EFFICIENCY / ↓
LESS ENERGY
PER TASK.
Smaller models, better hardware, inference optimisation and smarter scheduling can reduce the resource cost of useful work.
HACK4EARTH 2025 · GREEN AI
RETROSPECTIVE / 01FROM IDEA
TO PROOF.
A visual recap of how Green AI moved from concept into public experimentation.
In 2025, HACK4EARTH treated Green AI as a question to test in public. Builders were invited to either reduce the footprint of AI itself or apply AI to a measurable sustainability problem — and make the work visible.
MISSIONS2
Build Green AI / Use AI for Green Impact
TEAMS44
Competition teams recorded on Kaggle
ENTRANTS267
Public Kaggle entrant count
SUBMISSIONS169
Competition submissions recorded on Kaggle
01 / THE STATE OF GREEN AI
THE SIGNAL IS NOT ONE-DIRECTIONAL
MORE EFFICIENT.
MORE DEMAND.
Green AI is now best understood as a systems question: efficiency gains matter, but so does total demand.
New models, hardware and software stacks can reduce the energy needed for a single task. But AI is also spreading into more products, workflows and infrastructures. That is why the field still needs measurement, proportionate design and honest reporting rather than easy claims.
LESS ENERGY
DEMAND / ↑
MORE AI
EVERYWHERE.
As AI expands across sectors and use cases, total infrastructure demand can still grow even when individual tasks become more efficient.
Reference context: IEA · Key Questions on Energy and AI
02 / WHAT GREEN AI MEANS IN PRACTICE
THREE LAYERS.
ONE SYSTEM.
A compact reading of the field without collapsing everything into one headline claim.
COMPUTE
Do the same useful work with less.
Reduce unnecessary computation through right-sized models, lower runtime cost and more efficient processing choices.
INFRASTRUCTURE
Run workloads with better operational choices.
Consider utilisation, scheduling, hardware, electricity mix and the footprint of the surrounding infrastructure.
OUTCOME
Ask whether the application improves a real system.
The strongest claims are attached to observable results in domains such as food, water, energy, mobility or waste.
03 / 2025 · FROM IDEA TO PROOF
NOT “PROOF THAT AI IS GREEN”
PROOF THAT THE
QUESTION COULD BE BUILT.
The 2025 edition turned an abstract concern into a build brief with visible outputs.
01QUESTION
Can AI create measurable value while making its own footprint part of the design discussion?
02BRIEF
Two missions framed the challenge: build greener AI systems, or use AI for measurable environmental impact.
03BUILD
A public, open experimentation format created space for prototypes, comparisons and technical exploration.
04EVIDENCE
Projects made ideas inspectable through code, outputs, comparisons and documented implementation choices.
05CONTINUE
The lasting value was not a slogan but a clearer basis for what responsible technical work might look like next.
THE SHIFT
2025 asked: “Can AI be greener?”
2026 asks: “What does responsible engineering look like when whole systems are under pressure?”
04 / SUBMISSION TRAIL
SEE THE
WORK.
Minimal copy, direct routes.