Federated AI in PlaceMarch 18, 2026

Black Gold and Open Hands: An Invitation to Build the Tools Africa Needs to Pull Carbon from the Sky

AI Has Made Climate Change Worse. That Needs to Be Said.

# Black Gold and Open Hands: An Invitation to Build the Tools Africa Needs to Pull Carbon from the Sky

By Man Withahat | AI Almost Daily | March 2026

Published under CC BY-SA 4.0 — share freely with attribution.

*This essay is an argument, an invitation, and a challenge. It argues that corporate AI — for all its noise and promise — has made the climate crisis measurably worse. It invites a different approach: decentralized, locally owned, community-federated AI tools that build resilience instead of dependency. And it challenges developers, educators, students, and researchers to build those tools now, with Africa as the proving ground — a continent already demonstrating that the partnership of AI, Humans, and Nature, applied efficiently and cheaply, can pull carbon from the sky while feeding children and rebuilding soil. Every reference in this essay is a building block. What gets built from them is the question.*

The Number That Matters

The number is approximately 424. That is the current concentration of carbon dioxide in Earth's atmosphere, measured in parts per million. Before the industrial era it was roughly 280. The Paris Agreement talks about reducing emissions. That is necessary but not sufficient. What the planet actually needs is to bring that number down — not just slow its rise, but reverse it. That is carbon drawdown, and it is a fundamentally different challenge from emissions reduction.

Most climate technology conversation focuses on the supply side: cleaner energy, electric vehicles, greener cement. These matter. But they do not remove what is already up there. For that, we need to pull carbon out of the atmosphere and lock it somewhere stable for centuries or millennia. There are expensive, energy-hungry engineered approaches to this. And then there is biochar — ancient, low-tech, scalable, and sitting quietly in the middle of a continent whose time is coming.

Corporate AI Has Made Things Worse. That Needs to Be Said First.

Before anything else, the record needs to be set straight — clearly, with data, and without apology.

AI has been sold as an answer to almost everything — climate change included. Venture capital pitches, government reports, and conference keynotes position AI as a civilizational force for good. AGI — artificial general intelligence — is presented as the coming solution to humanity's hardest problems. For all its hype and promise, corporate AI has done nothing measurable to address climate change. It has made it worse.

MIT Technology Review (2025) found that "the adoption of AI and the accelerated server technologies that power it has been the primary force causing electricity demand from data centers to skyrocket after remaining stagnant for over a decade." By 2028, AI-specific electricity use in the United States alone is projected to reach 165–326 terawatt-hours per year — "enough to power 22% of US households each year."

A peer-reviewed Nature Sustainability study from Cornell University (2025) calculated that current AI growth would generate "additional annual carbon emissions from 24 to 44 Mt CO₂-equivalent between 2024 and 2030" — the equivalent of adding 5 to 10 million cars to U.S. roads — while also draining 731 to 1,125 million cubic meters of water per year.

Goldman Sachs Research forecasts that about "60 percent of the increasing electricity demands from data centers will be met by burning fossil fuels, increasing global carbon emissions by about 220 million tons."

The Shift Project warned that by 2030, "the data center sector could contribute two to three times more to climate disruption than in 2020."

The pattern is consistent and damning. Centralized, cloud-dependent AI runs on hyperscale data centers powered substantially by fossil fuels, serving profit extraction for a small number of corporations, creating dependencies rather than capabilities in the communities it claims to serve. A Cornell researcher put it plainly: "The AI infrastructure choices we make this decade will decide whether AI accelerates climate progress or becomes a new environmental burden."

That choice has already been made by the corporate sector. The question is whether a different choice can be made by everyone else.

A Different AI Is Possible — And Already Being Built

The essays collected at AI Almost Daily point toward a different model — not AI as a centralized oracle owned by corporations and monetized through subscriptions and data extraction, but AI as a community tool: decentralized, locally deployed, offline-capable, long-memoried, and federated across cooperating communities rather than dependent on distant servers.

