ThinkLabs AI Raises $28M with NVIDIA Backing to Solve Power Grid Crisis for AI Data Centers

2026-03-31T12:53:23.487Z·2 min read
ThinkLabs AI, a startup building physics-informed AI models to simulate electric grid behavior, has closed a $28 million Series A round led by Energy Impact Partners, with participation from NVIDIA...

ThinkLabs AI, a startup building physics-informed AI models to simulate electric grid behavior, has closed a $28 million Series A round led by Energy Impact Partners, with participation from NVIDIA's NVentures and Edison International.

The Problem

U.S. electricity demand is projected to grow 25% by 2030, driven by:

This demand is crashing into a grid designed decades ago for fundamentally different requirements. When utilities need to understand grid impact of connecting a large data center or EV charging cluster, engineers must run power flow simulations that take weeks or months using legacy tools from Siemens, GE, and Schneider Electric.

ThinkLabs' Solution

MetricLegacy ToolsThinkLabs AI
Study durationWeeks to monthsUnder 3 minutes
Scenarios per run110 million in 10 minutes
AccuracyBaseline99.7%+

The company uses physics-informed AI models trained on engineering simulators to calculate grid behavior orders of magnitude faster than traditional approaches.

Notable Investors

CEO Quote

"We are dead focused on the grid. We do AI models to model the grid, specifically transmission and distribution power flow related modeling." — Josh Wong, CEO

Analysis

ThinkLabs sits at the intersection of two mega-trends: AI infrastructure buildout and grid modernization. The round was "way oversubscribed" per the CEO, indicating that utilities and investors see physics-informed AI for grid management as a critical bottleneck to solve. As every major tech company races to build new data centers, the companies that can accelerate grid interconnection approvals and studies will become essential infrastructure providers in their own right.

← Previous: DeepSeek Suffers Three Consecutive Days of Outages, Longest Downtime Exceeds 10 HoursNext: UBTech Revenue Surges 53% to 2 Billion Yuan as Humanoid Robots Become Top Revenue Source →
Comments0