**** ## **Introduction: A New Blueprint for Climate Action** Climate change poses an unprecedented challenge that demands urgent, collective action on a global scale. Yet our predominant approaches – relying on market incentives and top-down policies – have proven insufficient in the face of this crisis ([Bowen, Dietz, Hicks, 2014](https://www.lse.ac.uk/granthaminstitute/explainers/why-do-economists-describe-climate-change-as-a-market-failure/#:~:text=The%20core%20one%20is%20the,effect%20of%20economically%20valuable%20activities)) ([Kanh and Munira, 2021](https://link.springer.com/article/10.1007/s10584-021-03195-w#:~:text=enhanced%20at%20scale%3F%20As%20a,in%20promoting%20the%20proposed%20framing)). Climate change has even been described as “the greatest market failure” ([Murphy, 2009](http://www.econlib.org/library/Columns/y2009/Murphyclimate.html)), where greenhouse gas emissions and climate risks are treated as externalities beyond the reach of normal market forces. The result is a glaring gap between what is scientifically necessary and what is being implemented on the ground. No single entity – whether government, corporation, or NGO – can tackle this alone. Indeed, **meaningful and lasting impact requires collective action** that brings together diverse perspectives, resources, and expertise ([Vijakumar, 2025](https://www.weforum.org/stories/2025/01/collective-action-is-key-to-climate-resilience/#:~:text=In%20an%20era%20marked%20by,diverse%20perspectives%2C%20resources%20and%20expertise)). This whitepaper presents _“Ten Million Projects by 2030”_ as a bold, distributed model for climate action, enabled by advances in artificial intelligence (AI) and grounded in local action. This vision calls for millions of community-scale mitigation and adaptation projects worldwide, all interconnected through an open socio-technical network (“**EarthNet**”). By 2030, these ten million initiatives aim to form a scalable mosaic of climate solutions – from clean energy microgrids in villages to urban greening campaigns and resilient housing upgrades – each tailored to local needs but enriched by shared global knowledge. In contrast to one-size-fits-all strategies, this approach is **polycentric**, operating at multiple levels with local, regional, and national stakeholders working in concert ([Ostrom, 2016](https://papers.ssrn.com/sol3/papers.cfm?abstract_id=1494833#:~:text=alternative%2C%20the%20paper%20proposes%20a,for%20assessing%20the%20benefits%20and)). Such an approach encourages experimentation, learning, and innovation across different contexts, building trust and momentum from the ground up. Crucially, this blueprint is not just a technological fix but a new **socio-technical model** for climate action. It integrates cutting-edge AI tools, knowledge graphs, and digital networks with the power of communities of practice, public investment, and grassroots participation. By leveraging AI in service of collective action, EarthNet can help overcome structural barriers and **structural traps** that have kept society locked into climate inaction – from short-term economic thinking to fragmented, siloed efforts ([Peter Søgaard Jørgensen, 2023](https://royalsocietypublishing.org/doi/10.1098/rstb.2022.0261)). In this paper, we detail this vision and its rationale: why a local distributed approach is needed; how EarthNet and the [Model Context Protocol (MCP)](https://modelcontextprotocol.io/introduction) enable coordination and rapid learning; and what it looks like in practice through an in-depth case study of a national climate-resilient housing initiative. We draw on academic research, policy analyses, and real-world precedents to show that _Ten Million Projects_ is both achievable and necessary. To meet the climate challenge we must go beyond the limitations of markets and isolated projects. We must foster a global tapestry of locally-led solutions, supercharged by AI and knowledge-sharing, to build resilience and cut emissions in every community. The “Ten Million Projects” initiative provides a blueprint for doing so, aligning policymakers, philanthropic funders, and AI researchers behind a common framework. The goal: a climate-resilient society by 2030, achieved through millions of collective actions that together form a greater whole. ## **The Limits of Market-Driven Approaches and the Need for Collective Action** #### **Market Mechanisms and Their Shortcomings** Climate change is often cited as a textbook example of market failure ([Bowen, Dietz, Hicks, 2014](https://www.lse.ac.uk/granthaminstitute/explainers/why-do-economists-describe-climate-change-as-a-market-failure/#:~:text=The%20core%20one%20is%20the,effect%20of%20economically%20valuable%20activities)). The costs of greenhouse gas emissions – in the form of extreme weather, sea-level rise, and ecosystem loss – are not borne by the emitters, leading to a systemic under-investment in mitigation. Moreover, many benefits of climate resilience (like a safer community or protected watershed) are public goods that markets struggle to value. As a result, **market-driven approaches alone have proven inadequate for catalyzing the needed scale of climate action**. Carbon pricing and emissions trading, while important, have been slow to spur deep decarbonization and virtually _cannot_ address adaptation, which often has no profitable revenue stream. _Adaptation in particular suffers from under-funding because it’s framed as a local or national public good, revealing the inefficacy of market mechanisms in this arena_ ([IPCC, 2018](https://www.ipcc.ch/site/assets/uploads/2018/02/WGIIAR5-Chap17_FINAL.pdf))_._ In short, when left to market forces, we collectively underinvest in long-term resilience and emissions reduction, as short-term profit motives dominate decision-making. #### **The Collective Action Imperative** Climate change is the **quintessential global collective action problem** – every nation and community’s efforts contribute to the whole, yet no single actor has sufficient incentive to act at the level needed ([Ostrom, 2016](https://papers.ssrn.com/sol3/papers.cfm?abstract_id=1494833)). Over the past decades, humanity has relied on slow global negotiations and voluntary corporate actions has led to insufficient progress. It has become clear that _truly transformative climate action requires coordinated effort across all levels of society_. This means governments at all levels aligning policies, businesses and investors actively supporting decarbonization, civil society mobilizing communities, and scientists and engineers providing technological innovation. **“Collective action is not just about partnerships; it’s about fostering a mindset that prioritizes a shared purpose over individual agendas,”** as one industry leader noted ([Vikajumar, 2025](https://www.weforum.org/stories/2025/01/collective-action-is-key-to-climate-resilience/#:~:text=Collective%20action%20is%20not%20just,building)). When diverse stakeholders pool their efforts, they can accelerate progress and achieve what none could alone – whether developing scalable solutions for the Sustainable Development Goals or addressing localized crises like water scarcity ([Vikajumar, 2025](https://www.weforum.org/stories/2025/01/collective-action-is-key-to-climate-resilience/#:~:text=Collective%20action%20is%20not%20just,building)). Policymakers and economists increasingly recognize that _addressing climate change demands going beyond the invisible hand of the market to the guiding hands of coordinated policy and community engagement_. For instance, the European Central Bank has argued that broad collective action – by governments, firms, and financial institutions – is required to correct the market’s failure to price climate risks ([Schnabel, 2020](https://www.ecb.europa.eu/press/key/date/2020/html/ecb.sp200928_1~268b0b672f.en.html)). Nobel laureate Elinor Ostrom likewise contended that a single global agreement would likely be “inherently weak” due to free-rider problems, and instead advocated a **polycentric approach** with many centers of initiative working in tandem ([Ostrom, 2016](https://papers.ssrn.com/sol3/papers.cfm?abstract_id=1494833)). Such an approach allows actors at different scales to innovate and reinforce each other’s efforts, building trust and reciprocity from the ground up. _In practice, this means empowering local governments, community groups, and networks of practitioners to take bold actions within their domain – while sharing knowledge and aligning with others through overarching frameworks_. #### **Structural Traps and Climate Inaction:** Beyond economics, there are deeper structural and cognitive barriers that trap societies into inaction. Research on “Anthropocene traps” identifies how _short-termism_ and other systemic dynamics impede sustainable behaviour ([Jehn, 2018](https://florianjehn.github.io/Societal_Collapse/2024-06-20-anthropocene_traps/#:~:text=The%20main%20temporal%20trap%20is,have%20easy%20access%20to%20it) and [Schippers et al., 2024](https://doi.org/10.3389/fsoc.2024.1194597)). One key **structural trap is