On September 16, 2026, Emerald AI, Google and NVIDIA announced the AI Energy Management Alliance (AEMA), a coalition to accelerate interconnection of data centers that can dynamically adjust electricity use in response to grid conditions. The alliance spans the full AI and power value chain — AI platforms, infrastructure providers, data center operators, technology companies, power producers, utilities and regional grid operators — to develop technical and operational approaches for flexible AI factories.

Power has become a defining constraint on U.S. AI infrastructure expansion. Traditional interconnection processes were designed for facilities with flat, static demand, not for computing infrastructure that can respond intelligently when the grid is stressed. A flexible data center adjusts its draw by shifting workloads, discharging storage, using paired generation or responding to contingencies, serving as a controllable resource rather than an inflexible load.

Confirmed

AEMA is technology-neutral and performance-based, focusing on measurable services — response speed, duration, predictability and emergency behavior — rather than specific hardware or software. Its principles call for defining ride-through, curtailment and contingency-response obligations before a facility connects; standardizing technical requirements, performance metrics and operational data sharing; creating faster, risk-adjusted pathways for customers making credible, verifiable flexibility commitments; and allocating interconnection costs to reflect actual system impacts and benefits such as avoided upgrades and improved ramping capability.

Google has committed 1 gigawatt of reducible power demand through utility agreements nationwide, said Tyler Norris, Google's head of advanced energy market innovation. The coalition includes 20 companies and organizations across AI and energy sectors, including Anthropic, utilities National Grid and AES, and power producers Constellation, NRG and RWE. NVIDIA and Emerald AI are already working with energy and infrastructure leaders on AI factories that respond to grid conditions in real time.

Emerald AI's pilot deployments demonstrate technical feasibility: a 25% power reduction at an Oracle data center in Arizona during peak demand, load shifting between Virginia and Chicago sites within milliseconds, and multi-MW deployments in California that cut power use during peak CAISO demand. The company says intelligent flexibility can unlock up to 100 gigawatts of capacity on the existing U.S. grid for AI.

Unknown

Whether regulators will create the fast-track interconnection pathways AEMA advocates remains uncertain. Federal regulators in June directed regional grid operators to examine new options for connecting large, flexible power users, but no rulemaking timeline has been set. Facilities should qualify for faster or larger connections only if their ability to cut demand is "verifiable and enforceable," said Emerald AI CEO Varun Sivaram. The 100 GW grid-scale claim has not been independently verified.

Community acceptance is another hurdle. New AP-NORC/University of Chicago polling shows 84% of Americans are concerned about data centers' impact on local electricity prices, and more than half are extremely or very concerned about AI's environmental impacts, up from 41% a year ago. Roughly equal shares of Democrats and Republicans — 79% and 76% — support requiring data center developers to pay for grid upgrades needed to support them.

Our take

AEMA reframes data center flexibility as a grid service with standardized performance metrics, not an ad-hoc operational choice. If regulators adopt performance-based interconnection rules, leverage shifts to operators who can demonstrably deliver verifiable flexibility — favoring platforms like Emerald AI's Conductor that orchestrate workload shifting, battery dispatch and spatial load migration. Google's 1 GW commitment sets a tangible benchmark; the test is whether utilities translate that into faster queues and cost allocations that reward flexibility.

Sources