### The Concept: ygw.local.coop

Imagine a system — call it ygw.local.coop, for "your good world, local cooperative" — that runs on modest hardware in a village, a farming cooperative, a school, or a community health clinic. It holds the community's own knowledge: its soil records, its crop histories, its health observations, its governance decisions, its cultural knowledge. It does not send that data to a corporate server. It learns from the community and for the community. It can share — selectively, on the community's terms — with neighboring nodes in a federated network of similar systems. It runs when the internet is down. It speaks the local language. It remembers what worked last season and the season before. It is, in the fullest sense, a community commons for intelligence.

This is not science fiction. The technical architecture — federated learning, edge AI, small language models, peer-to-peer data sharing — exists and is maturing rapidly. What is missing is the will to build it for communities that cannot pay corporate subscription rates, and the specific tools that make it immediately useful in those communities.

The research literature on federated learning is precise about what this architecture offers. A 2025 survey in Frontiers in Computer Science describes decentralized federated systems where "client devices are connected in a peer-to-peer network, enabling direct communication between them and eliminating the need for a central server." A comprehensive 2025 survey of federated learning for edge AI frames the stakes directly: "Heavy dependence on centralized infrastructure creates single points of failure and exacerbates the carbon footprint of repeated long-haul transfers and always-on connectivity, conflicting with emerging goals in green and resilient AI."

The alternative is already being deployed in Africa. In Nigeria, Integrated Aerial Precision runs agricultural AI tools where "AI models are compressed, quantised, and deployed via edge computing, making high-tech accessible even in areas without internet." In Malawi, Opportunity International's Ulangizi AI chatbot delivered real-time agricultural guidance to thousands of smallholder farmers in its first year. FurtherAfrica (2025) reports that "AI-enabled soil sensors and imaging devices can provide farmers with real-time insights on crop health, irrigation, and disease management without requiring constant internet access."

This is the Keep Building Back Better cycle: modest tools, locally owned, compounding over seasons and years, building a community commons of knowledge and capability that no corporation can extract and no connectivity outage can destroy.

The Partnership That Actually Works

The climate crisis will not be solved by any single actor, technology, or ideology. What the evidence points to consistently is a three-way partnership: AI, Humans, and Nature working in close collaboration — each doing what it does best, none substituting for the others.

Nature does the actual carbon sequestration. Soil microbes, fungal networks, plant roots, and the slow chemistry of mineralization lock carbon away in forms stable enough to matter climatically. No algorithm does this. Nature does this. The role of technology is to support and amplify what nature already knows how to do.

Humans provide the labor, local knowledge, cultural context, and sustained attention that no algorithm can replicate. A farmer who has watched a particular field for decades knows things no sensor will capture for years. That knowledge is the most valuable input in the system. It provides the moral anchor: the goal is not carbon credits on a spreadsheet. The goal is children who are well, soil that lasts, communities that can feed themselves through a destabilizing climate.

AI — the right kind of AI — contributes pattern recognition, long memory, and coordination capacity at scales humans cannot manage alone: which soil types respond best to which biochar feedstocks, what the soil pH was three seasons ago, what the cooperative decided at its last meeting, how to estimate carbon sequestration for a community too small to access carbon markets. These are real and bounded contributions, meaningful only when grounded in human knowledge and natural systems.

A 2025 paper in npj Climate Action warned that AI-driven climate solutions risk "cognitive offloading and epistemic stagnation" — communities handing their problem-solving to algorithms and losing the local adaptive knowledge most needed as climate conditions change unpredictably. The antidote is partnership, not dependency. The ygw.local.coop model is designed specifically around this principle: AI as a memory and coordination layer for human and natural intelligence, not a replacement for either.

What the Science Shows: Biochar as the Carbon Engine

Within this three-way partnership, biochar — charcoal produced from organic material burned in low-oxygen conditions and applied to soil — is the carbon engine. When biomass decomposes or burns normally, its carbon returns to the atmosphere as CO₂. When it becomes biochar, that carbon is stabilized in a form that persists in soil for centuries to millennia.

The sequestration numbers are not trivial. A 2025 synthesis in Communications Earth & Environment reviewed 19 published studies and found:

"Biochar is a carbon dioxide (CO₂) removal strategy that supports food security, sustainable land management and the circular economy... Recent independent assessments estimate sustainable mitigation potential of biochar systems at 2.6–10.3 Pg CO₂ equivalent yr⁻¹."

