Agentrys

Agentrys is an applied research company focused on advancing chip design through artificial intelligence. The firm is developing a platform known as Agentic Design Automation (ADA), which utilizes self-improving AI agents and a specialized, agent-native toolchain to move the semiconductor industry toward fully autonomous chip design. The company's stated goal is to address the increasing complexity and talent scarcity in the chip design sector by automating manual, labor-intensive tasks and shortening design cycles.

The company was founded by Mark Ren, who serves as CEO. Before establishing Agentrys, Ren was the director of design automation research at NVIDIA, where he led research efforts into AI for chip design. The founding team includes researchers and engineers with experience from prominent technology companies and research institutions such as NVIDIA, Meta FAIR, AMD/Xilinx, Samsung, Oracle, and IBM. This collective expertise spans frontier agent research and end-to-end chip design, from hardware-software co-design to physical design engines.

Agentrys' core product, Agentic Design Automation (ADA), is designed to overcome the limitations of traditional Electronic Design Automation (EDA) tools, which are built for human interaction. ADA employs AI agents that can work with existing EDA tools as well as new agent-native tools to autonomously manage design tasks. These agents are capable of setting goals, executing tools, interpreting results, and iterating until an objective is met, thereby reducing the manual effort required for analysis, optimization, and debugging. The system is built on three compounding learning loops: learning from existing design data, continuous refinement through expert feedback, and self-improvement from its own operational data. This approach allows the system to encode and amplify institutional knowledge with each use.

The business model appears to be centered on providing its ADA platform to design teams, EDA practitioners, and semiconductor organizations. Agentrys aims to help clients automate a range of engineering tasks, including RTL entry, verification, debugging, and performance optimization. By leveraging generative AI, agentic systems, and GPU-accelerated engines, the company seeks to deliver not just faster task execution but also measurably better design outcomes. The company is backed by undisclosed strategic investors and industry angels.

Keywords: Agentic Design Automation, ADA, autonomous chip design, semiconductor AI, self-improving agents, agent-native toolchain, Mark Ren, electronic design automation, hardware design, RTL generation, design verification, PPA optimization, generative AI in silicon, AI-driven EDA, autonomous systems, frontier AI research, deep learning, reinforcement learning