Envision launches Dubhe AI energy foundation model with Masdar

Dubhe is a notable example of AI moving from purely digital tasks into real-time control of physical infrastructure. By framing it as an “energy foundation model” that manages generation, storage, grid dispatch and demand using massive streams of live data, Envision is essentially building a domain-specific AGI for power systems. If it works at scale, AI stops being just a load on the grid and becomes an active manager of the grid itself, which is exactly the kind of cyber‑physical control loop advanced AI will require.([prnewswire.com](https://www.prnewswire.com/ae/ar/news-releases/u0065u006Eu0076u0069u0073u0069u006Fu006Eu002Du062Au064Fu0637u0644u0642u002Du0064u0075u0062u0068u0065u060Cu002Du0644u062Au0634u0643u064Au0644-302663079.html))

The strategic partnership with Masdar is equally important. Abu Dhabi is rapidly becoming a global hub for both AI compute and energy projects; combining Envision’s AI stack with Gulf capital and large renewable deployments could create a powerful testbed for AI‑managed energy systems under real‑world constraints. For the race to AGI, this tightens the feedback loop between AI compute demand and clean energy supply. Whoever can reliably feed hyperscale training clusters with low‑carbon, dynamically optimized power gains a structural advantage in scaling models without hitting political or physical energy ceilings.