Dubhe by Envision
The AI Era’s Energy Future
A different kind of AI announcement
Most AI announcements feel like they belong to screens: faster models, smarter assistants, better text, better images. This one belongs to the world under your feet.
Because here’s the uncomfortable truth we’re all circling around: AI doesn’t run on inspiration. It runs on electricity. And the world is walking into an energy moment that’s both thrilling and fragile. Thrilling because we’re seeing renewable capacity scale. Fragile because the next decade of AI growth is going to press on power systems in ways most people still underestimate.
So when Envision stood in Abu Dhabi on 15 January 2026 and announced Dubhe — a “world’s first Energy Foundation Model,” designed to shape what it calls the AI energy system — it wasn’t just a product launch. It was a statement about what the next rules of global infrastructure might look like.
And the name choice tells you the mindset: Dubhe is the guiding star of the Big Dipper, used for navigation for thousands of years — and Envision is deliberately positioning this model as a navigation layer for the AI era.
Not “just another model
The world has two simultaneous transitions happening:
The energy transition (renewables, storage, green hydrogen, smarter grids).
The intelligence transition (AI moving from software to systems, from “digital brain” to “physical governance”).
What Dubhe suggests is that these transitions are no longer separate. They’re converging into a single system question:
Can we build an energy world that scales fast enough, stays stable enough, and becomes cheap enough to support an AI-heavy civilization without breaking planetary boundaries?
Envision’s framing is bold: it calls this shift a move from language-driven AI toward Physical Artificial Intelligence — AI that governs physical systems essential to human survival and prosperity.
That’s the key distinction:
Language models are brilliant at symbols.
Physical systems demand prediction, coordination, resilience, and real-time decisions.
Energy isn’t a chat box. Energy is a living network of generation, storage, grid constraints, and demand that changes by the minute. And renewables make it even more dynamic because they’re weather-shaped.
Dubhe by Envision
An “operating brain” for energy
According to the release, Dubhe sits at the core of Envision’s Physical AI architecture, analyzing real-world energy data streams to orchestrate renewable generation, storage, grids, and demand in real time.
Read that again slowly. The ambition isn’t analytics. The ambition is orchestration.
If we translate that into business reality, the promise is something like:
A wind farm that doesn’t just produce power — it negotiates with storage, price signals, grid constraints, and demand forecasts.
A battery system that doesn’t just charge/discharge — it behaves like a strategic asset, optimizing stability and cost.
A grid that doesn’t just distribute — it becomes adaptive, predictive, and coordinated.
A demand system (cities, factories, data centers) that doesn’t just consume — it becomes responsive, flexible, and financially smarter.
This is how you approach “AI energy”: not as a new power plant, but as a new intelligence layer over everything that already exists.
The new rules of energy data
Here’s the part most people miss: foundation models don’t become useful because they’re large. They become useful because they’re trained on trusted, high-quality, continuously flowing data.
In the energy world, that data isn’t social media. It’s grid telemetry, wind turbine performance, battery health, weather models, pricing, carbon factors, demand curves, industrial loads, outages, and maintenance schedules.
Which raises uncomfortable but necessary questions:
Who owns energy data when it affects public stability?
How do we share it without exposing critical infrastructure?
How do we standardize it across countries, utilities, and private operators?
How do we prevent “data monopolies” from becoming the new energy geopolitics?
If Dubhe (or any energy foundation model) is going to work at scale, the world will need new governance models for:
Interoperability (systems speaking the same language),
Trust (proof of integrity),
Privacy/security (critical infrastructure protection),
Fair value exchange (data contributors benefiting, not being extracted).
This is where the UAE and Abu Dhabi become a meaningful context, not just a launch location. The region is actively shaping how big systems get built: energy, hydrogen, grid upgrades, and strategic partnerships.
Dubhe + Tianji
weather becomes the “hidden parent” of renewables
Envision points out something simple and true: renewable energy depends on the weather.
That’s why Dubhe works in tandem with Tianji, Envision’s large-scale weather foundation model, providing predictive intelligence so Physical AI can anticipate weather-driven variability and keep the system efficient and reliable.
This matters for customers because the weather isn’t just “forecasting.” In modern energy systems, weather becomes:
price behavior,
volatility risk,
curtailment decisions,
storage timing,
grid congestion,
and sometimes — political pressure when reliability is threatened.
When prediction improves, decision-making improves. And in energy, better decisions mean fewer blackouts, lower balancing costs, and higher renewable utilization.
“Near-zero marginal” energy
A provocative idea, and why it’s seductive
The release says Dubhe aims to unlock renewable abundance and drive costs toward near-zero marginal levels, enabling energy systems to scale with AI and meet unprecedented energy demands.
Let’s be careful here: “near-zero marginal cost” doesn’t mean the system becomes free. It means the incremental cost of producing an additional unit of energy, once infrastructure exists and is optimized, can trend downward.
That’s seductive for three reasons:
It makes electrification easier (transport, industry, cooling).
It makes AI growth less constrained.
It offers a path to prosperity that doesn’t require breaking climate limits.
But it also raises the systems challenge: the cheaper the marginal energy, the more demand rises — and the more orchestration matters.
What does this change for customers?
(real-world business impact)
If you’re a customer — a government agency, a utility, a data center operator, a heavy industry company, or a renewable developer — you don’t care about poetic language. You care about outcomes.
Here’s what an AI energy orchestration layer can mean in practice:
Data centers and AI compute
Better matching between compute schedules and renewable availability.
Lower cost per MWh through smarter bidding/dispatch.
Improved reliability and backup strategy.
Cities and public infrastructure
Adaptive demand response that feels invisible to citizens.