the short-term focus** of our institutions and economies: political and business cycles prioritize immediate growth or quick fixes, **overshadowing long-term sustainability**. Investments in resilience or emissions reduction – which may not pay off for years or decades – are often shelved in favour of projects with instant returns. Similarly, a **connectivity trap** exists whereby those who make decisions (or benefit from high consumption) are insulated from the environmental consequences, which are displaced in time or space. For example, urban residents may not directly see the ecosystems being degraded to support city life, weakening their impulse to act. These traps are self-reinforcing and contribute to what has been called the inertia or _lock-in_ of high-carbon, vulnerable development pathways ([Jørgensen,2023](https://doi.org/10.1098/rstb.2022.0261)). Our energy and infrastructure systems have massive sunk costs (“infrastructure lock-in” ), and our behaviours and expectations have normalized unsustainable practices, making change difficult. Overcoming these structural traps requires **deliberate collective intervention**. Market signals alone won’t break the cycle of short-term thinking – we need policies and norms that extend decision-makers’ time horizons (for instance, climate risk disclosure, long-term planning mandates, or education that instills future-focused values). We also need platforms for shared learning to make the hidden connections between actions and impacts visible to all. In short, _structural traps call for structural solutions_: rethinking incentives, information flows, and governance models at a fundamental level. The _new socio-technical model_ we propose is designed to do precisely this. Engaging communities in tangible projects roots climate action in local self-interest (escaping the trap of distant consequences). Networking these projects globally through AI and knowledge-sharing helps expose the links between local actions and global effects, building a more informed and altruistic collective mindset. And by iterating rapidly and adaptively, it counters the inertia of existing systems with an agile, experimental culture oriented toward the long term. #### **Why Policy and Philanthropy Must Lead** Because of these market limitations and structural barriers, **public policy and philanthropic funding are pivotal** in kick-starting and scaling climate resilience efforts. Government policies can correct market failures – for example, by funding critical R&D, implementing regulations like building codes, or subsidizing early adoption of clean technology. Likewise, philanthropies can take risks on innovative approaches and reach underserved communities where private capital might not venture. _Both are essential for supporting collective action initiatives that have broad social payoffs but may not yield immediate financial returns._ The Ten Million Projects framework relies on such leadership: policymakers to create enabling environments for local projects (through grants, incentives, and removal of regulatory hurdles), and governments and philanthropists to fund the development of shared infrastructure like EarthNet and to seed projects in vulnerable or low-income regions. Together, they can help society escape the current traps by essentially **“flipping” the system** – valuing long-term community resilience over short-term profit, and collaboration over competition. In sum, the status quo approach to climate change – heavy on market-based tools and fragmented efforts – has not delivered the needed results. If we continue down that path, global temperatures will rise well beyond safe limits, and communities will remain unprepared for climate extremes. To course-correct, we must emphasize **collective action, public goods, and structural change**. This sets the stage for a radically different approach: building a distributed network of millions of local projects, supported by technology and new institutions, to drive climate adaptation and mitigation from the bottom up. In the next section, we outline this new model and how it addresses these shortcomings. ## **A Socio-Technical Model for Climate Resilience: Local Projects in a Global Network** #### **From Centralized Efforts to Polycentric Networks:** The enormity of climate change might tempt us to seek one grand solution – a single global treaty, a revolutionary technology, or a trillion-dollar investment plan. History suggests, however, that complex global problems are better met with **polycentric and multi-level action** ([Ostrom 2009](https://papers.ssrn.com/sol3/papers.cfm?abstract_id=1494833#:~:text=alternative%2C%20the%20paper%20proposes%20a,for%20assessing%20the%20benefits%20and)). In a polycentric system, numerous semi-independent initiatives take place in parallel, each at the scale where it is most effective, but with mechanisms to learn from and reinforce one another. This creates resilience through diversity: a failure in one approach can be compensated by successes elsewhere, and innovations can spread rapidly through the network. The _Ten Million Projects by 2030_ initiative is fundamentally polycentric. Rather than a few huge projects, it envisions countless local projects – _energy cooperatives, community farms, retrofit teams, ecosystem restoration brigades, youth climate clubs_, and more – each initiated by those with the most at stake and local knowledge. These initiatives would be distributed across neighbourhoods, cities, regions, and countries worldwide, reflecting the varied contexts of the climate challenge (from drought-prone rural areas to flood-vulnerable megacities). What binds these myriad efforts together is a **common framework and supporting infrastructure**. Imagine a global platform where each community or group that wants to contribute to climate action can register their project, access technical resources, find partners, and share results. EarthNet (detailed in the next section) can serve as this backbone, ensuring that _local efforts add up to more than the sum of their parts_. In essence, it enables a _“network of communities”_ – a modern, AI-enabled extension of what Elinor Ostrom described: _small- to medium-scale governance units linked through information networks and monitoring at all levels_ ([Ostrom 2009](https://papers.ssrn.com/sol3/papers.cfm?abstract_id=1494833#:~:text=alternative%2C%20the%20paper%20proposes%20a,for%20assessing%20the%20benefits%20and)). Such linkages are vital: they allow successful strategies in one locale (say, a novel agroforestry method or a finance model for retrofits) to be quickly discovered and adapted elsewhere. They also foster a sense of global solidarity and shared mission, which bolsters the political and social support for action. #### **Key Principles of the New Model** Several core principles guide this model: - **Local Ownership, Global Solidarity:** Each project is **locally led and context-specific**. Communities identify their own needs – whether it’s securing water supplies, cooling a heat-stressed city block, or shifting to solar energy – and design solutions that fit their cultural, economic, and environmental context. This local ownership is crucial for legitimacy and effectiveness. However, unlike isolated community projects of the past, these initiatives are networked globally. They benefit from **global solidarity** in the form of shared funding pools, knowledge exchanges, and technology support. This ensures equity: even poor or remote communities can leapfrog by accessing world-class tools and expertise through EarthNet, and their on-the-ground insights are valued as part of the collective intelligence. - **Bridging Mitigation and Adaptation:** Traditional climate policy often separates mitigation (emissions reduction) from adaptation (resilience-building). The _Ten Million Projects_ approach intentionally bridges the two. Many local projects will produce co-benefits: for example, urban tree-planting both sequesters carbon and reduces heatwaves; restoring a mangrove forest captures carbon and buffers storm surges. By **integrating mitigation and adaptation** at the project level, the initiative breaks out of the silos that sometimes plague climate funding. It recognizes that communities care about holistic outcomes – cleaner air, safer homes, stable jobs – and that climate strategies should deliver on those immediate needs while also contributing to global goals. The result is a portfolio of projects that collectively advance both adaptation and mitigation, increasing political and public buy-in. _(In the case study on housing, we will see how a single program can cut emissions and shield people from extreme weather.)