AiV Launches AiVision, a Manufacturing-Specialized Vision Foundation Model (VFM) Inspection Solution

AiVision addresses the manufacturing industry's data shortage by generating and learning from defect data using only normal data.

AiV, a Physical AI and AI vision company accelerating autonomous manufacturing, today announced the official launch of AiVision, a Vision Foundation Model (VFM)-based inspection solution purpose-built for manufacturing environments.

As manufacturers increasingly adopt AI, one of the industry's greatest challenges remains the lack of high-quality training data. Gartner predicts that by 2026, 60% of AI projects will be abandoned due to insufficient data readiness, highlighting the difficulties manufacturers face in collecting and managing production data.

Traditional AI inspection systems require large volumes of defect images along with extensive metadata—including machine IDs, timestamps, and batch numbers—to train reliable models. In addition, conventional supervised learning approaches often struggle to detect previously unseen defects, while the manual collection and labeling of defect data remains costly and time-consuming.

AiVision addresses these challenges by fundamentally changing how industrial AI models are trained.

Built upon AiV's proprietary unsupervised learning technology, AiVision enables AI models to learn exclusively from normal (non-defective) products while automatically generating synthetic defect data for training. As a result, manufacturers can deploy highly accurate inspection models without the need to collect and label large quantities of defect images.

The solution incorporates AiV's proprietary technologies, including:

• Anomaly Detection capable of identifying previously unseen defects

• AI-powered Defect Generation for automatically creating synthetic defect data

• Unsupervised learning algorithms that dramatically reduce dependency on labeled datasets

These innovations allow manufacturers to respond more effectively to rare or newly emerging defects while significantly reducing AI deployment costs.

The underlying algorithms have demonstrated state-of-the-art (SOTA) performance on internationally recognized benchmarks including OpenOOD and Papers with Code, validating AiV's technological leadership in industrial AI.

With the launch of AiVision, AiV has become the first company in Korea to provide both a manufacturing-specialized deep learning inspection solution and an industrial MLOps platform, AiVOps, as a fully integrated offering.

Together, AiVision and AiVOps support the complete AI lifecycle—from data management and model training to deployment, monitoring, and continuous operation. This tightly integrated ecosystem enables manufacturers to rapidly adapt AI models to production line changes and new product introductions while improving both inspection efficiency and scalability.

By minimizing human intervention throughout data generation, data collection, model development, and AI operations, AiVision delivers a truly end-to-end deep learning inspection system. The solution helps manufacturers overcome growing labor costs, aging production facilities, and productivity challenges while strengthening long-term manufacturing competitiveness.

Sung Minsoo, CEO of AiV, commented:

"One of the biggest barriers preventing manufacturers from adopting AI has been the inherent limitations of applying general-purpose AI models to highly customized factory environments. AiVision addresses this challenge by commercializing Korea's first manufacturing-specialized unsupervised AI model tightly integrated with an industrial AI platform, enabling both exceptional real-world usability and operational efficiency."

Since its founding, AiV has raised a total of KRW 37 billion in funding from leading investors including LB Investment, Kiwoom Investment, and Samsung Securities.

The company is currently delivering AI projects with global manufacturers such as Tenneco, HD Hyundai Samho, Hankook Tire, and Poongsan across the automotive, shipbuilding, secondary battery, and defense industries.

AiV has also filed and registered more than 50 patents related to industrial AI technologies and continues to advance its vision of achieving fully autonomous quality inspection while expanding its presence in global manufacturing markets.