Lower peak stress without sacrificing comfort.
Higher renewable penetration without instability.
Industrial and manufacturing players
Optimized load scheduling tied to price and carbon intensity.
Predictive maintenance on energy assets.
Less downtime from grid disruptions.
National energy strategy (especially in GCC/MENA)
Faster integration of wind, solar, storage, and hydrogen.
Higher system confidence when scaling renewables.
Better resilience planning for heat peaks and extreme weather.
Comparisons that matter
Dubhe vs. traditional “energy management software.”
Old systems often:
optimize within a single asset (one plant, one battery),
rely on static rules,
struggle with multi-system orchestration.
Dubhe is positioned as a foundation model layer that can coordinate across generation, storage, grid, and demand in real time.
Dubhe vs. generic AI models
Generic models can analyze and predict, but energy requires:
hard constraints,
safety thresholds,
physical causality,
real-time reliability requirements.
That’s why Envision frames this as “Physical AI,” not just AI in dashboards.
Dubhe vs. “weather tools.”
Weather tools predict weather. Tianji is positioned as a foundation model feeding predictive intelligence into the operating decisions of the energy system.
Leadership
The story Envision is telling about civilization
Lei Zhang, Founder and CEO of Envision, unveiled Dubhe at Abu Dhabi Sustainability Week and framed it as a guiding star of the AI era, supporting prosperity for a new civilization.
Whether you agree with the scale of that language or not, it signals something: Envision wants to be seen not as a vendor, but as an architect of the next infrastructure layer.
And Envision backs that narrative with the structure of its business:
Smart Wind Turbines
Energy Storage
Green Hydrogen Solutions
Plus a global R&D and engineering footprint across the US, UK, France, Germany, Denmark, China, and others.
Add to that sustainability commitments: Envision joined SBTi and committed to “Business Ambition for 1.5°C” (2021), achieved carbon neutrality across operations by 2022, and targets value chain carbon neutrality by 2028.
The Masdar partnership
Deployment matters more than headlines
A model can be brilliant and still become a lab trophy. Deployment is where truth lives.
Envision signed a strategic cooperation agreement with Masdar, witnessed by HE Dr Sultan Al Jaber (also noted as ADNOC Managing Director and Group CEO and Masdar Chairman), together with Lei Zhang.
Under the agreement, Masdar recognizes Envision as a strategic partner across renewable energy technology sectors including wind power, energy storage, and green hydrogen, to advance large-scale deployment of AI energy systems and accelerate the energy transition.
That matters because the UAE has become a place where new energy systems don’t just get announced — they get built.
FAQ
1) What is Dubhe by Envision?
Dubhe is Envision’s Energy Foundation Model designed to sit at the core of a Physical AI architecture, coordinating renewable generation, storage, grids, and demand in real time.
2) Why call it an “Energy Foundation Model”?
Because the intent is to create a reusable intelligence layer trained on vast real-world energy data streams — not a single-use tool.
3) What does “Physical AI” mean in this context?
Envision frames Physical AI as AI that governs physical systems essential to human survival and prosperity, beyond purely digital tasks.
4) How does Tianji relate to Dubhe?
Tianji is Envision’s weather foundation model. It provides predictive intelligence because renewables are weather-driven, allowing the energy system to anticipate variability and operate reliably.
5) Is this mainly about renewables?
Yes — the model is positioned to orchestrate renewable-heavy systems and make them scale reliably.
6) What problem is Dubhe trying to solve for AI growth?
The press release explicitly frames Dubhe as enabling energy systems to scale in step with AI by addressing unprecedented energy demands on global infrastructure.
7) What’s the customer benefit for utilities and grid operators?
Higher reliability, better balancing decisions, improved use of renewables, and more predictive operations.
8) What’s the customer benefit for data centers?
Better matching of compute loads with renewable supply, cost stability, and improved energy resilience.
9) Is Masdar involved in deploying this?
Envision signed a strategic cooperation agreement with Masdar to advance large-scale deployment of AI energy systems.
10) What technology areas does the Envision–Masdar agreement cover?
Wind power, energy storage, and green hydrogen are explicitly mentioned.
11) What kind of company is Envision?
A green technology company providing renewable energy solutions across smart wind turbines, energy storage, and green hydrogen solutions.
12) How credible is Envision’s sustainability track record?
The release cites EcoVadis Gold Medal, CDP “A-List,” Fortune “Change the World,” and top 10 smartest companies by MIT Technology Review.
13) Does Envision have climate targets?
Yes — it joined SBTi and committed to “Business Ambition for 1.5°C” (2021), achieved operational carbon neutrality by 2022, and targets value chain neutrality by 2028.
14) What’s the biggest challenge for energy foundation models?
Data quality, security, interoperability, and governance — because energy data is critical infrastructure data.
Envision’s Physical AI Blueprint for Renewable Power at Scale
Dubhe is interesting not only because it’s ambitious, but because it quietly shifts the conversation.
For years, we spoke about renewables as hardware: panels, turbines, batteries. Then we started speaking about grids as software: sensors, automation, and demand response. Now we’re being asked to speak about energy as intelligence — a system that doesn’t just respond, but anticipates, coordinates, learns, and protects stability while unlocking abundance.
If Envision’s bet is right, the next era won’t be defined by who has the smartest chat model. It’ll be defined by who can build the most reliable “brain” for the physical world — the thing that lets countries scale clean power without fear, lets industry electrify without disruption, and lets AI grow without turning energy into a choke point.
And maybe that’s the most hopeful part: the idea that the AI era doesn’t have to be powered by compromise — not for the planet, not for prosperity, not for human dignity. It can be powered by systems designed carefully, shared wisely, and governed like the public good they really are.