A companion 2025 paper in the same journal revealed an additional mechanism:

"All biochar increased soil carbon storage and meanwhile further sorb CO₂ for more carbon sequestration. The Ca or Mg originated in biochar could react with more CO₂ and eventually be converted into inorganic carbonate minerals in the amended soil."

The carbon sink is larger than the biochar alone. A ScienceDirect review (2024) puts the century-scale potential plainly: "Each year, biochar effectively captures an estimated amount of CO₂ ranging from 1 to 35 gigatons (GtCO₂) and 78 to 477 GtCO₂ over this century."

Professor David Laird of Iowa State University told Inside Climate News (2024): "It's going to take a major industry effort to actually remove a gigaton of CO₂ from the atmosphere. This is a pathway towards that. It can be done."

Biochar does not work in isolation. Its greatest power emerges when embedded in regenerative agriculture — a systems approach that rebuilds rather than depletes the living soil. A 2025 ScienceDirect review describes the combination: "Regenerative strategies, which prioritize biological and ecological balance, include a variety of measures, including conservation agriculture, crop rotation, cover-cropping, organic management, biochar, and agroforestry." Biochar is a scaffold. Regenerative agriculture is the living system that occupies it.

Africa: Where the Partnership Demonstrates Its Power

Africa contains roughly 60 percent of the world's uncultivated arable land. Much of it is degraded — acidic, thin, depleted. This is precisely where biochar and regenerative agriculture together show their most dramatic effects, and where a locally federated AI partnership has the most to offer.

An MDPI Agronomy review (2024) observed that "with most of the circular agricultural practices still central to food production in the developing African countries, the continent can spearhead a global return to circular agriculture." Africa is not a problem to be solved by outside technology. It is a repository of traditional regenerative knowledge that the industrialized world abandoned and now desperately needs to relearn. It is also the continent best positioned to demonstrate — at scale, with the youngest and fastest-growing population on earth — that the AI-Human-Nature partnership can work without corporate extraction.

In the Democratic Republic of the Congo's Yangambi Engagement Landscape, CIFOR-ICRAF researchers ran biochar trials across three growing seasons in Farmer Field Schools between 2024 and early 2025. The baseline was stark: "Baseline soil tests revealed highly acidic soils with very low natural fertility." Farmers produced biochar in situ using bamboo, cassava stems, and maize cobs in Kon-Tiki pit kilns. Results came quickly: "Since July 2024, 86 percent of trained households have adopted kitchen gardens." The lead researcher said: "What touched the team and I when we first visited the landscape was the physical malnutrition in children. Nobody needs to tell the story."

The RESADE project, coordinated by the International Center for Biosaline Agriculture across Botswana, Liberia, Mali, Senegal, Sierra Leone, Togo, and Zimbabwe, found that biochar at 20–30 tonnes per hectare increases available soil water content by 16–40 percent compared to untreated soils. Farmer David Kpogbeh in Liberia put it plainly: "We get heavy rainfalls here in Liberia and have sandy soils. So by applying this biochar, we can ensure that our crops get the nutrients and water they need rather than having it all wash away."

In Zambia's Choma District, youth farmer Muntinta Handondo captured the urgency: "We cannot continue farming the way we used to farm. We need to have a way of helping the soil to keep water for a longer time." Her Trywell Youth Club is building exactly the kind of locally led, regenerative system this essay describes.

A Scientific Reports study (2024) working with communities in Ghana, Côte d'Ivoire, Senegal, and Uganda found that the difference between a technology that takes root and one that withers is whether communities are centered as designers, not recipients — "locally led, sustainable development pathways within rural communities." That framing maps directly onto the ygw.local.coop architecture: tools owned by communities, built with communities, serving community-defined goals.

Africa's own agricultural waste streams — coffee husks, sugarcane bagasse, cassava stems, maize cobs, bamboo, wood prunings — are the feedstock. No imports needed. Convert what is being burned openly into biochar, embed it in regenerative soil systems, track it with a locally deployed AI memory tool, share what works across a federated network of cooperatives. That is the cycle. That is the demonstration the world needs.

Projects Already Seeding the Ground

Africa is vast, and generalizations fail it. The projects worth watching — and worth connecting with — are specific, small, and particular. Each one is a building block.

