Google, Nvidia, and Anthropic want Emerald AI to find space on the grid for more data centers
TechCrunch Tim De Chant ● Covered by 6 sources
Google, Nvidia, Anthropic, and Emerald AI are teaming up to squeeze more data centers onto the grid. The pitch: make AI servers flex their power use instead of just demanding more wires and plants.
Based on reporting by TechCrunch, Tim De Chant — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Emerald AI has pulled together Google, Nvidia, Anthropic, and several utilities into a new coalition aimed at one very specific problem: where the electricity for more data centers will come from. The group, called the AI Energy Management Alliance, wants demand response to become a normal part of building out AI infrastructure, not an emergency tool used once in a while.
The idea is simple enough. When the grid gets tight, data centers would pause noncritical work or move some compute elsewhere instead of leaning on backup diesel. Emerald AI says that kind of flexibility could make room for another 100 gigawatts of data centers on the grid. That is the sort of number that tells you this isn’t just a tinkering exercise.
Demand response itself is not new. Utilities have used it for decades, usually with factories that agree to cut back during peak demand and get paid for it. What’s new is trying to make data centers behave the same way, except faster, cleaner, and at a much larger scale. Emerald AI’s software sits between utilities and data centers so requests can be coordinated quickly, with loads shifted to other sites when the grid has spare capacity.
The company is not alone in this race. Google has its own tools, and Enel X lets data centers use uninterruptible power supplies to shave peaks. A Goldman Sachs study published last year said capping grid use at 90% for a few hours at a time could free up 76 gigawatts of capacity. Emerald AI also has fresh money behind it: a $150 million Series A led by Energize Capital and DCVC.
Still, this is not a magic fix. Ayse Coskun, Emerald AI’s chief scientist, said the tech can reduce how much the industry needs new generation, but not wipe that need out. The broader bet is that AI workloads are peaky enough, and data centers nimble enough, that the grid can be treated less like a brick wall and more like something negotiable. That may be overdue.
My take — AI-written commentary, not fact-checked reporting
This is the rare grid story that sounds sensible instead of like startup incense. Data centers have spent years acting as if the answer to every power problem is more power; Emerald AI is arguing for better manners instead. That won’t save the grid, but it might save everyone from building expensive steel around the idea that software can never be the thing that flexes.
Read more about this at: TechCrunch