Powering AI at gigawatt scale and doing it sustainably are not competing goals. The buildout needs abundant power, and the lowest-carbon path to that power runs through a deliberate pairing of on-site generation with firm grid capacity. Our approach treats sustainability as an engineering constraint to design around, not a label to apply afterwards.
On-site generation does the early and the marginal work. Solar and storage co-located with a site can shave peak demand, firm up supply, and lower the carbon intensity of the power a campus draws, while reducing the load that has to come down the wires. It is not the whole answer at this scale, but it is a meaningful and steadily growing part of it.
Grid capacity carries the base. A firm connection, underwritten by long-term agreements, supplies the bulk of a gigawatt campus and lets generation decarbonise around it as the wider network does. Securing that capacity early is what makes the sustainable design viable, because a clean supply you cannot connect to is no supply at all.
The result is power that is abundant, firm, and lower-carbon, by design rather than by accident. Pairing generation with grid capacity, site by site, is how we deliver energy at the scale the AI buildout demands without conceding on the carbon it produces.