Biochar Life / Carbon Village Model — A public benefit corporation with roots in Warm Heart Worldwide, working with smallholder farmers across Africa and Asia through a community-led framework integrating biochar, trees, and soil regeneration. Developed with PlantVillage at Penn State and Carbon4Good. CEO Jason Highberger: "We are poised to scale our efforts, drive meaningful change, and further our mission of environmental stewardship and community upliftment in Africa."

RESADE / International Center for Biosaline Agriculture — Active in seven sub-Saharan African countries, training farmer cooperatives to make and apply biochar from local waste. A replicable, documented model with measurable field results and farmer testimony — a natural partner for a federated knowledge network.

CIFOR-ICRAF Yangambi Biochar Initiative, DRC — Field-school-based biochar training using Kon-Tiki kilns in one of Central Africa's most nutritionally vulnerable regions. Peer-reviewed, documented, expandable — and exactly the kind of node a federated community memory system could support.

Sasakawa Africa Association (SAA) — Operating across multiple African countries with regenerative agriculture as its explicit organizing framework. Presented at ACAT 2025 in Kigali with biochar featured alongside soil restoration and climate-smart farming.

African Biochar Partnership (ABP) — A continental coordination body connecting biochar producers, researchers, and practitioners across Africa. An entry point for anyone wanting to locate existing networks rather than start from scratch.

Project Regeneration — Biochar Nexus — An open-access hub for biochar science, project examples, and regenerative agriculture connections globally — including links to indigenous biochar traditions on the continent.

GreenAgri South Africa — Covering biochar, regenerative agriculture, and the South Africa Carbon Removal Equipment & Suppliers Guide 2025 — the first practical biochar implementation guide tailored to the southern African context.

For researchers wanting the critical perspective — essential reading before building anything for African farming communities — the Journal of Peasant Studies analysis of biochar and "green grabs" examines how "carbon market logics are not only threatening large-scale land grabs for biochar feedstocks but also other forms of resource, labour and ecological appropriation." The ygw.local.coop model is explicitly designed to prevent this: community ownership of tools and data as the non-negotiable foundation.

A Direct Call to Teachers, Professors, and Their Students

\[Author's Perspective\]

This section is addressed specifically to educators and students at land-based institutions — agricultural colleges, environmental science programs, engineering schools, forestry departments, and community colleges with applied research missions.

The tools this essay describes are not beyond the reach of a well-supervised student project. They are exactly the kind of work that wins EPA P3 grants, fulfills senior capstone requirements, and produces publishable results. Here is what needs building — each one a potential node in the ygw.local.coop federated network:

1. An offline-capable biochar production guide for smallholder farmers — Available in major African languages, deployable on basic Android devices without internet, covering feedstock identification, simple kiln construction, production quality indicators, and application rates by soil type and crop. A computer science or UX student builds the app. An agronomy student supplies the content. A languages student handles translation and cultural adaptation. One interdisciplinary team, one semester, one tool that could be used for decades.

2. A local carbon sequestration calculator — Allowing a village, cooperative, or NGO to estimate how much CO₂ they are drawing down through biochar production and application. Simple inputs: biomass weight, feedstock type, estimated pyrolysis temperature, application area. Output: approximate tonnes CO₂ sequestered with a plain-language uncertainty range. An engineering student project. A data science thesis. A chemistry lab validation exercise.

3. An open-source community memory tool for smallholder farmer cooperatives — A locally deployable record system — a node in the federated network — tracking what was planted, when, what amendments applied, what yields resulted, soil pH before and after. Across seasons and across members. Owned by the cooperative. No ongoing subscription to a foreign platform. Long memory, local deployment, federated sharing on the cooperative's terms.

4. A community co-design toolkit — Building on the methodology of the 2024 Scientific Reports study — an open digital version of the Triple-Layered Business Model Canvas process for evaluating biochar business models on social, ecological, and economic dimensions simultaneously. A design school project. A business school capstone. A development studies thesis.

The FARA Africa funding database lists active small grants for exactly this kind of applied agricultural research in sub-Saharan Africa. The EPA P3 program — which funded community college biochar research for over a decade — remains open to student teams. These are real funding pathways, not aspirational ones.

