Breno Araujo, CEO and Co-FounderCurtailment has become a global constraint on renewable growth. In 2024 alone, roughly 100 terawatt hours of renewable energy were wasted. The same grid infrastructure causing that loss is also slowing the expansion of AI data centers, where connection waitlists can extend up to ten years in some regions.
"Enverge addresses this dual crisis not by optimizing the grid, but by bypassing it entirely," says Breno Araujo, CEO and co-founder.
Bypassing the Grid Constraint
Enverge approaches the curtailment problem from a different direction. Rather than focusing on grid optimization, the company installs AI micro-datacenters directly at renewable generation sites. These facilities operate behind the meter at solar, wind and hydro plants, allowing unused energy to be captured at the source.
-
Enverge addresses this dual crisis not by optimizing the grid, but by bypassing it entirely.
This design removes dependence on grid expansion while creating a new pathway for monetizing excess generation. Renewable producers can convert otherwise wasted electricity into high performance compute distributed through the internet.
The model also enables faster infrastructure deployment. Through an asset-light approach and modular micro-datacenter architecture, Enverge works with local partners to deploy projects globally within 60 to 90 days. Traditional data centers typically require years of development and permitting before becoming operational.
Converting Energy into High-Value Compute
Machine learning and AI workloads form the economic engine behind the model. By converting curtailed electricity into distributed GPU compute capacity, Enverge allows renewable energy providers to participate in the rapidly expanding AI economy.
Instead of selling excess power into volatile local markets, generators can convert that energy into computing services priced in stable international currencies such as USD, EUR and GBP.
The model also strengthens the financial viability of renewable projects. Energy producers gain two independent revenue streams rather than relying entirely on electricity sales through the grid. That flexibility can improve return on investment for new renewable developments, particularly in emerging markets where grid infrastructure may lag behind generation capacity.
Recovering Lost Revenue from Curtailment
A recent project illustrates how the approach works in practice. Enverge is currently running four proof-of-concept deployments around the world, including one with a billion-dollar energy company experiencing approximately $40 million in annual losses from curtailment.
Enverge's strategy involves deploying a micro-datacenter to convert the client's excess energy into AI compute capacity. Instead of selling curtailed power at spot prices or wasting it entirely, the company transforms it into higher-value digital infrastructure.
For projects at this scale, the financial impact can be significant. The deployment can generate up to $2,500 in net revenue per megawatt hour, with a projected payback period of about 32 months and an internal rate of return around 25 percent.
Even more notable is the efficiency of the model. Recovering the full $40 million in lost revenue requires only about two percent of the curtailed energy produced by the facility.
Demand for AI computing capacity is expanding rapidly. Energy consumption for AI workloads is projected to grow from eight terawatt hours in 2024 to more than fifty terawatt hours by 2026. Analysts estimate that global AI data center supply could face a deficit of fifteen gigawatts by 2030.
Enverge is expanding its network with projects underway across the United States, Brazil, Norway and Greece. Through distributed micro-datacenters deployed at renewable energy sites, the company is positioning curtailed energy as a new source of infrastructure for the expanding AI economy.
