Korea — Smart Farm Innovation Valley (Sangju, Gimje, Milyang, Goheung) — state-anchored four-site cluster under RDA / MAFRA

East-Asia (Korea; Republic of Korea, four provincial sites: Sangju-Gyeongsangbuk, Gimje-Jeollabuk, Milyang-Gyeongsangnam, Goheung-Jeollanam)

Korea — Smart Farm Innovation Valley (Sangju, Gimje, Milyang, Goheung)

Content

The Smart Farm Innovation Valley is the Republic of Korea’s signature state-anchored agrifood AI / smart-farming cluster programme. It is operational across four provincial sites (Sangju in Gyeongsangbuk-do, Gimje in Jeollabuk-do, Milyang in Gyeongsangnam-do, Goheung in Jeollanam-do) under the Ministry of Agriculture, Food and Rural Affairs (MAFRA) and the Rural Development Administration (RDA), with the broader FAO Digital Villages Initiative recognising Sangju as a Digital Village (the corpus’s cleanest international anchor).

The programme’s mandate addresses the ageing-farmer succession crisis that Korea, like Japan, faces: per the Act on Fostering and Supporting Smart Farming (enacted), the target is to convert 30% of agricultural output to smart-farming systems by 2027.

Cluster composition (per site):

Each cluster has:

• Greenhouse complexes (operational agricultural facilities)
• Rental smart farms (entry-point for young farmers without infrastructure)
• Test centres (where research institutions and businesses develop / test / demonstrate new smart-farming equipment)
• Startup Incubation Centres (SIC) — the mentorship-and-skill-development core; the “education, training, capital, and resources” pipeline
• Distribution centres (export + domestic-market support)

Trainee pipeline (Sangju + Gimje, then expanded):

Site-by-site specialisation (FAO-anchored):

Governance structure: The services are jointly operated and managed by individual farmers, research institutions, and private-sector businesses — covering production, quality control, processing, sales/marketing. Upstream and downstream industries are integrated in the cluster design.

Statutory and infrastructure frame (parallel / related):

• Act on Fostering and Supporting Smart Farming (Korea) — sets the 30%-by-2027 target. (Encoded as the state-policy unit units/korea-act-fostering-smart-farming.md.)
• Smart Farmland Distribution Centers expansion: 14 locations in 2023 → 26 in 2024 → planned 100. (FFTC policy brief on MAFRA Major Policy Plans for 2024.)
• 2025 MAFRA designated four smart-farming fostering zones at the si/county level — adds administrative machinery beyond the original four Innovation Valley sites.
• KT Corp. AX Platform Company initiative (Nov 2024 launched upgraded AI-powered digital farming advisory platform; July 2026 $5.4B / 18 trillion won overall announcement) — corporate-telecom-side parallel infrastructure.

What this unit is doing in the taxonomy

Sector position. The cluster straddles multiple sector-position tags: on-farm-production-protected is primary (controlled-environment agriculture inside the cluster is the primary activity); distribution-and-retail (the Smart Farmland Distribution Centers); consumption (export support and market integration).

AI technique class. predictive-ml is the named cluster ML surface (insect pest/disease forecasting, remote temperature/humidity control, smart-farming solutions). sensors-and-iot-ml (smart-farm IoT sensors). generative-ai-llms is the emergent-class surface (not yet published at deployment scale but Korea’s industrial direction).

Purpose. The cluster serves multiple purposes simultaneously: food-security-sovereignty (state-stewarded agricultural-output target), market-access (export support infrastructure), supply-chain-efficiency (distribution-centre integration), climate-adaptation (Goheung tropical-crop R&D), worker-conditions (rental smart farms + ageing-farmer succession target).

Region tag. East-Asia (Korea; Republic of Korea, four provincial sites: Sangju-Gyeongsangbuk, Gimje-Jeollabuk, Milyang-Gyeongsangnam, Goheung-Jeollanam) — explicit site listing in the region tag captures the federated nature of the cluster.

Data governance. mixed — state-anchored cluster with vendor-ICT-stack controlled at site level. The specific per-vendor data flows are not publicly documented, so the broader governance claim stops at “state-stewarded cluster; vendor operation intra-site” rather than claiming a specific data-rights posture.

Data rights framework. state-stewarded — the cluster’s state-anchored franchise extends to the data-rights layer, though the per-vendor implementation is opaque.

Why it matters for talks

The Smart Farm Innovation Valley is the corpus’s anchor for “state-anchored agrifood cluster programme” patterns — a category distinct from both India’s federated DPI substrate (AgriStack) and China’s state-vendor provincial-autonomy hybrid. Korea’s pattern is state-anchored cluster-incubation: a national programme funding four provincial sites, each specialising in different crops, with FAO international recognition as the validation layer.

Two presentations:

1. The Korean state-anchored agrifood pattern. Surface alongside Japan’s industrial-automation vendor pattern (Layer A — Yamaha, Spread) to construct the “state-anchored cluster vs equipment-vendor concentration” binary that the meta-pattern unit (units/japan-korea-agrifood-ai-pattern.md) develops. This binary is one of the corpus’s clearest additions from this cycle.
2. The FAO Digital Villages Initiative as international validation. Korea being a named Digital Villages country is the corpus’s cleanest example of multilateral institution validation of an East-Asia agrifood AI cluster. The Sangju profile is the unit’s principal citation anchor.

Three secondary presentations:

• Young-farmer recruitment as a smart-farming policy frame. Korea’s 30%-by-2027 target is explicitly framed as a young-farmer recruitment mechanism. This is the inverse of Japan’s ageing-farmer framing — the corpus’s clearest “structural difference between two East-Asia clusters” observation.
• Site specialisation as a national-programme design pattern. Four sites, four crop specialisations, with each site driving a different export-market or R&D direction. This is materially different from India’s “AgriStack everywhere as foundation for many service layers” pattern.
• The Smart Farmland Distribution Center scaling. 14 → 26 → planned 100 (FFTC 2024) is a physical infrastructure expansion alongside the digital smart-farm cluster. The corpus’s only analog is Bayer / Climate FieldView’s dealership infrastructure in NA — same kind of physical meets digital playing out differently.

Critical context

• The 30%-by-2027 target is statutory, not measured. MAFRA announced the target; OECD 2025 cites it; no independent audit confirms current achievement against the target. V0 on the progress metric; V1 on the target statement itself.
• The trainee pipeline (104 / 81 / 208 / 168) is MAFRA-reported. FAO cites the figures from MAFRA materials; the percentage graduation outcome (81 / 104 = 78%) is meaningful but small-cohort.
• Cluster-coordination specifics are opaque. The Smart Farm Innovation Valley’s internal coordination across vendors, research institutions, and farmers is not publicly documented. The “jointly operated and managed by individual farmers, research institutions and private-sector businesses” framing (FAO) is the structural claim; the mechanism is opaque.
• The cluster’s international validation is Sangju-focused. Gimje, Milyang, and Goheung are named in FAO and FFTC materials, but Sangju is the only one with full Digital Villages Initiative profile treatment. The Sangju-as-Digital-Village recognition is the corpus’s anchor — other three sites have less multilateral validation surface.