Emerald AI raises $150M at $1.05B valuation

Emerald AI, a US startup developing software that allows AI data centres to adjust their electricity consumption according to grid conditions, has raised $150 million in an oversubscribed Series A round at a valuation of $1.05 billion.

The financing was co-led by Energize Capital, an investment firm focused on the energy transition, and DCVC, a deeptech venture capital firm. The round also included strategic and financial investors such as NVIDIA, Samsung Ventures, Siemens, Aramco Ventures, Salesforce Ventures, GE Vernova, RWE, and JERA Ventures.

Emerald AI plans to use the new capital to expand commercial deployments with AI companies, data-centre operators and electric utilities worldwide. The company said its software is already operating at multi-megawatt, full-data-centre scale, following five demonstrations in Arizona, Illinois, Virginia, Oregon, and London.

The company was founded by Indian-origin energy expert Varun Sivaram, who is its CEO and founder. Sivaram has previously worked as chief technology officer at India’s ReNew Power, chief strategy and innovation officer at Ørsted, and as a senior US diplomat focused on clean energy.

Emerald AI’s proposition centres on a problem becoming increasingly important as AI computing expands. Data centres require large and often concentrated amounts of electricity, while building new generation and transmission infrastructure can take years. The firm’s Emerald Conductor software coordinates computing workloads and onsite energy resources so a data centre can reduce or adjust its power consumption when the grid is under pressure, while maintaining critical workloads.

Sivaram said the company was founded on the belief that “the intelligence driving the AI revolution could solve its own greatest bottleneck: power”. He said its demonstrations showed that data centres could adjust electricity use when the grid needed relief without compromising critical computing, and that the new funding would allow the technology to be deployed more widely.

The development comes amid a broader change in the economics of AI infrastructure. The International Energy Agency expects data centres to account for nearly half of the growth in US electricity demand through 2030. Grid operators are already responding to the pressure. In June, the Federal Energy Regulatory Commission ordered six regional grid operators to justify or reform rules governing connections for large electricity users such as data centres.

PJM, the largest US grid operator, has also warned of growing supply shortages as data-centre demand rises. It has proposed measures to track large electricity users and address capacity gaps, highlighting why technologies that can make loads more flexible are attracting attention.

Emerald’s investors are making a similar argument. Energize Capital managing partner John Tough said, “The binding constraint on AI is no longer chips or capital; it is power.”

DCVC co-founder Zachary Bogue pointed to the company’s technology that could make flexibility a permanent feature of how data centres are powered. Siemens USA chief executive Ann Fairchild, meanwhile, noted better coordination between AI workloads and the grid would be important as computing demand rises.

The challenge is not confined to the US. India is rapidly expanding its own AI and data centre infrastructure. Reliance has announced a roughly $110 billion AI infrastructure investment plan, while the Adani Group has outlined a $100 billion investment programme. HCLTech and Sarvam AI are developing a $1.48 billion AI data centre in Odisha, while AirTrunk has committed $30 billion to develop 5GW of data-centre capacity in India.

Indian AI companies are also raising substantial capital, including Sarvam AI which raised $300 million at around $1.5 billion valuation, while AI infrastructure company Neysa secured backing for up to $600 million in primary equity and plans to raise a further $600 million in debt.