AI for Earth grants
Join hundreds of grantees and empower your work with an AI for Earth grant. Next application cycle ends on April 8, 2019, 11:59 PM, Pacific Time.
AI for Earth grants
Visit our interactive map to see how grantees are using AI for Earth research grants to empower their work
AI for Earth awards grants to support projects that change the way people and organizations monitor, model, and ultimately manage Earth’s natural systems. Depending on project need, our grants can award Microsoft Azure cloud computing resources (including AI tools) and/or data labeling services.
To date, we have awarded 202 grants to projects with impact in 57 countries, and we are committed to growing this community of grantees.
AI for Earth grants provide access to advanced Microsoft Azure cloud computing resources to support projects that change the way people and organizations monitor, model, and ultimately manage Earth’s natural systems.
What types of grants do you offer?
Currently AI for Earth provides two types of grants. Each round, you can apply to either, or both, depending on your project needs.
- Data labeling services: An important prerequisite for most AI projects is having an accurately labeled dataset to train your model on. Thus, we provide grants to help label key datasets in our four environmental focus areas. All datasets that are labeled through our grants program are hosted on Azure and made publicly available to other organizations and individuals for training models. The amount granted is dependent on size of dataset and difficulty of labeling.
- Azure compute credits: If you already have access to a labeled dataset and are ready to start computing in the cloud and accessing Azure AI tools, this grant provides you with Azure compute credits worth $5,000, $10,000, or $15,000 (depending on your project scope and needs). By being a member of the AI for Earth grantee community, you also have access to additional resources (technical advice and support, online Azure training materials, as well as invitations to the AI for Earth Summit for networking and education opportunities).
What environmental areas does the program focus on?
AI for Earth awards grants to projects that address four areas that are vital for building a sustainable future:
Agriculture: To feed the world’s rapidly growing population, farmers must produce more food, on less arable land, and with lower environmental impact. AI can help in areas like:
- Land-use planning and management
- Natural resource conservation
- Sustainable supply chains
- Climate resilient agriculture
Biodiversity: Species are going extinct at alarming rates, driving the decay of key ecosystem services like pollination on which humans depend. AI can help in areas like:
- Habitat protection and restoration
- Sustainable trade
- Invasive species and disease control
- Pollution control
- Realizing natural capital
Climate change: Extreme weather events, rising sea levels, higher global temperatures, and increased ocean acidity threaten human health, infrastructure, and the natural systems we rely on for life itself. AI can help in areas like:
- Climate resilience
- Extreme weather and climate modeling
- Sustainable land-use change
- Ecosystem services (including carbon sequestration)
Water: In the next two decades, demand for fresh water—for human consumption, agriculture, and hygiene—is predicted to dramatically outpace supply.
- Water supply (including catchment control)
- Water quality and sanitation
- Water efficiency
- Extreme event (droughts, floods, disasters, etc.) management
- Healthy oceans
How do I apply for a grant?
Submit your proposal via the online application form.
The grant application includes an online form with two sections, one for data labeling services and one for Azure credits. Fill out either or both, depending on the resources you need for your project. For Azure compute grants, you will also fill out a written project proposal (details below) that is uploaded in the Files section of the Online Application System.
We accept proposals on a rolling basis and review them four times a year. Our next evaluation will be for all proposals submitted by 11:59 PM Pacific Time, April 8, 2019. This is the second deadline for 2019.
Applicants can be affiliated with an academic institution, nonprofit organization, government entity, environmental start-up, or an innovative project within a company. For the Azure compute grants, we recommend that the main applicant has a demonstrated background in environmental science and/or technology, and that at least one member of the team has strong enough technical skills to complete the project successfully. Applicants should be close to or done with their data collection and ready to start with computation and model building.
Esri will also consider AI for Earth grant recipients who are affiliated with academic institutions, nonprofit organizations, and start-ups for a sponsored subscription to ArcGIS Pro. ArcGIS Pro is a leading geospatial software for creating maps, performing spatial analysis, and managing data.
The most important part of each Azure compute grant application is a project proposal that describes the environmental challenge addressed, the datasets used for analysis, the technical solution proposed, and its potential impact. Proposals should be written in English and not exceed three pages. While there is no specific format for proposals, a guideline is below that outlines information that is important to include:
1-2 sentence summary of your project
Cloud computing and data science tools
Grantees get access to Microsoft Azure cloud computing resources and a variety of data science and machine learning tools, including powerful assets in the the Microsoft Cognitive Toolkit and the GeoAI Data Science Virtual Machine (DSVM). Grantees will receive 12 months of free resources worth $5,000, $10,000, or $15,000, depending on project needs.
