You are currently viewing AI for Climate and Nature Grand Challenge

AI for Climate and Nature Grand Challenge

The Bezos Earth Fund is exploring new ideas for multiplying the impact of climate and nature efforts using modern AI.

The AI for Climate and Nature Grand Challenge is a global $100 million initiative from the Bezos Earth Fund. The first round of awards will focus on sustainable proteins, power grid optimization, and biodiversity conservation, in addition to embracing visionary wildcard solutions for climate and nature.

Focus Areas
  • Sustainable proteins
    • While the current generation of sustainable proteins may fall short of convincing consumers of their ability to deliver a product equal or better to animal-sourced meat, they have only touched the surface of the possibilities in terms of optimizing ingredients and processing. New approaches to discovering and unlocking the myriad of ingredients and processing methods that deliver a product that best mimics the complexity of meat will bring about a new generation of sustainable protein products that can meet expectations of consumers. Some of the opportunity areas for using AI-based approaches to enhance the quality and cost of sustainable proteins include:
      • End-product formulation: The complexity of food ingredients like “protein isolates,” which contain thousands of molecules, make it difficult to have predictive capacity to determine which formulation adjustments will deliver the desired effects. Instead of a scientist doing this manually, with so many combinations of variables, AI can help them try and fail many times, quickly.
        • For example, scientists have recently used AI to find an isolated protein in mung beans that has similar properties to scrambled eggs. Without AI, it would have taken years to identify this solution. 
      • Protein engineering: The production of proteins themselves is central to the shift to sustainable proteins. Proteins have evolved over billions of years and are comprised of chains of amino acids that fold in complex three-dimensional shapes. AI can be used to design such patterns by modifying genetic sequences to optimize for certain properties. AI allows protein formulation to move from an iterative art to more of a predictive science. 
      • Strain/cell line development: Regardless of the organism being used to develop a sustainable protein product (plant crops for plant-based meat, microbial strains for fermentation-derived proteins, or animal cell lines for cultivated meat), AI can improve the understanding of what modifications to these organisms would facilitate greater growth efficiency or bias them toward desirable traits. 
      • Feedstock optimization and cost reduction: For fermentation and cultivated meat, feedstocks for cell growth are a major contributor to unit economics and are highly influential in the overall sustainability of the production process. Optimizing feedstocks and culture media formulations is a highly complex process that requires multi-variate analysis and nuanced trade-offs between performance, cost, availability, and many other factors. Next-generation growth media can use AI to test thousands of novel combinations of ingredients and rapidly adapt feedstock formulations to changes in feedstock cost/availability. 
      • Process development and efficiency improvement: Digital twin technologies can facilitate rapid process development improvements for greater efficiency using manufacturing process simulations. Biosensors and process monitoring technologies coupled with AI can inform process tweaks for subsequent runs without the cost- and time-intensive need to test multiple parameter tweaks in an actual pilot- or demo-scale run.
  • Power grid optimization
    • AI emerges as a potential game-changer in addressing some of the challenges faced by electricity grids globally. By optimizing grid operations, anticipating and mitigating disruptions, and supporting the seamless integration of variable renewable energy (VRE) into grid operations, AI can accelerate the clean energy transition while enhancing grid resilience for communities. AI-driven solutions have the potential to optimize grid operations, saving an estimated $300 billion in efficiency gains within this decisive decade alone, according to one study.
    • Recent advancements in AI technologies, such as predictive modeling, real-time data analysis, and intelligent automation, offer unique opportunities to transform the grid system. These advancements can be leveraged to optimize transmission siting, enhance power flow, improve renewable energy forecasting, and efficiently coordinate distributed energy resources such as rooftop solar installations on homes. By harnessing the power of AI, they can create a more efficient, sustainable, and cost-effective grid that seamlessly integrates renewable energy sources, as many countries are striving to decrease the reliance of their grids on fossil fuels.
    • Opportunity areas include:
      • Grid optimization: How might AI optimize power flows, reduce congestion and enable seamless integration of renewable energy sources to create a more efficient, sustainable and cost-effective grid? By leveraging predictive models, AI can provide a solution to optimize transmission siting, enhance power flow and dynamically increase transmission capacity. These optimizations can lead to significant cost savings and facilitate the accelerated integration of variable renewable energy sources, such as solar and wind, into the grid. Specific use cases of AI applications in grid optimization include predictive maintenance, renewable energy forecasting, and dynamic grid control. 
      • Energy management: How might AI enable intelligent demand response, improve renewable energy forecasting, and efficiently coordinate distributed energy resources to balance supply and demand in real time? AI can analyze vast amounts of data from smart meters, weather forecasts, and consumer behavior to predict energy demand accurately. This enables the grid to intelligently shift loads to match renewable energy supply, optimize energy storage and coordinate distributed energy resources, such as electric vehicles and battery storage, to maintain a stable and efficient grid. 
