingly, developers are selecting sites based on where reliable power can actually be delivered on the required timeline."
Osnato noted that even executed power purchase agreements and interconnection requests do not guarantee timely delivery, as transmission constraints, permitting, fuel availability, and execution risk can still delay projects. "It changes where the risk sits rather than removing it," he said.
National Grid CEO Zoë Yujnovich framed the investment as an infrastructure play tied to long-term contracted electricity demand from AI and other large-load customers. "Our investment in Joulent is a disciplined, partner-led investment in contracted critical infrastructure for the AI-driven large load economy," she said. "Through National Grid Ventures, we are gaining exposure to a major source of electricity demand growth that diversifies our regional U.S. exposure and is supported by strong partners."
The announcement expands Joulent's strategic relationships with GE Vernova and Chevron. GE Vernova Chief Commercial & Operations Officer Pablo Koziner said National Grid's investment demonstrates confidence in the collaboration among the companies to develop large-scale generation for AI compute.
For operators, the emergence of companies like Joulent underscores that securing power is becoming a development discipline in its own right. Access to land and capital still matters, but the ability to deliver reliable electricity on schedule is increasingly determining where AI campuses can be built.