Global Winners
Six student teams from around the world. Two tracks. One mission — building real solutions for a sustainable future.
6
Winning Teams
14
Countries represented
5
World Regions
Awards Ceremony · Garching, Germany
The awards ceremony in Munich
The Global Sustainability Challenge Global Finals wrapped in Garching, Germany, and the range of problems these teams chose to tackle tells you something important about this generation.
They see the sustainability challenges in their own communities — from peatland degradation to post-harvest fish loss to microplastics in household wastewater.
With 3,000+ students, 800+ teams, 90+ countries, and 40 finalists from five world regions — every one of them is a winner.
Sustainable Energy
Develop scalable and impactful solutions that reduce emissions while improving affordability and reliability in the energy sector. Solutions may address clean energy covering generation, transmission, distribution, or storage.
Solutions included
- solar-powered cold storage for fishing communities with no refrigeration access
- AI-native solar microgrids anchored at rural telecom towers
- dual-phase flood energy from waterwheels and magnesium-air fuel cells
- LiFePO4 battery systems delivering pay-per-use energy as a service
- LSTM load forecasting and reinforcement learning for off-grid power dispatch
- emergency power for communities when disaster takes the grid down
8 countries represented
CoolFish
Kazakhstan · Bosnia & Herzegovina · Pakistan · Indonesia · Bangladesh
Preserve the catch. Power the Community.
In coastal fishing communities across Indonesia, Africa, and the Caribbean, up to 80% of the catch is lost before it can be sold — not because fishers fish poorly, but because there's nowhere cold to store it. Unreliable grids and expensive diesel make conventional refrigeration out of reach. CoolFish's answer is a solar-powered cold storage unit (400W PV, LiFePO4 battery, 300kg/day capacity) delivered as a service: fishers pay ~$1/crate/day, no ownership required. The compact, foldable design ships to even the most remote islands. Each unit serves ~100 fishers, projects a 60% reduction in post-harvest waste, and offsets 7–10 tons of CO₂ annually — while creating women-led operator jobs in the communities it serves.
Team · 6 members
Maral Shagatay
Aldijana Hodzic
Baleegha Fatima
Matthew Justin Lesmana
Md Atiq Aziz
Fadel Zikrillah
PowerBlocks
Jordan
AI-native solar microgrids for communities living with unreliable or no electricity
730 million people lack access to electricity — 96% of them in Africa and South Asia. PowerBlocks is building the infrastructure to change that: containerized solar microgrids (35 kWp generation, 50 kW capacity, 100 kWh storage) anchored at rural telecom towers, with excess energy distributed to schools, clinics, and households nearby. What sets PowerBlocks apart is the software layer — an AI system using LSTM load forecasting and reinforcement learning to intelligently dispatch power across Normal, Scarcity, and Crisis modes. The hardware deploys in 2–4 weeks versus 6–12 months for competitors. They're in active discussions with Orange Jordan x UNRWA for a pilot at Baqa'a Refugee Camp (80,000 refugees), and with Sudan's Ministry of Telecommunications to serve as energy backbone for a 4G rollout across 200 villages reaching up to 1.4 million people.
Team · 2 members
Yara AlSiddig
Dana AlSiddig
Junipero Energy
USA · Greece
Clean emergency power from floodwater itself — for communities when the grid goes down
When floods hit, the grid goes down — and the most common backup options (diesel generators, battery packs) either can't operate in floodwater or can't recharge. By 2030, 15 million people will be exposed to flood conditions annually. Junipero Energy's answer is HydroVolt: a dual-phase system that turns the disaster itself into an energy source. A 3D-printed waterwheel (1:30 gearbox ratio, 24V 350W motor) mounts in roof gutters to capture flood currents. Simultaneously, shelf-stable magnesium-air fuel cells use seawater as electrolyte — no charging required, works in both still and running water. In testing, the system achieved 3.76W using carbon cloth with MnO₂ catalyst, surpassing the threshold for phone charging, LED lighting, and emergency radio. Initial deployment targets are Guam and Houston. The team are Stanford Class of 2027 engineers spanning materials science, mechanical, electrical, and management science.
Team · 4 members
Daniella Fenster
Panagiotis Papanastasiou
Miles Bliey
Connor Hoffman
Adaptation & Resilience
Develop a solution that enhances the resilience and ability of people, infrastructure, or natural systems to withstand and adapt to the intensifying impacts of climate change. Solutions may address disaster preparedness, resilient infrastructure, climate-smart agriculture, water security, or other critical adaptation needs.