      • Resilience and inclusivity: How might AI enhance grid resilience, predict and mitigate outages, and ensure reliable, affordable, and clean electricity access for all, particularly in the face of increasing climate change-related disruptions? AI can analyze historical data, weather patterns, and grid performance to predict potential outages and better inform future preventive maintenance. By automating fault detection, isolation, and service restoration, AI can significantly reduce the duration and impact of outages. Moreover, AI can help identify and prioritize underserved communities, ensuring that the benefits of a modernized grid are distributed equitably and that no one is left behind in the transition to a clean energy future.
  • Biodiversity conservation
    • AI applications offer a range of tools and technologies that significantly contribute to monitoring, analyzing, protecting, and restoring diverse ecosystems. One key application is the use of AI in data analysis, where machine learning algorithms can process vast datasets, identify patterns, and predict changes in biodiversity in space and time that can support conservation efforts. This aids researchers in monitoring species populations and trends, tracking migration patterns, and assessing ecosystem health. It can additionally help the conservation community identify the highest conservation priority areas, informing the design of wildlife corridors and protected areas. Furthermore, machine learning algorithms can assist in identifying endangered species, combating illegal poaching, and managing invasive species. Integrating AI into conservation strategies enhances the precision and speed of decision-making, ultimately contributing to the preservation of biodiversity in the face of numerous environmental challenges. It’s also worth noting that the abovementioned applications need not apply only to wild species. They can also be used on domesticated species, most notably cattle. Computer vision and facial recognition platforms would avoid the need for physical tagging, and AI can also leverage satellite imaging and machine learning algorithms to trace cattle movement across large tracts of land.
    • Some of the potential tools and uses that could be used with AI for biodiversity conservation include:
      • Endangered species monitoring: Important species that are targets for conservation can often be elusive, hard to track, and unknown in terms of behavior and population numbers. Camera traps have opened a new frontier by capturing images of elusive species in different ecosystems, and today there are millions of images that can be used to identify and monitor biodiversity, but they don’t have the capacity to analyze them. 
      • Discovery of hidden biodiversity: The vast majority of species that they conserve are neither seen nor heard and come in the form of microbes and fungal communities. Environmental DNA is a tool that is transforming the ability to do conservation science. One notable example is the use of AI in eDNA metabarcoding analysis. Machine learning algorithms can process large volumes of genetic data extracted from environmental samples, identifying and cataloging species present in ecosystems more rapidly and accurately than traditional methods. This accelerates biodiversity assessments, aiding in the monitoring of endangered or elusive species. Additionally, AI can contribute to the development of predictive models that analyze eDNA data alongside environmental variables, helping researchers understand how different factors impact biodiversity. 
      • Bioacoustics tools: These tools allow them to build soundscapes, comprised of the diverse sounds produced by ecosystems, that offer valuable insights into the health and dynamics of natural environments. This type of tracing can aid conservation and rewilding efforts by furthering understanding of population numbers, preferred habitats, and human interactions. AI algorithms, particularly those designed for audio analysis, can efficiently process large volumes of acoustic data, distinguishing and identifying individual species’ vocalizations. This capability enables researchers to monitor biodiversity, assess population dynamics, and detect changes in ecosystems over time. AI-driven acoustic monitoring systems facilitate real-time data collection, aiding in the early identification of disturbances or threats to wildlife. Additionally, AI can contribute to the creation of responsive soundscapes, mimicking natural conditions to attract and support specific species. By harnessing AI in the analysis and management of soundscapes, conservationists can better protect and restore nature.
  • Wildcard solutions
    • Climate change and nature loss are complicated, urgent issues — and they want to understand how modern AI can help identify and accelerate solutions. Do you have an idea for driving impact within climate change mitigation or nature protection using AI? If you and your team have a transformative idea that doesn’t fit within other first-round focus areas, they encourage you to submit it in the wildcard category.
Funding Information
  • The Grand Challenge will offer up to $100 million in total funding over three rounds.
  • Each round will include the following two phases: In Phase 1, up to 30 Seed Grantees will receive $50,000 each and an invitation to join the Phase 2 Innovation Sprint. At the end of Phase 2, up to 15 Implementation Grantees will receive up to $2 million each.
How the Grand Challenge will unfold?
  • The Grand Challenge will unfold across three rounds and offer up to $100 million in total funding. The first round will include the following two phases.
    • Phase 1: Proposals
      • The first phase will award up to 30 seed grants for promising ideas that multiply impact using modern AI. Phase 1 is open to all eligible applicants.
      • During the first phase:
        • Potential applicants will have access to a virtual information session, webinars introducing AI and discussing the focus areas, and a collection of curated resources as they develop their proposals.
        • A submission should articulate the proposed solution and the problem it addresses, the near-term and long-term impacts the solution will have, the potential for scaling the solution, the resources needed for development, and the approach to responsible development of the solution.