_ - **Rapid Learning and Iteration:** An advantage of having millions of projects is the opportunity for **massive parallel experimentation**. Each project can be seen as a “trial” of what works in climate action, generating data and lessons. With proper knowledge management, the network as a whole can learn extremely fast. EarthNet’s design (with its knowledge graphs and communities of practice) is to make these learning feedback loops as tight as possible. If a coastal village in Bangladesh discovers an effective new way to raise homes above flood levels, that knowledge can be packaged (through how-to guides, AI-assisted design templates, etc.) and transferred to other flood-prone regions within days. The model context protocol (MCP) can help encapsulate the context of that solution – the type of housing, materials, cost, cultural factors – so that an AI assistant elsewhere can adapt it to a new locale. This is analogous to open-source software development: many independent contributors, continuous improvement, and rapid iteration cycles. In the realm of climate resilience, such agility is invaluable because conditions are changing quickly and yesterday’s assumptions may not hold tomorrow. By design, _Ten Million Projects_ seeks to **accelerate innovation diffusion** and avoid the slow, linear scaling of pilot projects that we often see. - **Structural Change through Collective Action:** While each project tackles immediate issues, together they aim to **shift structural conditions** that underlie climate inaction. As more projects bloom, they start to create new markets and norms – for example, widespread community demand for green building materials could stimulate industries to supply low-carbon products at scale, overcoming current cost barriers. Millions of people engaged in local climate solutions also builds social capital and a shared sense of purpose that can influence politics. In other words, this bottom-up movement can push top-down change: local success stories make it easier for national governments to enact ambitious policies (since they can see proof of concept and public support), and large institutions are held accountable by an informed citizenry that is actively involved in solutions. We move from a vicious cycle (where systemic inertia thwarts local action) to a **virtuous cycle** where local actions cumulatively erode the very structural traps we discussed earlier – short-termism is challenged by communities planning for their children’s future, and fragmentation is countered by networked collaboration. To illustrate how this model contrasts with the status quo, consider **Table 1** below, which summarizes the shift: | **Aspect** | **Conventional Approach** | **Ten Million Projects Model** | | ------------------------ | ------------------------------------------ | -------------------------------------------------------------------- | | Primary Driver | Market incentives, top-down targets | Collective action, community initiatives | | Scale of Action | Centralized, large projects or policies | Distributed, millions of small-to-medium projects | | Knowledge Flow | One-way (experts to locals), siloed | Two-way & networked (global knowledge commons) | | Adaptation vs Mitigation | Often addressed separately | Integrated in local projects (co-benefits emphasized) | | Innovation | Slow pilot programs, proprietary solutions | Rapid experimentation, open-source solutions | | Equity & Inclusion | Mixed – many communities left out | Built-in via self-organization, local ownership and shared resources | | Examples | Carbon markets, single mega wind farm | Thousands of community solar co-ops, climate resilient gardens | _Table 1: Comparison of Conventional Climate Action Approach vs. the “Ten Million Projects” Distributed Model._ As shown, the distributed model aims to be more inclusive, faster to learn, and structurally transformative. It complements, rather than replaces, national and international policies and initiatives – global carbon pricing or emissions standards remain important, but _Ten Million Projects_ supplies the grassroots foundation that makes those policies effective and just. It also unlocks the power of **AI and digital technology** in a way that respects local contexts, as we explore next. ## **EarthNet: A Distributed AI Network Empowering Communities** A cornerstone of this blueprint is **EarthNet**, an AI-enabled platform and network that connects and supports the ten million projects. EarthNet can be thought of as the “digital nervous system” of this distributed climate action movement, linking brains (AI models and human experts), muscles (project implementers on the ground), and memory (data and knowledge repositories) into one cohesive whole. Its design principles are openness, interoperability, and augmentation – _augmenting human intelligence and collaboration, rather than replacing or centralizing it_. Here we detail key components of EarthNet and how they function: #### **1. Model Context Protocol (MCP) for Distributed AI:** At the technical heart of EarthNet is the **Model Context Protocol (MCP)**, a recently introduced open standard that allows AI models to seamlessly integrate with external data sources, tools, and each other ([Anthropic, 2024](https://modelcontextprotocol.io/introduction)). Traditionally, deploying AI for a specific community project (say, optimizing a town’s microgrid or simulating local flood scenarios) required significant custom coding and data wrangling. The MCP simplifies this by providing a standardized way for different services and databases to connect to AI systems. In EarthNet, every community or project can run an **MCP-compatible client** – for example, a local climate data hub, a sensor network, or a document repository of traditional ecological knowledge – which exposes relevant information in a common format. AI tools (the servers in MCP terms) can then plug into these clients without bespoke integration code ([Anthropic, 2024](https://modelcontextprotocol.io/introduction)). What does this achieve? It means an _AI assistant can access local context in real-time_ – climate data, socioeconomic data, project records – and combine it with global knowledge to provide tailored insights. For instance, an EarthNet AI agent helping a coastal mangrove restoration project could automatically pull the project’s latest drone imagery and sensor data via MCP, analyze it against global climate models and best-practice databases, and suggest adaptive strategies if it detects early signs of mangrove stress. All this without the community needing an in-house AI team. The **client-server architecture** of MCP also allows multiple AI services to work together : one might handle data visualization, another performs simulation, another moderates community discussions, all sharing context. Currently, MCP supports both local and remote resources – meaning communities can keep sensitive data locally (enhancing trust and privacy) while still benefiting from remote AI expertise when needed. Anthropic and others are actively developing secure, authenticated ways to use MCP in organizational settings , which EarthNet will leverage to ensure data is shared only as intended by communities. In essence, MCP turns EarthNet into a **federated AI ecosystem**. Each node (community) can contribute data and pick which AI tools to engage. The protocol ensures a common “language” so that an improvement made for one community (like a new analysis plugin) can be easily applied by others. This openness avoids vendor lock-in and encourages an **ecosystem of AI research partners** to build tools that plug into EarthNet – from predictive models of crop yields to natural language processors that translate technical guidance into local languages. Over time, as more projects come online, EarthNet’s AI capabilities grow richer, enabling more sophisticated coordination and insight generation across the network. #### **2. Communities of Practice and Social Networking:** Technology alone does not create change; people do. Thus, EarthNet is not just an array of data feeds and AI models – it’s also a **social platform for communities of practice (CoP)** focused on climate solutions. Participants in the ten million projects (e.g. local project leads, volunteers, domain experts, municipal officials) join thematic or regional groups where they can exchange experiences, ask for advice, and collaboratively solve problems. These groups might be organized by sector (“Coastal Resilience Practitioners Network”, “Forestry Adaptation Network”), by geography (“Southern Africa Drought Adaptation Forum”), or by function (“Youth Climate Corps”, “Climate Caucus”). The idea draws from successful knowledge-sharing networks in the resilience field. For example, the Resilient Cities Network and similar initiatives have **communities of practice for urban practitioners to exchange real-time knowledge** on challenges like heatwaves or flooding ([Resilient Cities Network, 2023](https://resilientcitiesnetwork.org/communities-of-practice/) ). EarthNet scales this up dramatically: every one of the ten million projects can find a home in one or several communities of peers tackling analogous issues. Through EarthNet’s interface (imagine a hybrid of a professional social network and a wiki), a user might ask the AI assistant: “Has anyone implemented rainwater harvesting in low-income neighbourhoods? Looking for design