Solutions included
- 7-day-advance pest infestation prediction using hyperspectral satellite imagery and AI
- microplastic filtration at source using chitosan-coated banana fibre and activated carbon
- peatland restoration funded by Verra-verified carbon credits and satellite-verified payments
- bark beetle outbreak monitoring across 280+ traps validated to 86% accuracy
- community-driven restoration tasks dispatched by peat topology to the nearest metre
- laundry microplastic capture from 100nm to 100 microns — broader than any competitor
7 countries represented
PlaNX Sense
Germany · China · Egypt · India
7-day-advance pest infestation risk prediction for ecosystems
Pests destroy 40% of global crop yields annually and have damaged 72% of timber in Germany. When an outbreak hits, land managers find out too late — experts confirmed that critical information routinely arrives up to 7 days after the window to act has closed. PlaNX Sense flips that: their AI prediction engine fuses weather forecasting, multi- and hyperspectral satellite imagery, and field observations to generate location-specific infestation risk maps 7 days in advance. Forest managers get enough lead time to focus ground monitoring, time harvests, and target spraying — preserving timber value and reducing pesticide use. Trained and validated on 280+ bark beetle traps across Bavaria, the model achieves 86% accuracy on 7-day-advance predictions (84.2% mean cross-validation accuracy, April–October 2024). They're currently in product deployment with the Bavarian State Ministry for Food, Agriculture and Forestry and TUM.ai, backed by ESA's Business Incubation Centre, and accelerating into agriculture with a SaaS model targeting large land holdings.
Team · 5 members
Janina Fraas
Miaoxi Chen
Heba Gaballa
Ankur Arun Giri
Adesh Sanjay Phalphale
EcoMolecule
India · Indonesia · Germany · United Kingdom
Stopping microplastics at the source — before they leave your washing machine
35% of all microplastics in the ocean come from laundry. Every wash cycle releases synthetic fibres too small to see — and the average person already consumes 50,000 microplastic particles per year. France mandated microplastic filters in all new washing machines from January 2025, with the EU, UK, US, and Australia following. EcoMolecule's answer is the ChitoTrap: a retrofit filter using chitosan-coated banana fibre mesh and granular activated carbon to capture microplastics from 100 nanometres to 100 microns — achieving 95–97% reduction in lab validation. Competitors like PlanetCare and Cleanr capture particles above 10–50 microns and cost €120–300. ChitoTrap captures a broader particle range at €59.90, with no electricity required and easy cartridge replacement. Their first prototype worked at Hostel 7, IIT Bombay. In the six weeks before the GSC Global Finals, they received 10 validations from Bosch, LG, Whirlpool, and Reliance — and secured 3 signed letters of intent, including one from Whirlpool's lead engineer for new product development.
Team · 4 members
Harsh Raj
Veronica Maggie Koesnadi
Madina Makhmudkhodjaeva
Alice Shiyu Han
PeatGuard
Singapore
Restoring peatlands. Paying communities. Turning degradation into verified carbon credits.
Peatlands store twice the carbon of all the world's forests combined — and Indonesia's 13.4 million hectares make it ground zero for both climate risk and opportunity. 7 million of those hectares are already degraded, releasing carbon at scale. The problem isn't awareness — it's incentives: existing carbon credit markets reward passive protection, not active restoration, leaving local communities with no mechanism to get paid for the work of healing the land. PeatGuard changes that with a four-step platform: Detect → Restore → Pay → Generate. Satellite L-band InSAR and VV-backscatter imagery map peat subsidence to 0.6mm/year accuracy and quantify carbon emissions to 0.97 tCO₂/ha/yr — within Verra MRV requirements, validated by Prof. Sang Ho Yun at the Earth Observatory of Singapore and published in Nature Portfolio 2025 against 6 years of field data across 8 sites in Central Kalimantan. A mobile app (Peat Guardian) dispatches geo-tagged restoration tasks — blocking canals, drilling wells, placing sensors — to exact coordinates derived from peat topology. Satellite verification triggers real-time payments to villagers via Bank Indonesia's BI Fast system, linked to their phone number. Verified restoration generates carbon credits under Verra VM0027, sold to banks and transition finance buyers. Their pilot in Jawai and Teluk Keramat, West Kalimantan, has a signed Letter of Intent with Sungai Nilam Village and projects 300,000 tCO₂e of annual sequestration potential.
Team · 4 members
Swedha Prabakaran
Jia Yuun Chia
Delfina Poernomo
Louise Goh Ruo Si
Inspired?
Applications for the 2026–27 Global Sustainability Challenge are opening soon.
Pre-register for 2026–27