Emerald AI Raises $150M for Grid-Responsive Data Centers

In a significant development for the energy-intensive world of artificial intelligence, Emerald AI has successfully closed a $150 million Series A funding round. The investment, co-led by DCVC Energy + Climate, aims to deploy the company's innovative technology that allows data centers to dynamically adjust power consumption in response to grid needs, thereby alleviating strain on energy infrastructure.

The escalating demand for AI computation presents a critical bottleneck for the industry, placing immense pressure on power grids, data center operators, and local communities. Emerald AI's platform, dubbed 'Conductor', addresses this challenge by enabling data centers to become responsive grid assets. This is achieved by intelligently shifting computational workloads, either temporarily or geographically, without compromising service level agreements (SLAs) for critical AI training and inference tasks. The technology has been validated through successful demonstrations across five commercial data centers in key locations including Arizona, Illinois, Virginia, Oregon, and London, with notable partners such as NVIDIA, EPRI, Oracle, Nebius, and National Grid.

“Our demonstrations globally confirmed that data centers can precisely modulate their energy usage when the grid requires support, all while maintaining the integrity of essential computing operations,” stated Dr. Varun Sivaram, Founder and CEO of Emerald AI. “This capital infusion will accelerate the commercial rollout of our technology to AI hubs worldwide.” The company's strategic position at the nexus of energy and compute provides a holistic view of the data center ecosystem, from utilities to end-users.

For utilities, Emerald AI's software facilitates the management and dispatch of flexibility signals to participating data centers during periods of grid stress. Data center operators and compute customers benefit from reduced power demand through intelligent workload rescheduling or spatial shifting to networked facilities. Furthermore, the platform enhances on-site energy resource management, such as dispatching battery power, and generates an auditable record of contributions to demand response events. This capability is crucial as data center builders and operators often face multi-year delays for grid interconnections due to insufficient generation and transmission capacity.

The traditional model of data centers requiring a constant, unwavering power supply (often referred to as 'firm' customers) necessitates utilities building substantial capacity to meet peak demand, even for infrequent hours of maximum stress. Research suggests that if large flexible loads could shed even a small percentage of demand (0.25% to 1%), the U.S. grid could absorb an additional 76 to 126 gigawatts of new capacity. Emerald AI's technology makes this level of dynamic power modulation feasible within data centers, transforming them from potential grid burdens into valuable assets that strengthen grid resilience.

With the Series A funding, Emerald AI is set to scale its operations significantly. The company is already moving beyond pilot programs into full-scale commercial deployments, serving a diverse clientele including electric power utilities, leading AI firms, and global data center operators. Future large-scale deployments are anticipated later this year. Notable initiatives include a partnership with Silicon Valley Power to launch the nation's first Flexible Load Interconnection Program and collaboration with Digital Realty and NVIDIA on the Vera Rubin AI Research Factory in Manassas, Virginia—a nearly 100-megawatt facility designed as the world's first power-flexible AI factory.