Emerald AI
Bottleneck owner
Grid software that turns AI data centers into flexible, demand-responsive grid assets.
Emerald AI builds the Conductor platform, software that makes AI data centers' power demand flexible so they can act as grid partners instead of pure load. Data centers running Conductor modulate consumption in response to grid stress, which Emerald says can unlock up to 100 GW of capacity on the existing U.S. grid while protecting reliability and keeping energy affordable for communities. Its Arizona demonstration results were peer-reviewed and published in Nature Energy. Emerald emerged from stealth with a $24.5M seed, raised a $150M Series A at a $1.05B valuation in August 2026, and has formed an AI-factory coalition with Nvidia, Google and Anthropic alongside power companies including AES, Constellation, Invenergy, NextEra Energy and Vistra. It ran a pilot with Silicon Valley Power in Santa Clara.
Research mapping
Physical AI / Grid, Power & Thermal Infrastructure (L2) / Bottleneck owner
Evidence boundary
Original analysis and independent corroboration remain under development. Commercial scale and durable technical advantage remain under review.
What makes it defensible
Governs how much power large AI loads draw moment to moment - the demand-response control layer for AI data centers, validated in a peer-reviewed Nature Energy demonstration.
Position in the stack: Sits between hyperscale data centers and utilities/grid operators, owning the control and orchestration software that determines how much load each site can flex. Hard to copy because it requires site-level validated demonstrations plus simultaneous utility and AI-lab relationships.
Who buys it
Named customers: NVIDIA, Oracle, Digital Realty, Nebius, National Grid, Silicon Valley Power
Founding team
Dr. Varun Sivaram — Founder and CEO — Energy technologist and former senior clean-energy policy official; named to the TIME100 Next and TIME100 Climate lists, MIT Technology Review 35 Innovators Under 35, and World Economic Forum Young Global Leaders
SourceFinancing detail
- Last round
- Series A ($150M, Aug 2026)
Investors: Energize Capital, DCVC, NVIDIA, Siemens, GE Vernova, Samsung Ventures, Aramco Ventures, Salesforce Ventures, RWE, Radical Ventures, Lowercarbon Capital, Energy Impact Partners, In-Q-Tel
Mapped funds and investors
Direct relationships supported by published evidence.
Reported holdings as-of
No sourced relationships mapped yet.
Showing up to four relationships in each group. ETF weights describe the fund’s portfolio, not ownership of this company. Private investment history does not establish current ownership.
Evidence & media
-
Data center power startup Emerald AI raises $150M at $1.05B valuation
What this source supports: $150M Series A at a $1.05B valuation (co-led by Energize Capital and DCVC).
What it does not establish: No revenue, backlog, or signed utility contract values.
Source -
Emerald AI and Nvidia aim to offer the fast pass for data center grid connects
What this source supports: Emerald AI and NVIDIA pitch a 'fast pass' for data-center grid interconnects with power producers.
What it does not establish: Describes the partnership; no quantified signed utility deals.
Source -
NVIDIA and Emerald AI Join Leading Energy Companies to Pioneer Flexible AI Factories as Grid Assets
What this source supports: NVIDIA (an Emerald AI investor via NVentures) and energy partners announce flexible AI-factory pilots.
What it does not establish: NVIDIA/partner announcement; no independent load-flex performance data.
Source -
How Emerald AI Makes AI Factories Power-Flexible
What this source supports: NVIDIA showcases Emerald AI's grid-flexible AI-factory software (NVentures portfolio company).
What it does not establish: Investor/partner marketing; results are not independently verified.
Source -
Emerald AI coalition targets 100GW of grid space for data centers
What this source supports: Emerald AI launched the AI Energy Management Alliance (AEMA) with Google, Nvidia, Anthropic and utilities (AES, Constellation, National Grid, NRG) on 2026-09-16 to standardize data-center demand response.
What it does not establish: 100 GW figure is the alliance's own projection; not independently verified.
Source