CGIAR Breeding for Tomorrow + Excellence in Breeding (EiB) — public-platform AI plant breeding for the Global South; 700K+ germplasm samples + ICRISAT × CIMMYT AI-driven predictive breeding + Gates/FFAR/USAID funding substrate

(Global South — Africa, Asia, Latin America; with 13 CGIAR Research Centers worldwide)

Content

The CGIAR (formerly the Consultative Group on International Agricultural Research, now formally just “CGIAR” since 2008 rebranding) is the largest public-platform AI plant-breeding actor in the world as of 2026. The new 2025-2030 Research Portfolio launched in 2025 with “Breeding for Tomorrow” as one of its centrepiece programs — replacing the prior “CGIAR Genetic Innovation” structure.

CGIAR Breeding for Tomorrow explicitly targets 12 priority crops (per CGIAR rolls-out coverage): wheat, maize, rice, sorghum, millet, groundnut, common bean, cowpea, cassava, sweetpotato, yam, and banana. The program commits to 15-30% more genetic gain than current through AI-augmented breeding pipelines. Substantive deployment target: accelerated breeding through integration of genomic selection, speed breeding, AI-driven predictive breeding, and advanced data management.

Substantive 2026 anchors

ICRISAT × CIMMYT AI-driven predictive breeding initiative (pressroom.icrisat.org) — “The initiative combines artificial intelligence (AI)-driven predictive breeding, genomic selection, speed breeding, and advanced data management.” Targets dryland farmers in Africa + India. Substantively operational; this is the most concrete CGIAR-side AI-breeding project for Africa + South Asia we have surfaced.

CGIAR Genebanks — over 700,000 samples of crop, forage, and tree diversity conserved. The Genebanks Accelerator is part of CGIAR’s 2025-2030 portfolio. This is the world’s largest publicly accessible plant-genetic-resources infrastructure — substantively larger than any single state or corporate gene bank. The germplasm is held in trust under the International Treaty on Plant Genetic Resources for Food and Agriculture (ITPGRFA) Article 15 multilateral system.

Independent evaluation of EiB (CGIAR IAES 2022) — CGIAR Independent Evaluation Function evaluated the EiB Platform from August 2021 to April 2022. This is the first formal evaluation of EiB since its launch. Substantive: confirms EiB as a system-level coordination platform, not as a stand-alone breeding program; EiB’s primary delivery function is breeding-program support and modernization for the 13 CGIAR Research Centers.

CGIAR Breeding for Tomorrow — substantive scope:

• 12 priority crops committed to 15-30% genetic-gain uplift
• AI-driven predictive breeding + genomic selection + speed breeding + advanced data management
• Long-term holistic approach with shared data systems
• Proactive engagement with women + Indigenous communities + youth on breeding pipeline design

13 CGIAR Research Centers (per current structure):

• CIMMYT (Mexico, maize + wheat)
• ICRISAT (India, dryland crops: sorghum, pearl millet, chickpea, pigeonpea, groundnut)
• IRRI (Philippines, rice)
• ICARDA (Lebanon/regional, dryland crops: wheat, barley, chickpea, faba bean, lentils, lathyrus)
• CIP (Peru, potato + sweetpotato)
• IITA (Nigeria, African staple crops: cassava, yam, maize, cowpea, soybean, banana)
• AfricaRice (Côte d’Ivoire, African rice)
• WorldFish (Malaysia, fish)
• CIFOR-ICRAF (Indonesia/Kenya, forests + agroforestry)
• Bioversity International (Italy, agricultural biodiversity)
• ILRI (Kenya, livestock)
• ICRAF (Kenya, agroforestry)
• IFPRI (USA, food policy)

Substantive scope: 13 centers × ~12 priority crops × ~700,000 germplasm samples + ~AI-augmented pipelines + ITU multilateral-system governance = the world’s substantive public-platform AI plant-breeding actor.

