National Grid Invests $1.75 Billion in US AI Power Company Jouleant

On July 1, 2026, National Grid announced a strategic $1.75 billion investment in Jouleant, a US company that specializes in supplying electricity to AI data centers.

What Jouleant is and why invest $1.75 billion in it

Jouleant is a US company built for one specific job: delivering reliable electricity to AI computing facilities and next-generation data centers. Unlike traditional independent power producers, Jouleant designs its model around the needs of AI customers: high load density, steady supply and flexible scaling.

The investment by National Grid, one of the largest electricity and gas network operators in the United Kingdom and the Northeast US, is strategic rather than a financial portfolio play. National Grid CEO Zoë Yuinovich, speaking on Bloomberg Open Interest, described the deal as a bet on the inevitable: electricity demand from AI infrastructure will grow faster than the traditional industry can plan and build capacity.

• Investment amount: $1.75 billion
• Jouleant is based in the US and specializes in AI power
• Deal announced on July 1, 2026
• Strategic investor: National Grid, CEO Zoë Yuinovich

Why AI load has changed grid operators' calculations

Grid operators have traditionally forecast demand 10–20 years ahead, relying on demographics, industrial growth and historical consumption curves. AI has broken that predictability. Training a single large language model requires enormous computing power; serving billions of requests to AI services creates a constant, almost non-declining base load, unlike industrial or household consumers, whose usage dips at night and on weekends.

According to a number of industry estimates, total electricity consumption by US data centers could double by the end of the decade compared with the early 2020s. Big tech companies such as Microsoft, Google, Amazon and Meta have already moved beyond buying "green" certificates and become direct investors in power generation, from small modular nuclear reactors to large solar farms. A single large AI data center campus can draw more power than a small city, and the demand appears fast, within two to three years, while building a new substation or transmission line takes five years or more.

A new class of companies between the grid and the data center

The Jouleant investment reflects a broader structural shift: a new class of specialized intermediary companies is forming between traditional power generation and the end AI consumer. They take on technically and bureaucratically difficult tasks: obtaining construction permits, dealing with regulators, designing substations for specific loads and managing peak demand.

For tech companies, this removes the need to build their own energy expertise. For investors such as National Grid, it offers a way to share in the value created by the AI boom without building data centers directly. The $1.75 billion commitment, one of the largest transactions in this segment, confirms that the "AI power trade" has moved beyond stock-market speculation and become a real operating strategy for incumbents in the industry.

What this means

The National Grid–Jouleant deal is a signal to the market: traditional utilities see AI load as a long-term structural reality, not a temporary boom. Access to reliable electrical capacity is becoming as strategic a resource for the AI industry as semiconductors and compute chips, and those who have taken a position at this bottleneck today will shape the industry's growth conditions for years to come.

*Meta has been designated an extremist organization and is banned in Russia.