Daedong AI Lab — Three-Core-AI-Domains strategy for Korean agriculture (Mobility / Operations / Cultivation) + L4 autonomous tractor released April 2026

East-Asia (Korea; Republic of Korea national anchor)

Daedong AI Lab — Three-Core-AI-Domains strategy for Korean agriculture

Content

Daedong Group (Republic of Korea) operates through its AI-robotics software subsidiary Daedong AI Lab (CEO Choi Joon-gi; founded 2024; headquartered at the Daedong Seoul Office in Seocho-gu, Seoul). The group’s “Kioti” tractor brand is internationally marketed in the United States and other markets.

On July 17, 2025, Daedong AI Lab held a media briefing at the Daedong Seoul Office to present the company’s development strategy for three core agricultural AI domains:

The stated goal is to lead Korea’s agricultural sector through a full-scale “AI transformation” — “AI-driven machinery and robots can make independent decisions to carry out fieldwork, while cultivation AI can predict crop growth and recommend optimal strategies.”

Key products and dates:

• RT-100 transport robot — launched February 2025 — “Korea’s first government-certified autonomous transport robot for agricultural applications.” Fully electric, zero-emission. (Korean government certification per the Korea Institute of Agricultural Technology Promotion and MAFRA.)
• Flagship tractor with Level 4 autonomy — released April 2026. HX1400-AI (142 hp) and HX1200 AI (127 hp); six-camera vision AI plus precision positioning; “A-Motion Easy Mode” route setup from A-B straight lines, curved paths or azimuth bearings. Designated a new-technology agricultural machine by the Rural Development Administration on 1 September 2026 for its vision-AI-based Level 4 autonomous tractor platform (vision-AI recognition model + app-based remote control of autonomous work). The designation carries priority purchase by local governments and public agencies, subsidy and loan support, and preferential public procurement. Daedong frames it as Korea’s first commercial L4 autonomous-work tractor.
• RT-100 AI-Powered Voice Edition — announced subsequent (Korean-language AI interface).

Doosan Robotics + Daedong partnership (Oct 2025). Daedong and Doosan Robotics partnered to design and manufacture an autonomous mobile platform that uses agricultural field data and integrates field operations — combining Doosan’s robotic-arm engineering with Daedong’s agricultural-machinery heritage.

Saemangeum Agro-Bio Complex demonstration H2 2025. Daedong AI Lab planned precision-farming demonstrations at Saemangeum (Seoul-Jeollabuk-do reclaimed land designated for high-tech agriculture) — satellite data collection, drone swarm flights, smart attachments. Government support via the “Agricultural Robot Demonstration Support Project” led by the Korea Institute of Agricultural Technology Promotion and MAFRA’s precision-farming policy at Saemangeum.

Release confirmation and market follow-through (April-October 2026). The July 2026 retrieval had the release as imminent; it has since happened and been followed by three datable events:

• April 2026 — domestic launch. Since release the machine has been used by professional farmers and agricultural corporations for paddy and dry-field plowing, rotary tilling and ridge making. Daedong’s stated effect is reduced operator fatigue and reduced dependence on skilled workers — the labour-substitution reading is the vendor’s own, not an independent finding.
• August 2026 — first OTA update, adding autonomous-work convenience features. Daedong frames the product line as a Software Defined Tractor (SDT) running an MLOps loop: field data captured after launch feeds training and validation, improved models ship over the air, no hardware replacement.
• 1 September 2026 — RDA new-technology agricultural machine designation (see Key products above). Daedong’s development-division head framed the intent as moving Korean farm operations from one that depended on the experience and concentration of skilled workers to an AI-centred one — a direct statement of the substitution claim, from the vendor side.
• 6 October 2026 — free customer test-ride programme at 29 dealerships, running to year-end, for both HX models. Daedong’s domestic business head Choi Hyung-woo said customer criteria are shifting from mechanical performance to real-world work efficiency and how well the technology can be put to use. The programme exists because purchase decisions on autonomous equipment are not settled by spec sheets — evidence of an early, contested market rather than a settled one.

None of this is accompanied by unit sales, fleet counts or named customers. Six months after launch the volume question is still unanswered.

What this unit is doing in the taxonomy

Sector position. The three-core-domains strategy covers on-farm-production-open-field (Mobility AI for tractors), on-farm-production-protected (Cultivation AI for CEA-contexts), and animal-production (Korea is a notable livestock-density context; Daedong’s strategic range extends across sectors).

