According to a recent LinkedIn post from Emerald AI, founder and CEO Varun Sivaram participated in a Utility Dive feature alongside executives from NVIDIA, Heron Power, and Invenergy. The discussion reportedly focused on a shared proposal outlining three principles for how AI data centers should interact with the power grid.
The post highlights a framework in which data centers aim to avoid harming grid stability, operate flexibly when the grid is stressed, and contribute additional capacity over time. Emerald AI suggests that alignment among companies spanning chips, power electronics, generation, and software may indicate emerging cross-industry standards for data center grid behavior.
The LinkedIn content also references an estimate that modest load flexibility—reducing consumption for less than 0.5% of hours annually—could enable the existing grid to absorb roughly 100 GW of new demand. This perspective implies significant potential headroom for AI-related power growth, provided data centers adapt operations to grid needs.
For investors, the post points to the strategic importance of grid-aware AI infrastructure and to collaboration between technology and energy firms. If such principles gain traction, Emerald AI could be positioned to benefit from policy and market trends favoring flexible, grid-integrated AI data centers, potentially enhancing its relevance in both the AI and energy ecosystems.