templates and community engagement tips.” Within seconds, the AI can provide relevant people or organizations and create notifications via defined expert groups. Within hours, messages could come from around the world – perhaps an engineer in Brazil shares a blueprint they used, a NGO worker in India shares a guide on community workshops, and the AI assistant, continuing to monitor the discussion via MCP integration, chips in with a summary of best practices from the literature. These social interactions can be **captured and curated** to further enhance the knowledge base. EarthNet will employ _knowledge graph_ technology (discussed next) to organize insights from discussions, documents, and data so that they are easily discoverable. The **value of community-of-practice networks** is that they build trust and tacit knowledge exchange beyond what formal reports capture. They also help avoid duplication of effort – a key benefit when operating at the scale of millions of projects. If ten communities have tried something and found pitfalls, the eleventh can be forewarned even before they start. Moreover, these communities create a support system; undertaking a climate project can be daunting, but knowing that thousands of others are simultaneously engaged in similar work provides motivation and psychological safety to persist. In policy terms, this is building a _global social infrastructure for climate action_, where anyone from a mayor to a farmer can plug in and find solidarity and guidance. #### **3. Climate Action Knowledge Network: Knowledge Graph and Vector Database:** One of EarthNet’s most powerful features is its **knowledge management system**, built around a _global Knowledge Graph_ and vector database. A knowledge graph stores information that highlights relationships between knowledge concepts (like a giant network of connected facts and experiences). In the context of Ten Million Projects, the knowledge graph would encode relationships among _projects, people, places, problems, and solutions_. For example, it might link a particular project (say a solar microgrid in Kenya) to the technologies it used (solar panels, battery storage), to the outcomes achieved (reduced diesel generator use by X%, improved school electrification), to related projects elsewhere, and to key documents or tools. A vector database stores information representing knowledge entities as _vector embeddings_ (high-dimensional numerical representations of meaning), enabling _semantic search_ based on meaning rather than keyword similarity. For example, searching for "local renewable power projects" would return small-scale solar, wind, or geothermal projects regardless of any specific keyword. Together, these technologies create a rich, queryable web of knowledge. According to the Stockholm Environment Institute, **a global climate action knowledge network can break down barriers to the spread of needed knowledge and action** ([Kaltenböck et al. ](https://www.preventionweb.net/publication/creating-global-climate-action-knowledge-graph-leverage-artificial-intelligence-and#:~:text=This%20paper%20proposes%20and%20argues,this%20Knowledge%20Graph%20would%20enable)). It would enable rapid discovery of insights, efficient exploration of connections across domains, and better linking of science, policy, and practice. In EarthNet, the knowledge network allows an AI (or a user) to ask complex questions like “find all projects in tropical coastal regions that dealt with mangrove restoration and had budget under $50k” or “what building retrofit solutions improved indoor air quality during wildfires in communities with median income below national average?” The graph would quickly return relevant projects, contacts, and courses, whereas such cross-cutting queries would be nearly impossible with siloed information. **How is the knowledge network built?** Partly through manual input (project organizers adding their projects with specific connections, or curators linking related initiatives) and partly through automated extraction. As AI language models parse reports, discussion transcripts, and data sets flowing through EarthNet, they can populate the graph and vector databases with new nodes and links. For example, suppose multiple communities discuss a novel concept (like a “fog net” for water harvesting). In that case, the AI can identify that concept and map its emergence – linking it to projects that implemented fog nets, the people advocating it, and the measured results. The knowledge graph thus grows organically with every project and every piece of shared knowledge, becoming smarter over time. One can think of it as **the collective memory and brain of the climate resilience movement**. It will help avoid the common problem where valuable lessons learned in one project are lost when that project ends or the team moves on. Instead, those lessons remain in the graph to inform future actions. Additionally, because the graph links science and practice, it helps bridge the gap between research and implementation. For instance, academic findings (like optimal crop varieties for drought) can be linked to field projects validating or challenging those findings, creating a feedback loop that can guide further research. The importance of such an integrated knowledge system cannot be overstated – as the Stockholm Environment Institute noted, it sets the stage to leverage powerful AI applications by providing a solid, structured foundation of climate knowledge ([Kaltenböck et al. ](https://www.preventionweb.net/publication/creating-global-climate-action-knowledge-graph-leverage-artificial-intelligence-and#:~:text=This%20paper%20proposes%20and%20argues,this%20Knowledge%20Graph%20would%20enable)). In EarthNet’s case, AIs can navigate the knowledge graph and employ the vector database to provide evidence-based advice and identify patterns (say, common factors in successful projects) at a scale no human alone could. #### **4. Coordination and Rapid Iteration Enabled:** With MCP tying data and models together, communities of practice tying people together, and the knowledge graph and vector db tying information together, EarthNet creates an unprecedented capacity for **coordination and rapid iteration** in climate action. Coordination here does not mean top-down control; it refers to alignment and synergy. For example, EarthNet can highlight opportunities for **aggregated efforts** – if dozens of towns in different countries all express interest in building small solar farms, EarthNet could facilitate a joint procurement or a shared technical training program, lowering costs for all. Similarly, EarthNet might coordinate **phased project rollouts**: pilot something in a few diverse communities, evaluate outcomes, refine the approach, and then help replicate it widely through the network. This mirrors agile development in software – iterate in sprints and scale what works – but applied to social and infrastructural projects. The backing of AI makes this feasible, as progress can be monitored in near real-time (e.g. via remote sensing or IoT data feeding into EarthNet) and adjustments suggested. Another aspect is **knowledge transfer at speed**. Traditionally, transferring a successful climate adaptation intervention from one country to another can take years – relying on publication, conferences, and new funding cycles. With EarthNet, if a breakthrough occurs (say a new drought-resistant crop or an innovative financing scheme for rooftop solar), it can be broadcast through the network and literally millions of projects can consider adopting it the next day. The communities of practice help filter and localize these ideas (peers discuss how to adapt the idea to local language or culture), while AI assistants can instantly translate guidelines and even simulate expected results in the new context. This can compress the timeline of scaling solutions from decades to months or less. Finally, EarthNet’s integrated design helps measure collective impact. By aggregating data from projects, it can estimate the total emissions avoided, people protected, or other metrics – crucial for policymakers and funders to see progress. It can also identify gaps where few projects exist (maybe some region or sector is underrepresented) and mobilize resources or campaigns to fill those gaps, ensuring a comprehensive approach. **In summary, EarthNet turns the concept of “learning by doing” into “learning by doing, multiplied by millions, and shared instantly.”** This is a new socio-technical paradigm: one where AI and network technology amplify human collaboration and ingenuity to tackle the climate crisis. _(Having outlined the vision and infrastructure, we now ground these ideas in a concrete case study: a national program for climate-resilient housing, to show how local projects and EarthNet can deliver tangible results.)