A lot has changed since we last explained what Camus does. The grid has changed. The problem has changed. And the work we're doing has evolved. Here's the updated story.
The original question was simple: why did it take so long to connect new things to the grid?
Utilities reviewed interconnection requests one at a time, in sequence, using static planning tools built for a grid that didn't change much. Demand from EVs, solar, batteries, and data centers kept climbing. The queue backed up. Timelines stretched to five, seven, ten years. Projects died waiting.
That's the problem Camus was built to solve. It’s the same underlying technology utilities have trusted us with for years, applied and refreshed, to now solve an unprecedented global challenge. If we don’t get this right as an industry, the repercussions will be significant and take decades to correct. So, let’s make it happen.
When I joined Camus in 2024, something changed. AI data center construction exploded, and the single biggest question in the industry became: where does the power come from, and how fast can we get it?
Interconnection queues were already stressed and became a national bottleneck. Projects that would take five to seven years under traditional interconnection timelines were seeking power at any cost, financial or environmental. Lots have gone off-grid and fossil-fuel powered. The AI buildout has a grid problem.
We'd been working on that grid problem from the utility side for years. We had the data, the physics models, the utility relationships, and the platform. We saw what almost no one else could see: the bottleneck wasn't about how much power existed — it was about how utilities studied, approved, and operated new grid connections.
The insight that changed our trajectory came from working directly with utilities: the grid under-utilizes capacity, and a lot of it. Not all the time, but most of the time. A data center that agrees to reduce demand during the rare hours when the grid is constrained, 40 to 70 hours out of 8,760, can connect to grid power years before a utility could build out the necessary infrastructure upgrades to make a full firm connection a reality.
This is the core of flexible grid connection. Not a workaround. Not a compromise. A structurally better way to use what the grid already has, at no additional cost to ratepayers.
In partnership with a major PJM-area transmission operator and Google, we ran the first publicly available study to model flexible data center interconnection at scale — six real candidate sites, hourly analysis across all 8,760 hours of the year, real SCADA data and real transmission planning models.
The findings were unambiguous. Utilities can connect up to 3x more data center capacity in two years using flexible grid connections, compared to the traditional path. Flexible sites will receive grid power for more than 99% of all hours. No major transmission upgrades required to begin operations. The economics work for data centers: accessing power five years earlier generates $4.7–$5.5 billion in additional EBITDA per site. Data centers achieve net-positive returns when interconnection timelines are accelerated by 2 years or more.
For the first time, the industry has real numbers on what flexible grid connection can deliver at scale.
The full whitepaper is publicly available. Read the full report for free here.
Flexible grid connection is the category we've created and the one we lead.
The platform that makes it work is FlexConnect. It runs a comprehensive interconnection analysis, alongside the utilities’ traditional planning methodology, to determine exactly how much load the grid can serve at a candidate site, hour by hour, and then stays operational, enforcing the terms of the flexible connection every day. Day-ahead capacity limits. Real-time AC power flow forecasts. Automated coordination of on-site resources during constrained periods. The data center runs normally; the grid operates with steadfast safety and reliability.
We've deployed at utilities serving 22 GW of peak demand. New utility deployments go live in as little as 12 weeks. The platform is trusted by utilities because it was built with utility partners who validated against real planning tools, not tested in simulation.
The foundational technology is the Orchestration Platform, our operational data management system, which gives utilities a unified view of their transmission and distribution grid from SCADA, GIS, DERs, AMI and data sources such as ADMS or EMS. It's what makes FlexConnect possible: you cannot run flexible grid connections without knowing, hour by hour or minute by minute, what the grid can actually support.
Together, these two products cover the full arc from planning to operations. Flexible Grid Connection isn’t a marketing term; it’s our engineering standard.
Flexible grid connection isn't a data center story, but rather a community grid story.
The pressure data centers are putting on interconnection queues today is a preview of what clean generation will face tomorrow. Wind, solar, storage, all queued up, all waiting for upgrades that take years.
The approach we're proving with data centers to connect flexibly now is the architecture of the future. One where we upgrade the infrastructure over time and operate within real network constraints rather than worst-case assumptions.
The cost benefits run in three directions. Ratepayers avoid the rate increases that come with large infrastructure buildouts. Utilities turn existing infrastructure into a new revenue stream. Data centers access power years earlier — the kind of head start worth billions in additional EBITDA per site.
FERC, PJM, ERCOT, SPP, and MISO are all working through versions of the same question: how do you structure flexible service for loads and generation that can participate in grid management? The regulatory groundwork is being built in parallel, informed by our research. The technology was ready before the rules were.
Today, Camus works with more than 12 utility partners across the U.S., with platforms deployed at the millions-of-meters scale. We've co-authored research with Google, AES, and Princeton's ZERO Lab, been named a 2024 Technology Pioneer by the World Economic Forum, and closed our Series A.
The team includes engineers who built reliability systems at Google and Amazon, grid experts from NREL and GE, and operators who've spent careers inside utilities. We work with the people running the grid, not around them.
What hasn’t changed is our mission to decarbonize the grid with modern technologies that efficiently coordinate local resources. And today, FlexConnect is the coordination engine flexing when it matters.
Flexible grid connection is a real category solving a real problem at a real scale. The interconnection queue is the defining infrastructure bottleneck of the AI era. Utilities that move first on flexible connections will connect more load faster, generate more revenue from existing infrastructure, and reduce the risk of overbuilding due to phantom demand.
The data centers that move first will get to power years ahead of the competition. Camus builds the software that makes that happen, for both utilities and data centers.
Read the full Google/PPL study: Flexible Data Centers: A Faster, More Affordable Path to Power
