發生了什麼事
Virginia Business reports that Emerald AI raised $150 million in a Series A round co-led by Energize Capital and DCVC, giving the data-center power-management startup a reported valuation of $1.05 billion. The company plans to use the funding to expand commercial deployments globally.
Virginia Business reports that Emerald AI raised $150 million in Series A funding co-led by Chicago-based Energize Capital and Silicon Valley-based DCVC. The round reportedly values Emerald AI at $1.05 billion, making it a so-called unicorn. The article says the company was incorporated in Washington, D.C., has primary offices in Arlington County’s Clarendon area, and counts 12 Fortune Global 500 companies among its investors. Those investors are described as members of Emerald AI’s Strategic Advisory Board and as participants in product integration and commercial deployments. The funding details and valuation are reported by Virginia Business and are not independently confirmed in the supplied source by a public financing document or by the participating investors.
According to Virginia Business, Emerald AI was founded in 2020 and developed the Emerald Conductor software platform to help data centers reduce or shift electricity consumption when the power grid faces peak demand. The company says the system can support grid stability while preserving acceptable AI-computing performance. Virginia Business reports that Emerald AI emerged from stealth in summer 2025 with $24.5 million in seed funding, later raised $22.7 million earlier in 2026, and has backing from Nvidia and former Google executive Jeff Dean. The article says Emerald AI now considers its technology commercially deployed and intends to use the new capital to expand deployments with AI companies, data-center operators, and electric utilities.
Virginia Business reports that Emerald AI completed five demonstrations during the past year at commercial data centers in Virginia, Arizona, Illinois, Oregon, and London. The article says the company has begun scaling deployments, including a deployment across an entire data center in California. It also reports that Emerald AI is working with Digital Realty and Nvidia on a nearly 100-megawatt AI facility in Manassas that is expected to come online later in 2026. These are reported plans and company-related claims; the supplied article does not provide independent test data, customer performance measurements, contract terms, or confirmation that the Manassas facility has entered operation.
為什麼這很重要
Emerald AI is developing software intended to help data centers adjust electricity consumption during periods of peak grid demand while maintaining AI workloads. If the company’s reported deployments and estimates hold, the approach could affect how new AI data centers connect to constrained power systems.
The financing highlights a practical constraint surrounding AI infrastructure: access to electricity. Virginia Business cites the International Energy Agency’s projection that data centers could account for nearly half of U.S. electricity-demand growth through 2030. Within that context, Emerald AI’s product is aimed at the operating layer between large computing facilities and the power grid. Its stated purpose is to make electricity use more flexible at times when demand is high, potentially allowing data centers to respond to grid conditions without simply shutting down AI workloads.
The article reports that Emerald AI estimates widespread adoption of its technology could unlock more than 100 gigawatts of capacity on the existing U.S. power grid. That figure is an estimate from the company, not an independently validated result in the supplied source. Its significance depends on how much power individual facilities can shift, how quickly they can respond, what workloads can tolerate interruptions or slowdowns, and whether utilities recognize the response as reliable grid support. The source does not provide the methodology behind the 100-gigawatt estimate.
The reported funding also connects AI infrastructure investment with energy-sector participation. Virginia Business lists Nvidia, Samsung Ventures, Siemens, Aramco Ventures, Salesforce Ventures, GE Vernova, RWE, JERA Ventures, In-Q-Tel, and other financial or strategic investors in the round. If those relationships translate into deployments, Emerald AI could gain access to both data-center operators and utilities. At the same time, the article does not establish how much revenue the company generates, how many facilities currently use the software, or whether customers have achieved measurable cost or reliability improvements.
互動機制：它實際上是如何運作的
以互動方式探索這項發展背後的基礎技術。
A route planner searches possible journeys using explicit rules. What does this illustrate about AI?
接下來看什麼
The key questions are whether Emerald AI can scale beyond demonstrations, whether its software can maintain computing performance during grid events, and whether utilities and data-center operators adopt the system broadly. Virginia Business reports these developments, but the source does not independently verify the company’s performance claims, customer results, or valuation.
The immediate test is deployment at full data-center scale. Virginia Business reports that Emerald AI’s technology is operating commercially and that the company is deploying it across a California data center, but it does not identify the customer or provide before-and-after electricity data. Future reporting should establish the number and size of active sites, the types of AI workloads involved, and how the software performs during actual periods of grid stress.
The Manassas project is another milestone to monitor. The article says Emerald AI, Digital Realty, and Nvidia are working on a nearly 100-megawatt AI facility expected to come online later in 2026. The source does not say what portion of the facility’s electricity use Emerald AI will control, what operating targets have been agreed, or whether the project has received all required power and construction approvals. Those details will determine whether the project demonstrates a repeatable commercial model or remains a one-off deployment.
Investors and customers will also face questions about trade-offs. Virginia Business reports that Emerald AI says its demonstrations adjusted power use without compromising critical computing workloads, but the article provides no independent measurements of , uptime, completed jobs, or financial savings. Future evidence should show how the system handles different workloads, how much control data-center operators retain, how utilities compensate flexibility, and whether the reported valuation is supported by sustained commercial growth. Until then, the funding and deployments are concrete reported developments, while the broader grid and affordability claims remain company assertions.