_ ## **Case Study: A National Climate-Resilient Housing Initiative** **Context and Rationale:** Housing sits at the nexus of climate mitigation, adaptation, and social equity. Buildings account for a significant share of greenhouse gas emissions (through both construction materials and the energy used for heating, cooling, and lighting). At the same time, homes are where people experience climate impacts like heat waves, cold snaps, and wildfire smoke. Many countries face housing shortages or affordability crises – for instance, Canada is estimated to be **short 3.5 million homes** relative to need ([CMHC, 2023](https://www.cmhc-schl.gc.ca/professionals/housing-markets-data-and-research/housing-research/research-reports/accelerate-supply/housing-shortages-canada-updating-how-much-we-need-by-2030#:~:text=Our%20most%20recent%20projection%20is,in%20Ontario%20and%20British%20Columbia.)) – and much of the existing housing stock is aging, inefficient, and ill-suited for a changing climate. This case study explores a hypothetical but realistic national program that aligns with the _Ten Million Projects_ ethos: a **massive home-building and retrofitting campaign** that expands affordable housing, slashes emissions, and enhances climate resilience for residents. We will consider how such a program could be designed and implemented by leveraging EarthNet’s AI-enabled networks and the collective power of communities. **Program Overview:** Let’s imagine Canada launches the **“Resilient Homes for All” initiative in 2025**. The goals by 2030 are: (a) **Build** 3.5 million new affordable, net-zero-emission housing units nationwide, focusing on areas with severe housing shortages; (b) **Retrofit** 5 million existing homes and apartments for energy efficiency, electrification (no fossil fuel heating), and climate resilience upgrades (such as better insulation, ventilation, and cooling); (c) **Reduce** operational emissions of housing stock by 50% and embodied carbon in new construction by significant amounts; (d) **Improve** indoor environmental quality for residents, protecting them against extreme heat and wildfire smoke events; and (e) **Generate** hundreds of thousands of jobs and catalyze innovation in the construction sector. This is an ambitious undertaking, akin to a “Housing Marshall Plan,” and clearly a task beyond the scope of markets alone – it would need substantial public investment (potentially funded by a green recovery act or climate infrastructure bill) and coordination across federal, provincial, and local governments. It is precisely the kind of big push that policymakers and philanthropic funders might support, and it provides fertile ground for the _Ten Million Projects_ approach: the program would consist of **thousands of local projects** (community housing developments, city retrofit programs, non-profit initiatives) all under the national umbrella. **Mitigation and Health Benefits:** By expanding supply of **abundant, efficient housing**, the program addresses multiple crises. Each net-zero new home built means one less household relying on high-emission, high-cost energy. Each home retrofit – adding insulation, sealing leaks, installing heat pumps and efficient appliances – can dramatically cut energy use and bills for residents. For example, the innovative Dutch program, _Energiesprong_, has shown that deep retrofits can **achieve 70–80% reduction in a building’s energy use** and can be completed in as little as 7–10 days using prefab methods ([Wikipedia](https://en.wikipedia.org/wiki/Energiesprong)). Their success (over 10,000 retrofits across Europe and counting ) indicates that scaling up retrofits is feasible with the right approach. Our case study program would replicate such approaches nationally, with support from EarthNet to transfer know-how. The carbon mitigation potential is huge: millions of retrofitted homes, combined with clean electricity, means a steep drop in residential CO₂ emissions (which in Country X might account for ~20% of national emissions). In fact, a recent study in _Nature Communications_ found that adopting AI and smart technologies in buildings could reduce energy consumption by up to ~19% and carbon emissions by as much as 90% (with clean power) compared to business-as-usual by mid-century ([Chao Ding, 2024](https://www.nature.com/articles/s41467-024-50088-4#auth-Chao-Ding-Aff1)). Our program’s integration with AI (through EarthNet) would help realize these gains sooner by optimizing retrofits and operations (for instance, AI systems learning the most efficient way to heat and cool a home based on occupancy patterns and weather forecasts ([Nicholson, 2025](https://www.jll.com/en-us/insights/how-ai-is-boosting-efforts-to-cut-buildings-energy-use#:~:text=The%20value%20of%20AI%20lies,energy%20distribution%2C%E2%80%9D%20says%20Ramya))). Beyond mitigation, the **health and resilience co-benefits** are a major driving force. As climate impacts worsen, homes must become sanctuaries against hazards. **Extreme heat** is a growing killer; a well-insulated, shaded, and ventilated home can maintain safer indoor temperatures during heatwaves, especially if equipped with efficient cooling like heat-pump air conditioners. **Wildfire smoke** has turned air quality hazardous in many regions; here, housing improvements make a decisive difference. A tight building envelope (achieved by weatherization) and high-quality filtration can reduce indoor particulate pollution dramatically, protecting residents’ lungs. Research by NIST and ASHRAE confirms that **weatherization measures such as air sealing, combined with high-efficiency filters, can limit smoke particulate infiltration into homes** ([NIST](https://tsapps.nist.gov/publication/get_pdf.cfm?pub_id=934224#:~:text=infiltration%20%28Rajagopalan%20and%20Goodman%2C%202021%29,systems%20will%20help%20reduce%20the)). In low-income communities, these upgrades have the double benefit of reducing energy insecurity and exposure risk. The HUD Green and Resilient Retrofit Program in the U.S. is already showcasing such benefits: it funds affordable housing retrofits that _“make homes more resilient to climate risks and improve indoor air quality,”_ while also cutting emissions and energy costs ([HUD](https://archives.hud.gov/news/2024/pr24-190.cfm#:~:text=emissions%2C%20promote%20the%20use%20of,by%20improving%20indoor%20air%20quality)). For instance, grants under this program help properties add better ventilation systems, insulation, and renewable energy, specifically noting that these changes **“will reduce costs and increase the quality of life… by making homes more resilient to extreme weather events and by improving indoor air quality.”** ([NIST](https://tsapps.nist.gov/publication/get_pdf.cfm?pub_id=934224#:~:text=tied%20to%20enhanced%20energy%20performance,during%20the%20hot%20summer%20months)). In our Resilient Homes for All case, every retrofitted or new home would incorporate a set of **resilience standards**: e.g., ability to maintain safe temperatures during a multi-day power outage (through passive cooling or backup power), filtration to keep indoor PM2.5 below health thresholds during smoke events, and structural reinforcements if in storm or seismic zones. These features essentially climate-proof the dwelling. The improvement in public health could be quantified in reduced heat-related illnesses and respiratory issues during wildfire season. And as the COVID-19 pandemic taught us, better indoor air quality (ventilation, filtration) has broad benefits, potentially reducing transmission of airborne illnesses as well – truly making housing a foundation for community well-being. **Implementation via Ten Million Projects Framework:** Rolling out such a massive program can be daunting, but EarthNet and the distributed approach would make it manageable and adaptive. Here’s how it might work in practice: - **Local Project Nodes:** The national government sets targets and funding mechanisms, but actual projects are proposed and led locally. A city might propose “Retrofit 20,000 homes by 2030” as its project, a rural cooperative might propose building 50 new efficient homes for their members, a non-profit might focus on weatherizing senior citizens’ houses, etc. By having many implementers, we avoid bottlenecks. EarthNet registers each of these projects as part of the Resilient Homes initiative, enabling tracking and support. - **Knowledge Sharing and Training:** Before deployment, EarthNet’s communities of practice would ramp up training and knowledge exchange. Builders, architects, and planners join communities like “Green Building and Retrofits Network” and “Healthy Homes Network” to learn state-of-the-art techniques. Through EarthNet, they access design templates and case studies: perhaps the latest passive house designs, or how a community in British Columbia developed an effective smoke shelter retrofit for homes. They also get guidance on using AI tools – for example, an AI-driven retrofit planner that, given a house’s characteristics, suggests an optimal package of upgrades and predicts energy savings and improved indoor air metrics. **Technical standards and toolkits** developed at the national level (like model building codes, procurement guides for low-carbon materials) are