        • Phase 1 submissions are due by 5:59 p.m. Eastern Time (9:59 p.m. UTC) on Tuesday, July 30. Applicants must accept the Grand Challenge Agreement to submit a proposal.
        • An expert review panel and multidisciplinary judging panel will evaluate submissions according to Phase 1 selection criteria. Based on their evaluation, judges will recommend up to 30 awardees.
        • The Bezos Earth Fund will determine the final slate of Seed Grantees. After completing a Bezos Earth Fund grant agreement, each Seed Grantee will receive $50,000 and an exclusive invitation to participate in Phase 2.
    • Phase 2: Implementation plans
      • Seed Grantees will enter a virtual Innovation Sprint; these grantees will receive targeted support and resources from AI experts as they develop detailed implementation plans. Phase 2 is open exclusively to Seed Grantees.
      • The virtual Innovation Sprint will include:
        • Teaming support. Seed Grantees will build partnerships to enable development and implementation of their proposals.
        • Expert mentorship. Expert mentors will guide Seed Grantees as they refine their proposals and prepare for implementation.
        • Webinars. Seed Grantees will join exclusive webinars to hear from influential experts at the forefront of climate, nature, and AI.
        • Additional resources. Seed Grantees may receive access to additional resources, which may include computing infrastructure or access to relevant datasets.
      • Seed Grantees will submit their implementation plans and present them to the judging panel. Judges will evaluate submissions according to the Phase 2 selection criteria and recommend up to 15 awardees. The Bezos Earth Fund will select the final slate of Implementation Grantees. Phase 2 will culminate with the announcement of up to 15 Implementation Grantees; each team will each receive up to $2 million to close the gap between concept and viability.
    • Post-award implementation
      • Following Phase 2, Implementation Grantees will bring their solutions to life over the course of two years; during this time, they will test, iterate, and report on the effectiveness of their approaches. Awardees will also convene in person at the AI for Climate and Nature Summit in 2025.
      • The Bezos Earth Fund will announce more information on future rounds of the Grand Challenge at a later date. Sign up for the newsletter for announcements.
Eligibility Criteria
  • The first round of the Grand Challenge invites grant proposals from eligible organizations: U.S.-based 501(c)(3) entities and global academic institutions. In keeping with the Bezos Earth Fund’s commitment to equity and access, eligible applicants may collaborate with organizations all over the world to develop their proposals. Proposals from non-affiliated individuals are not eligible.
  • To establish and maintain eligibility for participating in the Challenge, each Lead Entity and Contributing Entity represents, warrants, agrees and covenants as follows:
    • It is a legal entity organized under the laws of a state, country or other governmental body that is not subject to U.S. sanctions as specified on the U.S. Department of Treasury website.
    • It satisfies all Challenge requirements specified on the website (the “Challenge Website”), the terms of which are incorporated herein by reference as though fully set forth, and this Agreement. 
    • All applications, whether submitted by a single entity or group of entities, must designate a single entity as the “Lead Entity,” that, if the grant application is selected, will be the Earth Fund’s grantee. The Lead Entity is responsible for fulfilling the obligations designated to such Lead Entity under this Agreement and the Challenge Website, including submitting the grant application, submitting this Agreement signed by each Contributing Entity, signing the grant agreement, receiving the grant award, and additional obligations identified from time to time by the Earth Fund in support of the administration and operation of the Challenge. For applications submitted by a single entity, that entity shall be the Lead Entity for that application.  A Lead Entity, whether applying alone or jointly, must be either (1) a domestic U.S. organization that is exempt from taxation under Section 501(c)(3) of the U.S. Internal Revenue Code, or (2) an academic institution from anywhere in the world. 
    • The Lead Entity must designate an individual point of contact who is affiliated with and authorized to act on behalf of the Lead Entity as the “Team Lead.” The Team Lead must be the older of: 18 years of age or the age of majority in the state or country where the individual resides. Team Leads will be responsible for all Challenge-related communications with the Administrators or Bezos Earth Fund. The Team Lead may not be a citizen of or reside in a country that is subject to U.S. sanctions as specified on the U.S. Department of Treasury website.  Any notices by the Earth Fund or either Administrator shall be deemed delivered once sent via email to the Team Lead using the email address provided in the grant application.
    • Each Contributing Entity agrees that the Lead Entity and Team Lead may undertake all tasks required to advance the grant application and in connection with this Agreement and all Challenge-related agreements and other obligations.  For the avoidance of doubt, any Contributing Entity acknowledges that the grants provided pursuant to this Challenge will be paid solely to entities that are academic institutions or U.S. entities that are exempt from taxation under Section 501(c)(3) of the U.S. Internal Revenue Code.   
    • No Grant Applicant may be owned, managed, or controlled by employees, officers, directors or board members of the Earth Fund or either Administrator. 
    • Phase 2 participation is restricted to Phase 1 awardees. To be eligible for Phase 2 grant awards, Phase 1 awardees must have complied with all requirements to participate in relevant activities for Phase 2 as described on the Challenge Website. 
    • Any agreement signed, or obligation undertaken regarding a Lead Entity or Contributing Entity’s participation in the Challenge that conflicts with this Agreement may make the Grant Applicant ineligible.

For more information, visit Bezos Earth Fund.

Leave a Reply