Professors: the researchers at CIFOR-ICRAF, ICBA, SAA, and Biochar Life are approachable, collaborative, and actively looking for partners. Your students' work could be deployed in Yangambi or Choma within a year of completion. That is not a hypothetical. It is an invitation.

Where to Learn More

The science of biochar and carbon drawdown
International Biochar Initiative | US Biochar Initiative | Project Regeneration Biochar Nexus

African field projects and organizations
biocharafrica.org | RESADE/ICBA | CIFOR-ICRAF Forests News | Biochar Life | Sasakawa Africa Association

Regenerative agriculture in Africa
GreenAgri South Africa | MDPI Agronomy — Regenerative Agriculture special collection

Decentralized and federated AI
Federated Learning for Edge AI — 2025 survey | Deep Federated Learning review, Frontiers (2025)

Critical perspectives (essential reading)
Journal of Peasant Studies — Green Grabs and Biochar

Funding for student and small projects
FARA Africa Funding Database | EPA P3 Student Design Competition

AI and climate — honest assessment
MIT Technology Review — AI energy use | npj Climate Action — AI and cognitive offloading

Standing on the Shoulders of Those Who Went First

This essay would not have been written without the foundational work of Roland Bunch — an agricultural development practitioner who has spent more than four decades building community resilience across more than 50 nations of Africa, Latin America, and Asia.

His book Two Ears of Corn: A Guide to People-Centered Agricultural Improvement (1982) established the foundational principle that runs through every section of this essay: that agricultural transformation works when communities lead it, and fails when outside experts impose it. Published in ten languages, it is one of the all-time best-sellers on the design of agricultural development programs in developing nations. It is not a book about technology. It is a book about listening.

His second essential book, Restoring the Soil: A Guide for Using Green Manure/Cover Crops to Improve the Food Security of Smallholder Farmers, synthesizes decades of field-based research with smallholder farmers across the world. The carbon sequestration implications of his work are striking. Bunch states that on at least 75% of the developing world's wastelands, farmers can be growing crops at double the national average within two to three years, and triple within four to six — while, as a free by-product, sequestering carbon in the soil at a rate of over 4 tonnes per hectare per year. At that level, the world's farmers and ranchers alone could sequester over half the carbon needed to meet the Paris Accord targets by 2100.

Bunch's diagnosis of Africa's agricultural crisis cuts through a great deal of noise. Presenting at Chico State in 2017, he stated that Africa's drought crisis is being caused primarily by soils that can no longer infiltrate water because they have lost all of their organic matter — and that regenerative agriculture, particularly the use of green manure and multi-species cover crops, is the solution.

The ygw.local.coop vision described in this essay is, in one sense, an attempt to build the memory and coordination tools that Bunch's community-centered method has always required but never had access to. His work proves the human and agricultural model works. The question this essay asks is: what locally owned, offline-capable, federated AI tools would make it scale?

Both books are available through ECHOcommunity.org. Restoring the Soil is also available as a free PDF from the publisher. Anyone serious about the Africa biochar and regenerative agriculture opportunity should read them before writing a single line of code.

The Scale of What Is Possible

To move the atmospheric CO₂ number meaningfully downward requires removing carbon at gigaton scale over decades. Yale University's Environment 360 quotes Darko Matovic, mechanical engineering professor at Queen's University: "The idea of biochar is really attractive, and in my opinion it is the only viable way of removing carbon dioxide from the atmosphere."

Africa's population will reach 2.5 billion by 2050. Its agricultural land base is vast, much of it degraded, all of it capable of regeneration. Its biomass waste streams are an untapped carbon capture resource of continental scale. Its farmers are already motivated. Its traditional regenerative knowledge is intact. Its young population is the most energetic demographic force on earth for the coming century.

What is missing is not motivation, not land, not biomass, not science, and not farmers ready to act. What is missing is appropriate, accessible, locally deployable tools — owned by communities, federated across cooperatives, built cheaply with the AI we already have, and focused on the AI-Human-Nature partnership that actually works.

Every reference in this essay is a building block. The projects listed are real, reachable, and looking for partners. The funding pathways exist. The science is solid. The farmers are ready.

