As AI infrastructure scales rapidly, a Vancouver-based startup is betting that the biggest constraint isn’t compute—it’s power.
Soma Energy has emerged from stealth with $7 million in funding to address what it describes as a structural bottleneck in the data centre ecosystem: the inability to access electricity quickly enough to support AI-driven growth.
While demand for data centre capacity is projected to surge in the coming years, the timelines required to build new generation, transmission, and interconnection infrastructure—often five to ten years—are fundamentally misaligned with the pace of AI deployment.
Soma’s approach is to bypass that constraint by unlocking capacity already available within the grid.
The company’s AI-powered energy intelligence platform operates across both supply and demand, optimizing how distributed generation, storage, and large loads interact with electricity markets in real time.
For data centre operators, the platform aggregates on-site generation, battery storage, and demand into a single control layer. This effectively transforms facilities into flexible grid assets, allowing operators to extract more capacity from existing infrastructure and significantly accelerate time to power.
For power producers, Soma provides real-time dispatch intelligence to determine when to generate, store, or trade energy across wind, solar, and battery assets—reducing costs and improving efficiency at the system level.
The founding team brings direct hyperscale experience to the problem. CEO Ath Caramanolis previously led AWS’s renewable energy optimization efforts, scaling the company’s portfolio to 10 gigawatts and negotiating more than $1 billion in deals. He is joined by CTO Mario Souto and Chief AI Scientist Henrique Hoeltgebaum, both of whom helped build machine learning systems to manage large-scale energy assets at AWS.
The company is already optimizing two gigawatts of electricity and working with multiple data centre customers, including deployments that have enabled access to capacity “significantly sooner than expected,” according to partners.
As the industry races to build AI infrastructure, Soma Energy is positioning itself as part of a new layer in the stack—one focused not on generating more power, but on orchestrating what already exists.
With fresh funding led by Category Ventures, the company plans to scale deployments across North America and expand its engineering and commercial teams.