Inspiration
Cities of the future cannot thrive without renewable energy. Yet today, renewable project planning is slow, expensive, and fragmented. It can take up to five years just to complete environmental assessments in Australia, and globally, most projects face years of delays. We were inspired to ask: What if this entire process could take minutes instead of years?
That idea became Griddy — a platform that leverages satellite data, weather patterns, and machine learning to find the best renewable energy sites instantly.
What it does
Griddy is an AI-powered renewable energy planning assistant. Users can type natural prompts like “I want a 100MW solar-powered data centre in Australia under $500M” and Griddy will:
Analyze solar, wind, and hydro potential in the region
Check land use, grid connectivity, and costs
Recommend the most viable sites in seconds
Provide alternatives with trade-offs (e.g., lower cost vs. better grid access)
In short, Griddy reduces renewable project planning from years to minutes — enabling governments, developers, corporates, and communities to act faster in building sustainable infrastructure.
How we built it
Satellite Data + Weather Patterns: Ingested global datasets of solar, wind, and hydro abundance.
Machine Learning Modelling: Trained models to predict energy generation and costs dynamically, under real-world and future climate conditions.
RAG + Prompt Interface: Built a conversational interface that connects natural language queries to our predictive engine, so users can ask questions like they would to an expert consultant.
Visualization Layer: Generated interactive maps and site overlays to make complex data instantly understandable.
Challenges we ran into
Integrating large-scale weather and energy datasets across regions while keeping the model fast and accurate.
Balancing accuracy vs. speed — feasibility studies are slow for a reason, and condensing that into seconds was technically challenging.
Designing a user interface that made highly technical modelling outputs intuitive for policymakers, developers, and non-technical users.
Hackathon time constraints meant prioritizing the MVP without losing sight of the big vision.
Accomplishments that we're proud of
Built a functioning MVP that can take natural prompts and return accurate site recommendations.
Created a novel ML model trained on global weather + energy data — enabling the most accurate predictive renewable modelling platform to date.
Validated the idea with shocking stats: planning delays of 3–5+ years and massive project bottlenecks.
Showed that Griddy can dramatically accelerate the renewable energy transition by removing planning as the bottleneck.
What we learned
The true barrier to renewable rollout isn’t technology — it’s planning and decision-making speed.
By combining AI, satellite data, and predictive modelling, it’s possible to transform feasibility studies into something real-time and interactive.
Storytelling is as important as tech — showing policymakers and developers the before vs. after impact is what gets people excited.
Hackathons force sharp focus: we learned how to distill a big global challenge into a tangible MVP.
What's next for Griddy
Expand the model beyond Australia to deliver fully global renewable siting intelligence.
Integrate economic modelling to compare ROI, financing, and carbon offset value of different projects.
Partner with governments, developers, and corporates to pilot Griddy on real renewable projects.
Build APIs so Griddy can plug into existing planning workflows, energy marketplaces, and policy tools.
Ultimately, make Griddy the essential software infrastructure that underpins the Cities of the Future.
Log in or sign up for Devpost to join the conversation.