The atmospheric CO₂ counter does not care about our institutional barriers, our carbon market inefficiencies, or our tendency to build tools for people who already have everything. It just keeps climbing.

Someone in this continent, on this soil, with this biomass, is ready to help bring it down.

Build them something useful. Keep building it better.

Man Withahat writes AI Almost Daily.
Published under Creative Commons CC BY-SA 4.0. Share freely with attribution.

Key References

### Foundational Works
- Two Ears of Corn: A Guide to People-Centered Agricultural Improvement — Roland Bunch (1982; available via ECHOcommunity.org)
- Restoring the Soil: A Guide for Using Green Manure/Cover Crops to Improve the Food Security of Smallholder Farmers — Roland Bunch (free PDF via ECHOcommunity.org)
- Roland Bunch — Center for Regenerative Agriculture and Resilient Systems, Chico State University
- Roland Bunch speaker profile — Biodiversity for a Livable Climate

### Biochar Science & Carbon Sequestration
- Estimates vary but credible evidence points to gigaton-scale climate change mitigation potential of biochar — Weng & Cowie, Communications Earth & Environment (2025)
- Biochar-amended soil can further sorb atmospheric CO₂ for more carbon sequestration — Gui et al., Communications Earth & Environment (2025)
- Biochar in global carbon cycle: Towards sustainable development goals — Sustainable Chemistry for the Environment, ScienceDirect (2024)
- Biochar for a sustainable future: Environmentally friendly production and diverse applications — Energy Conversion and Management: X, ScienceDirect (2024)
- Biochar carbon sequestration: Can it save us from climate change? — Inside Climate News (2024)
- Refilling the carbon sink: Biochar's potential and pitfalls — Yale Environment 360

### Regenerative Agriculture
- Recent advances in regenerative sustainable agricultural strategies for managing soil carbon and mitigating climate change consequences — Soil & Tillage Research, ScienceDirect (2025)
- Circular regenerative agricultural practices and their role in food security in Africa — MDPI Agronomy (2024)
- Biochar as a soil amendment for sustainable agriculture — MDPI Agronomy (2024)

### Africa: Field Projects & Research
- Biochar and soil resilience in Yangambi, DRC — CIFOR-ICRAF Forests News (2025)
- RESADE: Rehabilitating degraded soils in sub-Saharan Africa with biochar — ICBA (2024)
- How to use biochar to improve soil fertility — Zambia farmer case study, KCOA Africa
- Co-designing sustainable biochar business models with sub-Saharan African communities — Mohammed et al., Scientific Reports (2024)
- Sasakawa Africa Association at ACAT 2025, Kigali — regenerative agriculture program
- Biochar Life investment to accelerate biochar among small farmers in Africa — Carbon Herald
- Biochar opportunities in Africa — African Biochar Partnership overview, ClimateSAN
- Biochar Nexus — Project Regeneration
- GreenAgri South Africa — biochar and regenerative agriculture resources (2025)
- The new carbon colonialism? Examining land and carbon grabs — Journal of Peasant Studies (2013)

### Decentralized & Federated AI
- Deep federated learning: a systematic review of methods, applications, and challenges — Frontiers in Computer Science (2025)
- A comprehensive survey of federated learning for Edge AI: recent trends and future directions — Preprints.org (2025)
- Edge AI in Africa: Unlocking innovation beyond the cloud — FurtherAfrica (2025)
- How AI is revolutionizing smallholder farming in Malawi — Opportunity International / Ulangizi (2025)

### Corporate AI: Climate Impact
- How much energy does AI use? It's a surprisingly tricky question — MIT Technology Review (2025)
- Environmental impact and net-zero pathways for sustainable artificial intelligence servers in the USA — Xiao et al., Nature Sustainability (2025)
- Responding to the climate impact of generative AI — MIT News / Goldman Sachs (2025)
- AI and data centers: The Shift Project warns of unsustainable energy trajectory — DirectIndustry e-mag (2025)
- Embedding the frontiers in climate action: navigating the evolving earth-human ecosystem — Zheng et al., npj Climate Action (2025)

### Funding & Student Resources
- FARA Africa funding database for agricultural research
- EPA P3: People, Prosperity and the Planet student design competition