Tynapse Raises $3.17M Seed Round to Build Runtime Security Infrastructure for AI Agents

Tynapse, a Korean startup developing trust and reliability infrastructure for AI agents, has closed a seed funding round of approximately USD 3.17M (KRW 4.5 billion), bringing its total raised to USD 3.45M (KRW 4.9 billion) just six months after team formation.

The seed round was led by Mirae Asset Venture Investment, with participation from Mirae Asset Capital, Murex Partners, and Kakao Ventures. The raise follows an earlier USD 141K (KRW 200 million) angel round from serial unicorn founders and leading AI researchers, and a pre-seed investment from Mashup Ventures secured within three months of founding.

Tynapse is building what it calls an “AI Trust Layer” — a runtime security infrastructure that validates and controls the outputs and actions of AI agents at the moment of execution. The system uses a two-stage detection and judgment architecture to identify and block risks in real time, addressing threats such as hallucinations, data leakage, jailbreaking, and unauthorized privilege escalation. Every decision made by the system is automatically logged as a legally valid audit trail, enabling clear accountability and traceability in the event of an incident. The company is currently conducting proof-of-concept trials with major domestic banks ahead of commercial deployment.

As AI agents move from experimentation into live enterprise workflows, operational risks — including hallucinations, factual inconsistencies, data exposure, and privilege misuse — have emerged as the primary barrier to enterprise adoption. Tynapse plans to expand from financial services into healthcare, the public sector, and broader enterprise markets, positioning itself as the global category leader in AI security infrastructure.

The founding team is led by Minseung Kang, a former CTO in the financial sector with research and execution experience in AI trust and safety. The team also includes AI specialists from Google Research and engineers with large-scale traffic operations backgrounds. Tynapse has validated its technology on the global stage, reaching the top finalist ranks — as the only Asian entrant — at both the NVIDIA GTC 2026 AWS Startup Pitch and the 2026 Snowflake Startup Challenge.

“Tynapse is a team that combines deep hands-on experience in finance and security with world-class AI capabilities,” said Jinhwan Cho, Director at Mirae Asset Venture Investment. “We invested because we believe they can actually solve the AI reliability problem.”

“As AI moves into the execution layer, trust infrastructure has become a must-have category that determines the scalability of the entire industry,” said Jungho Shin, Senior Associate at Kakao Ventures. “We believe Tynapse will be the winner in this category.”

“In the age of AI agents, operational reliability and regulatory compliance matter just as much as model performance,” said Minseung Kang, CEO of Tynapse. “Our goal is to grow into a global standard infrastructure company — one that makes AI trustworthy, no matter where or what it executes.”