YATAV
Category: AI in Cybersecurity
YATAV is an AI safety and trust verification platform provider, offering enterprise-grade AI security, reliability verification, and work automation solutions for regulated industries. YATAV was founded in 2018. The company is led by Lee Seong-chan (이성찬) & Jo Eun-sang (조은상). Based in Seoul, South Korea. Team size: 11-50. Total funding raised: $600K. Latest round: Seed. Key investors include MYSC (엠와이소셜컴퍼니), Korea Social Investment (한국사회투자).
- Founded
- 2018
- Headquarters
- Seoul, South Korea
- Team size
- 11-50
- Total funding
- $600K
Value proposition
Enterprise AI reliability platform that combines multi-layered security (AEGIS), hallucination prevention, truth verification (TruthAnchor), and work automation (HomiClaw) into a unified pipeline — enabling safe and trustworthy enterprise AI deployment in regulated environments.
Products and solutions
AEGIS AI Security (enterprise AI security with 6-layer defense), TruthAnchor AI Reliability (4-layer hallucination prevention pipeline), HalluGuard (real-time hallucination detection), HomiClaw (autonomous AI work agent), HAECHI Security Platform v3.0 (unified security scanner for LLM/web vulnerabilities), AEGIS VLA (Physical AI safety for autonomous robots), AEGIS Personal (browser extension for individual AI safety), OpenHangulAI (HWPX document AI editor), AEGIS Essential (government/enterprise AI security browser extension)
Unique value
Defense-first AI infrastructure: 6-layer PALADIN system blocking 215+ attack patterns in under 2ms with Rust-based engine, combined with 4-layer truth verification pipeline for hallucination-free enterprise AI operations.
Target customer
Enterprise and government organizations in regulated industries including finance, defense, public administration, and manufacturing requiring safe AI deployment
Industries served
Finance, Defense/Security, Public Sector/Government, Manufacturing/Robotics, Healthcare
Technology advantage
6-layer PALADIN multi-defense system blocking 215+ attack patterns; Rust-based engine processing 50,000+ requests/sec at avg. 2ms; 4-layer TruthAnchor verification pipeline; Apache 2.0 open-source AEGIS DreamDojo Guard SDK for robot safety; World-model specific physical AI safety; 11 patents
How they differentiate
End-to-end AI safety pipeline covering defense, verification, and execution in a single platform — unlike fragmented competitors. Open-sourced AEGIS robot safety SDK (Apache 2.0). Specializes in both cyber AI security (prompt injection, data leakage) AND physical AI safety (robotics world-model attacks). Holds 11 patents.
Main competitors
Prompt Security (AI security platform), Palo Alto Networks Agentic AI Security, Protect AI (AI security for ML pipelines)
Key partnerships
SK Telecom (Glomo Award collaboration), NVIDIA (AEGIS DreamDojo integration), CMC Corporation (AI security collaboration discussions), Hunter College NYC (metaverse counseling seminar)
Notable customers
Public sector clients via B2G contracts (26 government procurement contracts), finance and defense sectors (closed network environments per VivaTech 2026 coverage)
Major milestones
2018-07: Company founded, 2024-12: Selected for TIPS R&D program and raised Seed round from MYSC & Korea Social Investment (₩800M total), 2025-01: Exhibited at CES 2025, 2025-02: Exhibited at MWC Barcelona 2025 (won Glomo Award with SKT), 2025-09: Selected for Micro DIPS (초격차 스타트업 1000+), 2026-02: MWC 2026 — launched AEGIS DreamDojo Guard SDK for robot safety (open-source Apache 2.0), 2026-03: Launched HomiClaw AI work agent, 2026-04: Launched AEGIS Mobile and HAECHI v3.0, 2026-06: Showcased AI trust verification platform at VivaTech 2026
Growth metrics
Team grew from startup to 16 employees (2026); cumulative government project contracts worth ₩2B+; 26 B2G contracts
Market positioning
Early-stage but fast-growing Korean AI security startup differentiating on comprehensive safety pipeline (defense + verification + automation) vs. point-solution competitors. Selected for multiple Korean government programs (TIPS, Global Ddiddol, Micro DIPS). Expanded from psychological counseling metaverse to enterprise AI security infrastructure since 2024-2025.
Geographic focus
South Korea (home market), expanding to Europe (VivaTech 2026, MWC Barcelona), US (CES)
Patents and IP
11 patents registered (as of Oct 2025); PALADIN architecture patent; Apache 2.0 open-source AEGIS DreamDojo Guard SDK released on GitHub
About Lee Seong-chan (이성찬) & Jo Eun-sang (조은상)
Co-CEO Lee Seong-chan: Founded YATAV in 2018; Counseling Psychology Master's degree; received Ministry of SMEs & Startups and Ministry of Health & Welfare Minister's Awards; 5 patents; TIPS selection. Co-CEO Jo Eun-sang: Co-founder since 2018; Coaching Psychology Master's degree; expertise in high-performance team building; cumulative ₩2B+ government project execution.
More AI in Cybersecurity companies
Official website: https://www.yatavent.com