Funding substrate (the substantive non-obvious observation)

CGIAR’s funding base is a multi-billion-dollar multilateral + bilateral + philanthropic + foundation mix with specific substantive observability:

• Bill & Melinda Gates Foundation — the largest single CGIAR donor historically; Gates Foundation’s agricultural development portfolio is structured largely through CGIAR as delivery channel for genetic-gain / smallholder farmer outcomes.
• USAID Feed the Future — large US-government funding for crop improvement research in target countries.
• FFAR (Foundation for Food and Agriculture Research) — US-based foundation (public-private partnership model established by 2014 Farm Bill); FFAR has dedicated Accelerated Crop Breeding programme + Crop-Animal Systems Research RFA ($4M funding pool as of 2026); explicitly aligns with CGIAR / National Plant Germplasm System / USDA efforts.
• GIZ (Deutsche Gesellschaft für Internationale Zusammenarbeit) — German federal enterprise for international cooperation; multiple GIZ-funded breeding-related projects in CGIAR partner countries.
• World Bank — historically a major CGIAR funder through IDA + IBRD lending instruments.
• UK FCDO (Foreign, Commonwealth & Development Office) — historic CGIAR supporter.
• Australian DFAT / ACIAR — Asian-region CGIAR support.
• Canadian AAFC — historical CGIAR supporter.
• Brazilian EMBRAPA — host of CGIAR-related programs.
• European Commission / EU Horizon Cluster 6 — recent funding.
• CGIAR Trust Fund — pooled funding mechanism since 2017 (Window 1 / Window 2 / Window 3 funding bilaterals).

This is a substantive multi-billion-dollar public-platform funding stack that does not appear at the same level of transparency or attention as the corporate-multinational seed industry funding. The total CGIAR annual budget is in the USD$900M-1.5B+ range across all centers, channels, and project funding; specific annual figures published in CGIAR financial reports.

CGIAR as the cross-cutting public-platform anchor

This is the substantive non-obvious observation:

The world’s substantive plant-breeding AI activity runs through:

• 4-5 multinational seed-corporations (Bayer, Syngenta Group [China-owned], Corteva, BASF, KWS etc.)
• Chinese state-orchestrated cluster (CAAS + Syngenta Group China + BGI/MGI + smart-breeding national plan)
• US land-grant university network (USDA-NIFA AI Institutes)
• Canadian provincial-commodity-led cluster (BeanGPT + Ontario Bean Growers + ARIO)
• Brazilian EMBRAPA + multinational-corporate-pipelined
• EU private-breeding-cluster (Limagrain + KWS + RAGT)
• CGIAR public-platform for the Global South — explicitly multilateral-system governance, ITPGRFA-trust framework, and substantive deployment of AI-augmented breeding in low- and middle-income countries.

CGIAR Breeding for Tomorrow is the only actor combining (a) substantive AI-augmented breeding pipelines, (b) multilateral-system governance (not corporate IP), and (c) 700K+ germplasm samples held in trust under ITPGRFA. No other actor combines these three.

Cluster shape comparison

The CGIAR cluster’s structural distinctiveness: multilateral-system + Global South orientation + 700K-germplasm public-domain substrate + 12 priority-crop mandate.

What this unit is doing in the taxonomy

The CGIAR unit is the public-platform cross-cutting anchor for AI plant breeding globally. Complements:

• units/plant-breeding-ai-methodology.md — methodology stack
• units/uog-bean-gpt-najafabadi.md — Canadian provincial-commodity-pipelined cluster
• units/brazilian-seed-ai-academic-research-led.md — Brazilian multinational-pipelined cluster
• units/longping-yuan-caas-china-seed-ai.md — Chinese state-orchestrated cluster
• units/bayer-syngenta-corteva-multinational-pipelines.md (forthcoming) — multinational-corporate-pipelined
• units/limagrain-kws-ragt-eu-private-plant-breeding.md (forthcoming) — EU private-cluster
• units/usda-ars-iowa-state-aiira-us-land-grant.md (forthcoming) — US land-grant

The CGIAR unit, combined with the forthcoming corporate-multinational + EU + US land-grant units, completes the substantive regional coverage of the AI plant-breeding global scan.

Why it matters for talks

1. CGIAR Breeding for Tomorrow is the substantive public-platform AI plant breeding actor globally. Talks on the regional-cluster-comparison archetype (06) should name CGIAR Breeding for Tomorrow, the 12 priority crops, and the 15-30% genetic-gain commitment.
2. The ITPGRFA multilateral-system governance frame is structurally distinct from the corporate-IP frame and from the academic-tier frame. Worth naming.
3. The 700K+ germplasm in trust is the world’s substantive publicly accessible plant-genetic-resources infrastructure. Substantively larger than any single state or corporate gene bank.
4. The ICRISAT × CIMMYT AI-driven predictive breeding initiative is the substantive operational example. Concrete dryland-crop AI-augmented pipeline for Africa + India.
5. The funding substrate (Gates + USAID + FFAR + GIZ + World Bank + multilateral) is the substantive multi-billion-dollar public-platform funding stack — worth naming for talks on agrifood AI funding architecture globally.
6. CGIAR’s substantive scoping past the US-PRC-EU axis — CGIAR anchors the Global South: ICRISAT (India) + IITA (Nigeria) + AfricaRice (Côte d’Ivoire) + IRRI (Philippines) + CIP (Peru) + ICARDA (Lebanon). The regional-cluster-comparison archetype (06) gains substantive Global-South coverage from naming CGIAR.

Critical context

• CGIAR is undergoing structural reform — the 2025-2030 portfolio restructuring is the substantive reorganization since the One-CGIAR 2019 reform attempt. The substantive test is whether the Breeding for Tomorrow program delivers 15-30% genetic-gain uplift by 2030.
• Multilateral-system governance is the substantive distinctiveness. The genetic resources held in trust by CGIAR Research Centers are under ITPGRFA’s multilateral system of access and benefit-sharing (Article 15). This is substantively distinct from corporate-multinational IP regimes (Bayer, Syngenta Group, Corteva, BASF) and from Chinese state-orchestrated IP-strengthening (2021 amended Seed Law).
• The CGIAR Independent Evaluation Function of EiB (2021-22) is the formal evaluation record. The substantive takeaway: EiB is a system-level coordination platform, not a stand-alone breeding program. The deployment outcome depends on each Research Center’s own capacity to use EiB’s toolkit.
• The Gates Foundation’s CGIAR-leaning portfolio is substantive — the Gates Foundation’s Agricultural Development program has historically directed major funding through CGIAR. The substantive critique (substantively voiced by GRAIN and others): Gates Foundation funding tilts CGIAR toward smallholder-farmer productivity goals over farmer-seed-sovereignty / biodiversity-conservation goals.
• The 12 priority crops substantively cover Global-South staples: cassava, yam, sweetpotato, cowpea, millet, sorghum, groundnut, common bean are the substantive Global-South staples that the EU / Chinese / North American clusters largely do not address. The CGIAR cluster’s crop orientation substantively delimits the AI-Augmented-Public-Platform coverage of the world’s food-system vulnerability hotspots.
• Funding-substrate transparency is uneven — Window 1 / Window 2 / Window 3 CGIAR Trust Fund has been substantively criticised for opacity since 2017. Gates Foundation’s CGIAR funding share has grown relative to OECD-DAC multilateral-channel funding.
• CGIAR’s AI-augmented breeding is operationally at speed-breeding + GS + AI-driven predictive breeding scale, not at the speed of multinational-corporate pipelines. The 15-30% genetic-gain commitment is the program’s substantive KPI; the deployment scale will be verified through 2025-2030.
• The substantive tension: CGIAR’s multilateral-system governance is structurally resistant to corporate-IP licensing — some CGIAR-bred varieties have been commercialised through multinational partnerships (e.g., IRRI × Bayer hybrid rice lines), but the germplasm-substrate remains in trust. AI augmentation of CGIAR pipeline outputs may complicate this — once a CGIAR-bred variety has been augmented by AI at CGIAR, who owns the AI-augmented IP? This is the substantive governance question for 2026-2030.