AWS opens applications for the Compute for Climate Fellowship 2026

In 2023, we launched the Compute for Climate Fellowship in partnership with the International Research Centre on Artificial Intelligence (IRCAI), an organization under the auspices of the United Nations Educational, Scientific, and Cultural Organization (UNESCO). The Fellowship was a first-of-its-kind global R&D funding program that provides grants in the form of AWS credits for groundbreaking POCs that think big and use advanced cloud computing to address the climate crisis.

Over three cohorts, the program has supported 35 startups around the world, from fusion energy companies in the United States to virtual power plant builders in South Africa and ocean biodiversity mappers in France to AI crop scientists in the United Kingdom. Each Fellow company builds a groundbreaking POC climate solution funded by AWS with technical and scientific advisors from both AWS and IRCAI.

In 2026, IRCAI and AWS will award up to $2 million USD in AWS credits to up to 10 companies from around the world, with individual grants of up to $200,000 per startup.

Startups can apply for the 2026 Compute for Climate Fellowship starting September 1, 2026. The deadline for proposals is September 30, 2026 at 11:59pm EST.

The program is global and startups from all countries can apply. The Fellowship selects proposals that think big, demonstrate innovative use of advanced cloud computing, and show the greatest promise of global impact.

Proposals must address at least one of the following solution areas in the fight against climate change:

1. Clean energy, grid flexibility, and fusion/nuclear simulation—generation, storage, distribution, and intelligent management of clean energy, including nuclear and fusion reactor simulation, renewable energy forecasting, virtual power plants, grid digital twins, and hydrogen production modelling.
2. Climate intelligence, earth observation, and risk analytics—data infrastructure and analytics systems that measure, monitor, predict, and communicate climate-related risks, including geospatial AI, climate and weather modelling, greenhouse gas measurement reporting and verification, methane detection, and multi-hazard warning systems.
3. Sustainable agriculture, food systems, and water security—technologies that transform food production and water systems through precision agriculture, alternative proteins, autonomous farming, physical AI and robot automation, drought resilience, and climate-smart supply chains.
4. Circular economy, industrial simulation, and low-carbon materials—technologies that decarbonize industrial processes and enable circular resource flows through generative AI materials discovery, industrial digital twins, manufacturing optimization, and critical minerals traceability.
5. Carbon removal and ecosystem restoration—technologies that remove CO2 from the atmosphere and restore natural ecosystems, including direct air capture simulation, enhanced weathering, biochar optimization, blue carbon, and automated MRV for permanence verification.
6. Low-carbon transportation, logistics, and autonomous systems—technologies that decarbonize the movement of people and goods through autonomous vehicle simulation, EV battery modelling, fleet electrification, route optimization at scale, and urban motility simulation.
7. Sustainable built environment and data center infrastructure—technologies that optimize the energy performance and climate resilience of buildings and data centers through cooling simulation, carbon-aware workload scheduling, thermal digital twins, and waste-heat recovery.

Across all seven solution areas, the Fellowship especially encourages community-led and Indigenous-led approaches that combine advanced technology with local knowledge, lived experience, and culturally appropriate mitigation, adaptation, and resilience strategies. The Fellowship recognizes that Indigenous Peoples are important leaders in addressing climate change and welcomes proposals from innovators in Indigenous communities worldwide.

Given the record-breaking heatwaves witnessed in 2026, we particularly encourage proposals that harness advanced computing and AI to anticipate heat risks, protect vulnerable communities, strengthen energy and infrastructure resilience, and enable more effective adaptation to a rapidly warming climate.

