Skip to main content

Fobecon

Category: AI Infrastructure

An AI-powered Con-Tech (Construction Technology) solution that automates the labor-intensive process of quantity surveying (Jeoksan) by analyzing construction drawings using Vision AI. Fobecon was founded in 2024. The company is led by Song Joong-seok. Based in Seongnam, South Korea. Team size: 10-20. Total funding raised: $1,100,000+. Latest round: Pre-A (Undisclosed, Jan 2026). Key investors include ["Xplor Investment (GS E&C CVC)","Magellan Technical Investment","Mark & Company","Yoonmin Creative Investment Foundation"].

Founded
2024
Headquarters
Seongnam, South Korea
Team size
10-20
Total funding
$1,100,000+

Value proposition

Reduces quantity surveying time from 10 hours to just 20 minutes (a 30x productivity boost) and lowers associated costs by approximately 75% while minimizing human error in material estimation.

Products and solutions

["Auto-Jeoksan 1.0 (Flagship Solution)","Window & Door Recognition Module","Indoor Space Recognition Module (Living rooms, Balconies, etc.)","Wet/Dry/Structural Wall Classification Module","Net Quantity (Jeongmi-suryang) Calculation Engine"]

Unique value

The only solution in the domestic market that combines Vision AI with high-performance GPGPU computing to extract 'net quantity' data directly from 2D CAD and PDF drawings without requiring complex 3D BIM modeling.

Target customer

Mid-to-large construction companies (General Contractors), specialized subcontractors, interior design firms, and architectural planning offices.

Industries served

["Construction Technology (Con-Tech)","Architecture & Engineering","Interior Design","Real Estate Development"]

Technology advantage

Leverages proprietary Computer Vision algorithms and CUDA-based image processing to achieve rapid segmentation and classification of structural elements (walls, openings, spaces), specifically optimized for the nuances of the Korean construction sector.

How they differentiate

Fobecon differentiates by utilizing Vision AI to extract 'net quantity' data directly from 2D CAD and PDF drawings, eliminating the need for complex 3D BIM (Building Information Modeling) which competitors often require. Its proprietary GPGPU-accelerated algorithms reduce quantity surveying time by 30x (from 10 hours to 20 minutes) while specifically targeting high-labor finishing works like masonry, tiling, and plastering.

Main competitors

["Changsoft I&I (The Billion)","TenEleven (Space AI)","Angelswing"]

Key partnerships

["Xplor Investment (Corporate Venture Capital arm of GS E&C)","Magellan Technical Investment","Mark & Company","Yoonmin Creative Investment Foundation","Seongnam City (Grand Prize winner of 2025 S-Startup Contest)"]

Notable customers

["GS E&C (Strategic Partner)","Gijun Construction (MOU Partner)"]

Major milestones

["Secured Pre-A funding from GS E&C's CVC (Xplor Investment) in Jan 2026","Won the Grand Prize at the 2025 Seongnam Startup Contest (S-Startup)","Selected for the 17th Shinbo Startup Nest program in 2025","Selected for the Ministry of SMEs and Startups 'TIPS' program in Aug 2024","Secured Seed investment and 'Little Penguin' status from KODIT in July 2024"]

Growth metrics

Productivity boost of 30x in quantity surveying; 75% reduction in material estimation costs; selected for the prestigious TIPS program and Shinbo 'Little Penguin'.

Market positioning

Specialized AI Con-Tech disruptor for automated material estimation

Geographic focus

South Korea (Primary focus on domestic construction nuances and CAD standards)

Patents and IP

Proprietary Vision AI algorithms for construction element recognition (Specific registration numbers not publicly disclosed)

About Song Joong-seok

Song Joong-seok holds a Ph.D. in Computer Software Engineering from Hanyang University and has over 17 years of experience in Computer Vision, GPGPU (CUDA), and Vision AI R&D. Before founding Fobecon, he established a strong reputation for commercializing Vision AI across various industrial sectors, focusing on high-performance computing and image processing.

Official website: