GridCARE, an AI startup focusing on expediting data center grid connections, has completed its first joint project with West Coast utility Portland General Electric (PGE). The project explored the possibility of accelerating large load interconnections by leveraging flexibility in data centers in Hillsboro, Oregon.

According to the companies, the project led to 80MW of incremental capacity to be freed up for additional data center load in 2026, part of a total of more than 400MW that will be energized by 2029.

"Utilities face a once-in-a-generation opportunity to meet exponential load growth," said Amit Narayan, CEO of GridCARE. "By combining generative AI with grid flexibility, we help utilities like PGE move at the pace the AI economy requires. This project shows what's possible when utilities and innovators partner closely."

To achieve this, PGE utilized GridCARE's patented DeFlex methodology to apply generative AI forecasting, detailed hourly demand modeling, and grid flexibility solutions. Through the use of the AI system, PGE was able to optimize the use of flexible resources such as batteries and onsite generation to interconnect multiple data center customers years earlier than initially expected.

"The pace of AI and cloud growth demands new thinking about how we power the digital economy," said Christian Belady, retired VP of Datacenter Advanced Development at Microsoft. "What PGE and GridCARE have accomplished in Hillsboro is the kind of innovation that not only enables AI and cloud expansion — it unlocks billions in new investment and accelerates returns across the entire North American ecosystem."

According to GridCARE, its generative AI solution offers numerous benefits for data centers and grid operators. These include faster interconnection for large loads without having to wait for grid upgrades, more efficient use of existing grid assets, and increased flexibility that can support both system reliability and energy goals.

"Delivering safe, reliable electricity at the lowest possible cost are our core customer commitments," said Larry Bekkedahl, SVP of Strategy and Advanced Energy Delivery, Portland General Electric. "By working with GridCARE, we've successfully implemented an AI-based approach that helps us serve fast-growing data center demand while keeping costs low and maintaining system reliability. This kind of beneficial electrification is a win for our customers, our communities, and Oregon's economy."

GridCARE launched earlier this year, following the successful raise of $13.5 million. The company was born out of the Stanford Sustainability Accelerator program, and through its platform, aims to leverage advanced generative AI-based analysis to detect pockets with geographic and temporal capacity on the existing grid. By doing so, the company claims, it can reduce the time-to-power of data centers to 6-12 months.

The company is headed up by Amit Narayan, who previously founded chip design automation firm Berkley Design Automation. Other co-founders of the company include Ram Rajagopal, Liang Min, and Arun Majumdar, all of whom are employees at Stanford University.

DCD spoke to CEO Amit Narayan about the company and its long-term goals in our latest Critical Power Supplement. Available here.