The 2026 Fellowship is also interested in proposals that make innovative use of frontier AI and advanced compute technologies, including:

Core compute technologies:

• AI/ML and generative AI—foundation models (including geospatial, weather, and domain-specific models), and agentic AI systems
• Physical AI and robotics—robot learning, simulation-to-real training, and embodied foundation models for field, industrial, and mobility applications
• High-performance computing—GPU-accelerated training, inference, and scalable simulation
• Physics-informed and scientific ML—neural operators, surrogate modelling, and scientific ML for physical systems
• Climate and weather modelling/earth observation—downscaling, scenario analysis, and geospatial AI
• Digital twins—for energy systems, infrastructure, industry, natural systems, and cities
• Quantum and quantum-inspired optimization—for climate modelling, materials discovery, routing, or energy systems

Enabling technologies:

• Automated measurement, reporting, and verification for emissions, carbon removal, biodiversity, and climate finance
• Large-scale optimization requiring distributed compute
• Time-series analytics, IoT, edge computing, and real-time monitoring at scale
• Privacy-preserving, secure, and interoperable climate data infrastructure

Over the past three years, the Compute for Climate Fellowship has supported 35 startups around the world. Here’s a few examples of how these companies are using advanced computing to solve some of the toughest climate and energy challenges:

• Asoba (South Africa)—building intelligence infrastructure for distributed energy markets in Africa, helping energy operators improve asset performance and unlock new revenue opportunities through AI-driven virtual power plants and microgrid optimization.
• Cosma (France)—mapping deep-sea ecosystems with swarms of autonomous underwater drones and AI, capturing up to 100,000 seabed images per hectare to monitor ocean biodiversity at scale.
• Smartex (Portugal)—using AI-powered computer vision to detect textile defects in real time, preventing fabric waste before it happens and helping manufacturers reduce their environmental footprint.
• Hum.AI (Canada)—building generative AI foundational models to track, predict, and protect the natural world, working closely with Canadian First Nations communities on coastal monitoring and ecosystem restoration.
• Phytoform Labs (United Kingdom)—developing AI models that invent novel crop genetics to create more climate-resilient plants, securing the food of tomorrow.
• Realta Fusion (United States)—developing compact magnetic mirror fusion energy systems and running first-of-a-kind plasma stability simulations in the cloud.
• Thea Energy (United States)—developing proprietary stellarator fusion technology with simpler, planar programmable magnets to reach commercial-scale carbon-free fusion energy faster.

Through the Fellowship, these companies have built POCs that accelerated their R&D and opened up new opportunities for growth.

“Advances in AWS cloud computing have paved a new pathway for Realta Fusion to expedite research in magnetic confinement fusion energy systems. Such intensive plasma physics simulations have never been done in the cloud before. The Realta Fusion team is grateful for this fellowship award and excited to use the power of cloud computing to advance fusion energy research.”

– Cary Forest, Co-Founder and Chief Scientific Officer at Realta Fusion

“The Compute for Climate Fellowship is absolutely crucial for Phytoform’s mission of improving sustainability in agriculture. With help from AWS and IRCAI we were able to really accelerate our solution to deploy a unique approach to enhance R&D and secure the food of tomorrow.”

– Nicholas Kral, CTO at Phytoform Labs

Selected companies will receive:

• Grants in the form of AWS credits to cover the cloud computing costs of the POC (up to $200,000 USD per startup)
• Three-month POC build period with personalized technical support and mentorship
• Access to advanced computing services including HPC, generative AI, and quantum computing tools
• Expert guidance in advanced computing, AI, sustainability, and ethics from IRCAI and AWS mentors
• Media and visibility opportunities including speaking opportunities and PR on a case-by-case basis
• A showcase of results across AWS channels, United Nations events, and international bodies

All POCs will be designed under the guidelines of UNESCO’s Ethics Impact Assessment for AI, ensuring each solution is built with safe, responsible technology.

Startups that apply but are not selected will have access to up to $5,000 USD in AWS credits and be invited to join the IRCAI Industrial Club.

• Applications open: 1 September, 2026
• Applications close: 30 September, 2026
• Review and finalist interviews: October-November 2026
• Winners announced: December 2026
• POC build period: January-April 2027
• Demo Day: April 2027

To apply for the Compute for Climate Fellowship and find out if your startup is eligible, visit https://ircai.org/ircai-aws-climate-fellowship-program-2026/.

The Compute for Climate Fellowship is a partnership between Amazon Web Services (AWS) and the International Research Centre on Artificial Intelligence (IRCAI).