disseminated interactively via EarthNet. This ensures consistency in quality while allowing customization. - **Supply Chain and Innovation Coordination:** A program of this scale could strain supply chains (for heat pumps, insulation, skilled labor, etc.) if not managed. Here EarthNet’s coordination features shine. The knowledge graph will reflect real-time data on project progress and resource use. If a shortage of a certain insulation material is flagged in multiple projects, the central coordinators (government or industry partners) get an early warning to ramp up production or find alternatives. EarthNet’s AI could even forecast demand for key components across all projects and help aggregate bulk orders to negotiate better prices, which small contractors can then tap into. On the innovation side, as different projects try solutions (maybe different brands of air filters or various wall panel systems), performance data flows back into EarthNet. Suppose one type of heat pump consistently performs better in humid climates based on IoT data from homes – EarthNet will surface that insight and projects can adjust specs accordingly. This **continuous improvement loop** means the program avoids lock-in to subpar technology and can adapt over its duration with evidence-based tweaks. - **Community Engagement and Equity:** Because this is not just top-down construction but a community-driven effort, local groups engage residents through EarthNet as well. They might use the platform to organize town hall meetings (virtual or hybrid), share infographics on the benefits of retrofits, and even allow residents to vote on design options (for instance, what kind of shade trees to plant in their neighbourhood as part of cooling efforts). Through a dedicated app, a household might track the energy savings from their retrofit and compare with neighbours in a friendly competition, fostering a sense of ownership. EarthNet’s social features ensure that the voices of residents (especially marginalized groups) are heard – via feedback forms, local forums, and integration with civic grievance systems if, say, contractors aren’t performing. This **bottom-up feedback** is crucial to ensure the program delivers not just numbers, but real improvements in people’s lives (no “fake retrofits” or shoddy work, as sometimes happened in past weatherization drives). - **Philanthropic and Private Sector Support:** While government funds the core, philanthropic funders could use EarthNet to identify where additional support is needed. For example, the knowledge graph might reveal that certain low-income or remote areas have fewer project proposals (maybe due to lack of grant-writing capacity or technical know-how). Foundations could step in to fund capacity-building programs in those areas or provide micro-grants for community orgs to start housing initiatives. Likewise, impact investors could use EarthNet’s data to invest in scaling up factories for prefab retrofit panels or startups that innovate in smart home technologies, knowing that a large market (all these projects) is guaranteed. The transparency of progress on EarthNet builds confidence for all stakeholders; it is visible how funds translate into outcomes. - **Measuring Outcomes:** By 2030, the initiative’s success can be measured in multiple dimensions: number of homes built/upgraded, reduction in carbon emissions, reduction in residents’ energy bills, health metrics (fewer heat-related ER visits in areas with high retrofit rates, for example), and community satisfaction. We expect to see that those neighbourhoods with many resilient homes fare significantly better during climate extremes. For instance, during the summer of 2028, a record heatwave and wildfire smoke event hits British Columbia. Data collected through EarthNet shows that in neighbourhoods heavily retrofitted, indoor temperatures stayed 5–10°C cooler without emergency AC (thanks to insulation and cool roofs), and indoor PM2.5 levels were **50-80% lower than outdoors** with the provided filters – a dramatic contrast to non-retrofitted areas, where many vulnerable people had to seek public cooling and clean air shelters. This quantifies the **life-saving impact** of the program. Indeed, studies have found that having air filtration and tight building envelopes _“not only makes buildings more resilient to wildfire smoke episodes but also reduces everyday pollution exposure”_ ([Bernstein, 2024](https://www.eli.org/research-report/wildfire-smoke-state-policies-reducing-indoor-exposure#:~:text=Wildfire%20Smoke%3A%20State%20Policies%20for,episodes%20but%20also%20reduces)). Finally, EarthNet facilitates the **storytelling and replication** of this success. The experiences from Canada’s housing program – the policies that enabled it, the workforce development, the financing models (perhaps some homes were financed via green mortgages or on-bill financing for retrofits) – all are documented and shared internationally. Other countries launching similar efforts (and many will, as housing and climate are universal issues) can plug into the same knowledge network. In effect, the national initiative itself becomes one of the “Ten Million Projects,” albeit a very large aggregation of them, contributing to the global pool of know-how. **Summary of Impacts:** In conclusion, this case study demonstrates how a _Ten Million Projects_ approach could tackle a concrete challenge. Through a multitude of local actions (each house built or retrofitted is a mini-project), guided and scaled by AI-driven coordination (EarthNet), we achieve national-scale outcomes: major emissions reductions, climate-resilient communities, and improved living standards. It exemplifies the power of integrating mitigation (cutting energy use and emissions) with adaptation (guarding against heat and smoke) and doing so in a way that engages people (creating jobs, involving residents in decisions, focusing on affordable housing). The strategy aligns economic recovery with climate action – e.g., jobs in construction and manufacturing of green materials – which increases its political palatability. Crucially, the initiative also highlights how structural traps can be overcome. The **short-term economic mindset** is countered by framing retrofits and construction as investments that pay back via energy savings, health benefits, and avoided climate damages – and EarthNet helps by aggregating data to prove those paybacks to budget officials. The **fragmentation trap** is overcome by the shared platform: instead of thousands of disjointed efforts, they consciously learn from each other and march in the same direction, creating a unified market for solutions (driving costs down). The **market failure** is addressed by public/philanthropic funding and coordination, but leveraged by market participants (companies innovating in response to the clear long-term demand signalled). And the collective action problem is solved in practice: every community contributes what it can (with support relative to need), and everyone benefits from the collective reduction in climate risk and improved quality of life. The case of housing is just one example. Similar logic can be applied to other domains: regenerative agriculture networks to fix carbon in soils and boost food security, community-led reforestation and fire management programs, local micro-mobility and transit solutions to reduce transport emissions, and so on. In each, EarthNet and the Ten Million Projects framework would operate in analogous ways, adapting to sector specifics. The key takeaway is that **AI-enabled, community-driven networks can implement large-scale climate solutions both faster and more equitably than traditional methods**. We next provide recommendations on how to turn this vision into reality. ## **Recommendations and Pathways to Scale** Achieving the _Ten Million Projects by 2030_ vision will require concerted effort across policy, funding, and technology domains. Below we outline practical recommendations for key stakeholders – policymakers, philanthropic organizations, and AI research partners – to initiate and nurture this socio-technical transformation. We also discuss near-term steps and milestones to measure progress. #### **1. Policy and Government Actions:** Governments at all levels should embrace and facilitate polycentric climate action as a complement to top-down policies. Concretely: - **Launch National “Million Projects” Programs:** Similar to our case study, national governments can identify priority areas (e.g. resilient housing, community energy, nature-based solutions) and create funding streams that empower local entities to start projects in those areas. This includes block grants or challenge funds that communities, cities, and NGOs can draw on with minimal bureaucracy, provided they commit to sharing data and lessons via EarthNet. Setting a numeric target (like funding 10,000 local climate projects in the next two years) can galvanize action and signal seriousness. - **Establish an EarthNet Task Force:** A public-private task force (including tech companies, research labs, community representatives, and agencies like an environment or climate ministry) should be convened to develop the EarthNet platform as a public good. This task force would oversee issues of governance (to ensure the network remains open and benefits the many), data standards, and integration with existing public data portals. Governments should consider funding the initial development of EarthNet as critical infrastructure – akin to funding ARPANET in the early internet days. - **Incentivize Knowledge Sharing:** Build requirements or incentives for any publicly funded climate project to participate in the knowledge network. For example, a city receiving federal money for green infrastructure could be required to document its project on EarthNet (to a reasonable degree) and to use open data standards. Conversely, exemplary sharing could be rewarded (e.g., an annual “Climate Action Learning Award” for communities that contributed valuable knowledge to others). This will help populate EarthNet and normalize the culture of open collaboration. - **Address Structural Barriers:** Policymakers should also work on removing barriers identified as structural traps. Extend planning horizons by incorporating climate risk and future value of resilience into cost-benefit analyses for public investments (so short-term cost fears don’t block good projects). Mandate or encourage corporate disclosure of climate risks and actions, aligning private sector behaviour with long-term community goals. Consider new institutional models (like city-region climate compacts or cooperative utilities) that empower local climate action beyond traditional jurisdictions. - **Global Cooperation and Funding:** At the international level, forums like the UNFCCC should recognize and promote this approach. Developed countries and multilateral funds could allocate a portion of climate finance directly to support grassroots projects and the digital networks connecting them. Adaptation finance, notoriously lagging ([Kahn, M. 2021](https://link.springer.com/article/10.1007/s10584-021-03195-w#:~:text=responsibility,2%29%20that%20it%20makes)), could flow more readily if seen as building global public goods through local efforts. The concept of adaptation as a _global public good_ lends itself to justifying investments in EarthNet and community projects internationally, since knowledge and benefits spill across borders. Governments might also share EarthNet’s development so that it becomes a collaborative international open-source project (imagine a scenario where engineers from multiple countries co-develop the platform, ensuring it’s multilingual and globally accessible). #### **2. Philanthropic and Civil Society Actions:** Philanthropies and NGOs are often the risk-takers and innovators in social change. They should: - **Fund Pilot Networks:** Provide grant support to pilot the EarthNet approach in a few thematic areas immediately. For instance, a foundation could fund the creation of a “Climate-Smart Agriculture Network” on EarthNet, linking hundreds of farming projects across different continents, and demonstrate the impact on yields and carbon sequestration from knowledge exchange. These pilots will serve as proofs of concept to attract larger support. - **Build Capacity in Underserved Communities:** Ensure that marginalized and climate-vulnerable communities are not left out due to digital divides or lack of resources. Philanthropies can fund local facilitators or “digital extension agents” who help communities come online to EarthNet, articulate project ideas, and apply for funding. This is analogous to agricultural extension services, but for climate projects and using modern tech. By doing so, philanthropy guards against the network being dominated by well-resourced actors and truly democratizes participation. - **Support Open Tools and Content:** Invest in the development of open-source tools for climate action that can plug into EarthNet. This might mean supporting an NGO or university to create an open tool for community energy planning, or a library of climate education materials that can be shared. Also, fund the translation and localization of content so that knowledge is accessible in local languages and culturally appropriate forms (e.g., visual guides, radio snippets, etc., not just technical papers). - **Leverage Convening Power:** Use the influence of philanthropy to bring together unusual partners – e.g., indigenous knowledge holders with AI scientists, youth activists with urban planners – under the umbrella of this initiative. Such convenings (virtual or physical) could accelerate the fusion of ideas that EarthNet can then propagate. Civil society networks like C40 Cities, Climate Caucus, ICLEI, or grassroots movements can align their efforts with EarthNet to avoid duplication and strengthen the overall ecosystem. - **Accountability and Transparency:** NGOs can also play a watchdog role, using EarthNet’s data to hold actors accountable. For example, if certain promised projects are not delivering or benefits are not reaching disadvantaged groups, the platform's transparency allows civil society to highlight these issues and advocate for course corrections. This fosters trust in the system as one that is collaborative but also accountable. #### **3. AI Research and Tech Community Actions:** The AI and tech sector is pivotal in building and maintaining the digital infrastructure: - **Contribute to EarthNet Development:** AI researchers should actively contribute to the design of EarthNet’s AI components. This could be through participating in open-source projects related to MCP, knowledge graphs, or developing specialized climate-action models. The tech industry can volunteer or sponsor talent for coding sprints to develop features like robust simulation tools or multi-lingual support for the platform. - **Advance the MCP and Interoperability:** Support the maturation of the Model Context Protocol by adopting it in AI products and ensuring it meets the needs of climate and sustainable development applications. This might involve creating new adapters for common data sources in climate work (e.g., satellite imagery databases, weather data APIS) so they can be readily used in EarthNet workflows. Encourage major AI model providers (like those developing large language models or geospatial AI) to make their systems MCP-compatible, thus easily accessible to EarthNet users. - **Focus on Explainability and Trust:** Develop AI tools that are **explainable, transparent, and community-friendly**. For instance, when an AI suggests a particular adaptation measure to a community, it should be able to explain the rationale in plain language, citing data – much like this whitepaper cites sources – to build trust. Research in _explainable AI for climate decision support_ can improve how recommendations are presented on EarthNet. Additionally, implement feedback loops where community users can rate or flag AI advice, helping to improve model performance and alignment with human values over time. - **Scenario Modelling and Systemic Insight: One area where AI can particularly aid policymakers via EarthNet is in integrated scenario modelling**. By aggregating data from millions of projects, AI can help simulate “what if” scenarios at macro scales: e.g., “What if we reach 10 million projects – what does that mean for global temperature by 2050 compared to current policies?” or “Where are we likely to fall short, and what kind of projects are underrepresented that could plug the gap?”. Developing these analytical capabilities will demonstrate the added value of the network and guide strategic decisions. AI can highlight synergies (like how improving housing efficiency reduces load on the grid, enabling more EV charging without new power plants) or unintended consequences early. - **Ethical Guardrails:** The tech community should also anticipate and mitigate risks – such as unequal access to AI, biases in data, or potential misuse of the platform (e.g., spreading misinformation). Incorporating ethical guidelines and community governance in EarthNet’s AI (perhaps via a diverse oversight board) will be important. This aligns with the broader AI for good movement, ensuring technology is aligned with human and ecological well-being. #### **4. Near-Term Milestones (2025–2027):** To keep momentum, some tangible milestones within the next few years could include: - By **2025**: Launch of CanAdapt, (the first implementation of the EarthNet platform) with at least 500 initial projects across Canada. Establishment of 5 key communities of practice (e.g., forestry, mining, energy, resilient housing, coastal adaptation). Release of an **EarthNet Whitepaper (technical)** detailing architecture. - By **2026**: Demonstration of a “rapid knowledge transfer” event – e.g., a heatwave hits a region and within weeks 50 new cooling center projects spin up through EarthNet in other cities as a result of lessons learned. Show a measurable policy impact – perhaps a government cites CanAdapt