AI technique class. Multiple:

• computer-vision (vision-based autonomous driving is the Mobility AI core; the 500K orchard / 3M video dataset)
• predictive-ml (cultivation AI for crop growth prediction)
• generative-ai-llms (AI Daedongi chatbot + Agentic AI extension)
• robotics-autonomy-ground (L4 autonomous tractor)

Purpose. Multi-purpose:

• yield-optimisation (cultivation AI for crop growth)
• worker-conditions (autonomous tractor reduces operator labour; L4 vision-based)
• input-reduction (precision-farming context: pinpoint application)
• market-access (international market through the Kioti brand; AI-powered products are Daedong’s stated bid for global market position)

Region tag. East-Asia (Korea; Republic of Korea national anchor) — Daedong’s deployment geography at retrieval date is Korea-anchored; the international market reach is via the Kioti brand and pre-existing Daedong dealers, not via the new AI portfolio.

Capital intensity. industrial.

Data governance / rights. proprietary / vendor-owned — Daedong’s training data is the company’s competitive infrastructure.

Maturity grade. activity-status: deployed (commercial release, in field use); maturity-scale: S1 and maturity-verification: V0 held. The July 2026 scan wrote the promotion rule as “confirm at release, then promote S1 to S2 with V1 vendor-customer naming.” The release half is now satisfied; the naming half is not — Daedong publishes no unit volumes and names no customers, so S2 (hundreds to low thousands) and V1 (named customers) are not earned by the evidence. Re-tag when sales or customer names appear in a primary source.

Why it matters for talks

Daedong AI Lab is the corpus’s anchor for the “Korean agricultural-machinery vendor pivoting to AI” pattern — a distinct actor-type from both:

• NA equipment vendors (e.g. John Deere’s recent Blue River / See & Spray acquisition; units/dji-agriculture-global-export.md is China-side aerial counterpart; Kubota Agri Concept 2.0 is Japan-side tractor counterpart).
• Korean state-anchored cluster (units/korea-smart-farm-innovation-valley-rda.md — the structural counterpoint to Daedong’s vendor-led strategy).

Two presentations:

1. The vendor-state parallel in Korea. Korea’s agrifood AI cluster has both a strong state-anchored cluster programme (Sangju et al.) and a strong vendor-led AI roadmap (Daedong). The pattern is materially different from Japan (where state plays infrastructure role, vendors play product role) and from India (where state-DPI is the substrate, vendors work above it). Daedong is the canonical “vendor-side AI roadmap” in Korea.

2. Vision-based vs GPS-based autonomy. Daedong’s Mobility AI explicitly positions vision-based autonomy as superior to GPS-based autonomy for agricultural contexts — “because conventional GPS-based autonomy has clear limitations when it comes to recognising field boundaries or responding to obstacles.” This is a substantive technical claim. Surface in any talk about agricultural autonomy / robotics.

Three secondary presentations:

• Daedong AI Lab as a corporate-strategy pivot. Daedong is an agricultural-machinery major (established tractor brand) extending into AI software — a corporate-strategy move that mirrors the NA “industrial conglomerate pivots to data-and-AI” pattern. Surface the corporate-strategy reading.
• The Saemangeum Agro-Bio Complex as a Korean state-vendor integration context. Daedong operating the precision-farming demonstration at a state-reclaimed-land site is a state-vendor integration pattern. Surface this observation.
• Doosan Robotics + Daedong. Cross-sector robotics partnership (October 2025) — Doosan is best-known for industrial / collaborative robot arms; Daedong for agriculture. The partnership is the corpus’s first explicit agricultural-industrial-robotics partnership in Korea.

Critical context

• H1 2026 L4 tractor release — confirmed, trigger closed (2026-10). Released in Korea April 2026; confirmed in Daedong’s own 1 September 2026 company release and in English-language trade coverage. The refresh obligation written into this unit in July has been met.
• Training-data figures are vendor-stated and inconsistent. ~500K images + 3M driving-video samples (July 2025 briefing) against “5 million-plus images” secured during development (Sept 2026 release). The two are not reconciled anywhere in the primary sources; treat both as self-statement and do not cite either as a verified count.
• Deployment volume and customer identity remain the open question (G-080). Six months post-launch, no sales figure, fleet count or named customer appears in any source — only the category descriptions “professional farmers and agricultural corporations”.
• “Korea’s largest agricultural dataset” is Daedong self-statement. V0 vendor-reported; no third-party comparison.
• RT-100 deployment volume is V0. February 2025 launch confirmed; deployment numbers not in primary sources.
• The three-AI-domains framework is corporate strategy, not verified deployment. Each “domain” is in a different stage of operational readiness; the framework is real (it’s how Daedong is structured), but the deployment-readiness of each component varies.
• Daedong operates in a regulatory frame. Korea’s Agricultural Robot Demonstration Support Project and MAFRA’s Saemangeum precision-farming policy are the institutional pathways. The unit’s T2-T3 translation grade is anchored in these institutional mechanisms.