Learn more about Microsoft tools that grantees can use:
- Overview of Microsoft AI platform: services, infrastructure, and tools
- Learn more about the GeoAI Data Science Virtual Machine
- Create a GeoAI Data Science Virtual Machine
- Microsoft Cognitive Toolkit is now generally available
- Learn more about Microsoft Cognitive Toolkit
- Download the Microsoft Cognitive Toolkit from GitHub (You can install it locally or use it from one of the data science virtual machines.)
- Learn about the beta release of Microsoft Cognitive Toolkit
- Watch a Microsoft Cognitive Toolkit video
- Learn how Microsoft AI is amplifying human ingenuity with intelligent technology
- Learn more about Azure Machine Learning Services
- Learn more about Jupyter notebooks on Azure
- Learn more about Cognitive Services, pre-trained models for common scenarios
Grantees affiliated with academic institutions, non-profit organizations, and start-up institutions also have the potential to access Esri ArcGIS Pro a next-generation, 64‐bit GIS desktop providing 2D and 3D mapping capabilities in an intuitive user interface. ArcGIS Pro features advanced visualization, analytics, image processing, and data management capabilities.
The new Geo AI DSVM integrates ArcGIS Pro into the Microsoft DVSM on Azure, to seamlessly tie geospatial data management, analysis and visualization capabilities with data science and machine learning tools. You can provision your own GeoAI instance and refer to this tutorial on land cover classification.
Submit your application via the online application form.
You will need to include a proposal written in English, no longer than three pages. This is the most important component of your application, and should include concise descriptions of the following:
- What datasets will you use for your analyses? Is this data already collected?
- Approximate timeline for key project milestones
- What Azure offerings will your project use?
- We are particularly interested in projects where researchers are making use of higher-level Azure services such as machine learning (in other words, not just virtual machines and storage). You can see what’s available at Azure Cloud Services.
- Do you have the resources and technical skills to complete the project successfully?
- How does your project align with the AI for Earth areas of focus (agriculture, food, biodiversity, and/or climate change)?
- How does your project help transform the way we address environmental challenges?
- How will the results of your project be disseminated (for example, academic publications and open-source solution) and be beneficial to a community of users?
Applicants can be affiliated with an academic institution, non-profit organization, environmental start-up, or even an innovative new project within a company. We recommend that the main applicant has a demonstrated background in environmental science and/or technology (such as a PhD degree), and that at least one member of the team has strong enough technical skills to complete the project successfully.
To qualify for access to ArcGIS Pro through Esri, the grant recipient must be affiliated with an academic institution, non-profit organization, or start-up organization.
To estimate the monetary value of Azure computing resources you need, use the Azure calculator. Select all the products you would need for your research (including storage, virtual machines, and so forth) and enter these specifications into the Azure calculator. For your data, consider both the options under “Storage” as well as the options under “Databases”.
- Determine which size of VM you need; if you require machine learning models that use large amounts of data, heavy graphic rendering, or video editing, reference GPU-optimized VM sizes.
- Determine which geographic regions have your VM of choice.
- Note: GPUs are often in high demand, and thus are not always available in each region. If you need a GPU-optimized VM, try selecting a few different geographic regions until you can find your desired VM type.
Data labeling services: An important prerequisite for most AI projects is having an accurately labeled dataset to train your model on. Thus, we provide grants to help label key datasets in our four environmental focus areas. All datasets that are labeled through our grants program are hosted on Azure and made publicly available to other organizations and individuals for training models.
Azure compute credits: If you already have access to a labeled dataset and are ready to start computing in the cloud and accessing Azure AI tools, this grant provides you with Azure compute credits worth $5,000, $10,000, or $15,000 (depending on your project scope and needs). By being a member of the AI for Earth grantee community, you also have access to additional resources (technical advice and support, online Azure training materials, as well as invitations to the AI for Earth Summit for networking and education opportunities).
You may apply for both grants
AI for Earth grant applications are reviewed by Microsoft employees who are directly involved in the AI for Earth grant selection process for the sole purpose of proposal review and to determine the grant level to be provided.
Applications are evaluated on their scientific merit, innovative use of technology, and potential impact.
This is an ongoing program, with proposals evaluated four times a year. Our next evaluation will be for all proposals submitted by 11:59 PM Pacific Time, April 8, 2019. This is the second deadline for 2019.
After each deadline, applicants will receive a decision by email within four weeks.
Yes. This award program is available worldwide.
Yes. On occasion, we offer special grants, such as the recently closed AI for Earth European Union (EU) Oceans Award and the AI for Earth Innovation Grant (offered jointly with National Geographic Society). Information on special grants will be available on this website.
If you have questions about the awards program or the application process not answered in this FAQ, please email us at email@example.com.
Occasionally, AI for Earth offers unique award opportunities. Current and past opportunities are listed below.
- There are no current special grant opportunities.