data in setting a new climate target or funding allocation. Achieve integration with major platforms (like data from Canada's ClimateData.ca or the World Bank’s Climate Knowledge Portal flowing into CanAdapt, or interoperability with national adaptation plan databases). - By **2027**: Scale to thousands of projects. Independent evaluation shows that communities using EarthNet have, on average, faster project implementation times or better outcomes than those that don’t (validating the approach). Secure long-term funding for EarthNet through a coalition of governments and foundations, moving it from pilot to permanent global infrastructure. Begin regional hubs (Africa, Asia, etc.) to adapt EarthNet to different contexts and languages fully. - **Beyond 2027**: approach the 2030 horizon with exponential growth in projects, aiming for that 10 million figure, knowing that even if we achieve a fraction of it by 2030, the world will be markedly better prepared and on a lower emissions trajectory than on our current path. ## **Conclusion: From Vision to Reality** The climate crisis often conjures images of overwhelming, insurmountable catastrophe. But as this whitepaper has shown, there is another story we can tell – one of empowerment, ingenuity, and solidarity at a global scale. _“Ten Million Projects”_ is more than a number; it is a paradigm shift in how we organize ourselves to respond to a planetary emergency. It acknowledges that **people closest to the problems** are often closest to the solutions, especially when given support and connected to others. It harnesses the unprecedented power of AI not to concentrate decision-making in the hands of a few, but to distribute knowledge and resources to the many. And it fundamentally aligns climate action with human development: every project is not just a climate intervention but also an investment in community prosperity and resilience. By emphasizing the **limitations of market-only approaches**, we have highlighted the necessity of collective action and new institutional models. Climate change will not be solved by a price signal alone; it requires cooperation on a scope and scale beyond anything in human history ([Vijayakumar, 2024](https://www.weforum.org/stories/2025/01/collective-action-is-key-to-climate-resilience/)). Yet, as we have also shown, cooperation is within reach if we make use of the tools and understanding now available. Structural traps that have stymied progress – short-term thinking, fragmented efforts, free-riding – can be overcome by changing the way we collaborate and share benefits. The blueprint offered here is essentially a way to _rewire the system_, creating feedback loops of positive action. The EarthNet model – with MCP, communities of practice, and knowledge graphs – is admittedly complex under the hood, but its purpose is simple: to make doing the right thing easier everywhere. It is about lowering the transaction costs of learning and coordinating, much as the internet did for information exchange or marketplaces did for commerce. When a small town can access the combined wisdom of the world’s climate experts with a few clicks, or an engineer’s innovation in one country automatically propagates to thousands of others, we have a force multiplier for change. **This is the promise of AI and networking in the climate fight** – not AI replacing human agency, but amplifying it. Our case study on housing put meat on these conceptual bones, showing that real issues can be tackled in this way. Imagine extrapolating that to every facet of climate solutions. We could see a future where communities compete in the best sense: a friendly race to resilience, where each tries to outdo the other in creative climate action, and then shares how they did it. The competitive spirit, normally channelled through markets, is here channelled through _emulation for the common good_. And where markets play a role, they are guided firmly by community needs and ethical guardrails, thanks to the transparency and collective oversight that EarthNet provides. To the **policymaker** reading this: We urge you to see local initiatives not as a side-show but as the main show, with your role being the ringmaster that keeps it all humming in harmony. Create policies that empower, fund, and connect – the people will amaze you with what they can do when given the chance. To the **philanthropist or funder**: This is your opportunity to back something transformative, beyond the incremental projects. Investing in this social-tech infrastructure will pay dividends across all the thematic areas you care about, from health to environment to social justice, because climate resilience underpins them all. And to the **technologist or researcher**: your skills have never been more needed. By building open systems and focusing your talents on climate and resilience challenges, you can help bend the arc of history towards survival and hope. In closing, the age of AI can be a boon for the age of climate action if we choose to direct it wisely. We stand at a crossroads where one path continues with business-as-usual and techno-centric fixes that might only deepen inequalities and leave many behind. The other path – the one illuminated in this paper – moves toward **collective intelligence and collective action**, harnessing every brain, every willing pair of hands, with AI as glue. Ten million projects by 2030 is an ambitious target, but it is the kind of ambition the climate crisis calls for. Achieving it would mean _millions of solutions unfolding, tailored to their place and time, yet linked to a global endeavour_. It would mean a generation of people who feel empowered rather than helpless in the face of climate change. It would mean that wherever a child is born in the late 2020s, they grow up in a community actively building a livable future. This is a future worth striving for. The blueprint is before us – now it is time to build, project by project, network by network, until the tapestry is complete and resilient. As a famous adage goes (often attributed to an African proverb): **“If you want to go fast, go alone. If you want to go far, go together.”** With the climate crisis, we must do both – go far and fast – and by weaving our efforts together with the thread of technology and the fabric of humanity, we just might succeed. ## **Sources:** Anthropic (2024). _Model Context Protocol (MCP) open standard announcement_. https://www.anthropic.com/news/model-context-protocol Bernstein, T. (2024). _Wildfire Smoke: State Policies for Reducing Indoor Exposure_. Environmental Law Institute. https://www.eli.org/research-report/wildfire-smoke-state-policies-reducing-indoor-exposure. Ding, C., Ke, J., Levine, M. _et al._ Potential of artificial intelligence in reducing energy and carbon emissions of commercial buildings at scale. _Nat Communications_ **15**, 5916 (2024). https://doi.org/10.1038/s41467-024-50088-4 Grantham Research Institute and Duncan Clark. (2014). _Why do economists describe climate change as a 'market failure'?_ The Guardian. https://www.theguardian.com/environment/2012/may/21/economists-climate-change-market-failure HUD (2024). _Green and Resilient Retrofit Program press release_. https://archives.hud.gov/news/2024/pr24-190.cfm Khan, M.R., Munira, S. Climate change adaptation as a global public good: implications for financing. _Climatic Change_ **167**, 50 (2021). https://doi.org/10.1007/s10584-021-03195-w Nicholson,R., (2025). _How AI is boosting efforts to cut buildings’ energy use_. https://www.jll.com/en-us/insights/how-ai-is-boosting-efforts-to-cut-buildings-energy-use NIST/ASHRAE (2022). _Wildfire Smoke Impacts on Indoor Air Quality – recommendation for weatherization and filtration_ . Ostrom, E. (2009). _A Polycentric Approach for Coping with Climate Change_. World Bank Policy Research Working Paper 5095. https://papers.ssrn.com/sol3/papers.cfm?abstract_id=1494833 Rockström, J., et al. (2023). Safe and just Earth system boundaries. Nature, 619(7968), 102–111. https://doi.org/10.1038/s41586-023-06083-8 Schippers, M. C., Ioannidis, J. P. A., & Luijks, M. W. J. (2024). Is society caught up in a Death Spiral? Modeling societal demise and its reversal. Frontiers in Sociology, 9. https://doi.org/10.3389/fsoc.2024.1194597 Søgaard Jørgensen et al. (2023). _Anthropocene traps_ – summarized in Florian Jehn (2024) . Søgaard Jørgensen, et al. (2023). Evolution of the polycrisis: Anthropocene traps that challenge global sustainability. Philosophical Transactions of the Royal Society B: Biological Sciences, 379(1893), 20220261. https://doi.org/10.1098/rstb.2022.0261. Stern, Nicholas. (2008). _The Economics of Climate Change._ American Economic Review 98 (2): 1–37. https://www.aeaweb.org/articles?id=10.1257/aer.98.2.1 Stockholm Environment Institute (2020). _Climate Action Knowledge Graph positioning paper_ . Vijayakumar, C. (2025). World Economic Forum. Collective action is the key to drive urgency in building climate resilience. https://www.weforum.org/stories/2025/01/collective-action-is-key-to-climate-resilience/