[For more diverse corporate information on startups, visit Unicorn Factory's Big Data platform 'Data Lab'.]
"In the past, manufacturing was dominated by small-variety mass production that failed to adequately reflect individual demand. Now, manufacturing must transform into high-variety, low-volume production and personalized manufacturing to survive."
Han Yo-han, CEO of Dimillion, said, "We are convinced that there is a huge market opportunity in the era of hyper-personalization where individuals pursue their own values, and we are focusing our efforts on solving the pain points of that market."
The company name 'Dimillion' is a compound of 'Data' and 'Millions.' Han Yo-han (CEO) explained, "It embodies our determination to supply smart manufacturing solutions based on tens of thousands to millions of pieces of data collected at manufacturing sites."
Dimillion's flagship solution is the industrial robot platform 'Flexibot.' It combines computer vision, robotic arms, on-device AI (artificial intelligence), and AMR (autonomous mobile robots) to optimize 'flexible production,' enabling fast and efficient execution of high-variety, low-volume manufacturing.
Flexibot analyzes its surrounding environment in real time and autonomously sets paths to perform tasks, actively responding to variables in the production process. It can be easily integrated into existing facilities or production lines, reducing initial investment costs and installation time while collaborating with human workers to simultaneously enhance productivity and safety.
Han (CEO) stated, "Traditional automation equipment was optimized only for producing specific products; when a product changed, the entire equipment had to be readjusted or replaced. In contrast, Flexibot allows robots to actively judge and assemble and inspect various products without separate teaching."
Combining OCR and LLM to Reduce On-site Knowledge Gaps
From a hardware perspective, there is the independently developed 'Flexible Feeder.' A feeder refers to a device that automatically supplies, aligns, and transports raw materials, components, or other materials to the next stage of the manufacturing process.
Traditional feeder devices could not be reused when product specifications changed, creating significant cost burdens. However, Dimillion's products can be reused even when components change, thanks to vibration and vision recognition technologies. This is said to yield substantial cost-saving effects in industries with diverse component types, such as cosmetics and medical devices.
Additionally, the company developed an 'AI Agent Worker' that utilizes OCR (Optical Character Recognition) and LLM (Large Language Model) technologies to automatically interpret unstructured work instructions exchanged via paper or fax and input them into MES (Manufacturing Execution Systems). This solution plays a major role in resolving knowledge gaps on-site and reducing quality variations.
Han (CEO) said, "We pursue technology that goes beyond simply controlling robots to understanding context. We are collecting comprehensive data from the perspective of '5M1E,' which adds Measurement (Measure) and Environment (Environment) to the four elements of production: Man, Machine, Material, and Method."
Dimillion focuses on 'standardization of data.' He explained, "Collected unstructured data is structured through the AAS (Asset Administration Shell) data standard. Only when data is standardized can higher-level AI systems function properly," he added.
Scaling Up Through Initial Funding Round
Dimillion currently collaborates with over 30 clients. Key industries include automotive, biotechnology, batteries, and semiconductors. Recently, the company conducted a PoC (Proof of Concept) related to AI agents with LG CNS and is also performing a PoC for an Indonesian factory operated by Jinus, a subsidiary of Hyundai Department Store Group.
Revenue has grown rapidly as well, reaching approximately 400 million won in 2024 and nearly 2 billion won in 2025. This year, the goal is to achieve 5 billion won. Han (CEO) stated, "We aim for 400% growth annually."
He analyzed, "Although the smart factory sector may appear as a red ocean with numerous competing companies, demand is explosive enough to surpass that red ocean. The challenges of skilled worker retirements and shortages of new personnel facing manufacturing are creating strong demand for AI solutions."
Dimillion has grown its business solely through internal revenue without significant external funding. To establish a clear growth trajectory this year, the company plans to secure initial investment in the first half of the year and has also been selected for TIPS (Technology Innovation Program for Startups), a government technology startup support program.
Han (CEO) said, "The funds raised will be used to establish and operate a 'Simulation Center' dedicated to continuously securing data from manufacturing sites. Through this center, we aim to acquire more data, advance our technologies, and provide greater trust to our customers."
Previously, Dimillion was selected for an initial startup package through the Gwangju Creative Economy Innovation Center and carried out the Startup Growth Technology Development Project 'Diddimdol (First Step),' receiving practical support for early growth. Last year, it also signed a memorandum of understanding with Gwangju Metropolitan City to create an AI semiconductor industry ecosystem.
Expanding from Vietnam as a Base to Conquer Asia and Reach North America
The company is also accelerating its expansion into global markets. Based on the branch office established in Ho Chi Minh City, Vietnam last year, it plans to expand local PoCs in Vietnam, Singapore, Indonesia, and India, and through these efforts, enter the North American market.
Han (CEO) stated, "Through a 'Mother Factory' strategy based on Korea's superior manufacturing competitiveness, we will disseminate success stories in a top-down manner." A Mother Factory refers to a plant that plays a central role among a manufacturer's global production bases and leads production.
As a new business initiative, the company is pursuing 'automated disassembly of waste batteries.' He said, "Currently, battery disassembly is done manually, which poses significant safety risks and has low efficiency. By automating this process, we aim to transform into a climate tech company that responds to the era's demands for carbon neutrality and ESG management."
He further added, "In the long term, we want to directly operate dark factories—fully automated facilities running 24/7 without human intervention—and create a structure where these factories themselves serve as massive showrooms and sources of data."
Dimillion's ultimate vision is to become a company dedicated to 'Innovation for Better Life.' Han (CEO) said, "We dream not of technology replacing people, but of a manufacturing environment where humans and technology grow and integrate together."
He emphasized, "We want to create a world where AI handles repetitive and dangerous tasks, allowing people to focus on more creative and valuable work. Contributing to sustainable industrial structures and better job environments is the value Dimillion pursues and its goal."
[MoneyToday startup media platform 'Unicorn Factory']
in FUTURE
