Agentrys Raises $24.5 Million to Transform Chip Design With Agentic AI
Agentrys, an emerging technology company focused on automating and continuously improving semiconductor design, has raised a total of $24.5 million in funding. The financing includes an oversubscribed $19.1 million seed round led by Etna Labs, following a $5.4 million pre-seed round led by MediaTek, which was Agentrys’ first strategic investor.
The new capital will support the company’s expansion across several areas, including hiring, development of agent-native engineering tools and deeper engagements with semiconductor customers. Agentrys plans to apply these resources to verification, physical design and other critical stages of the chip-development process.
The company is developing what it calls Agentic Design Automation (ADA), a new approach intended to move semiconductor engineering beyond traditional electronic design automation (EDA) tools. Instead of automating isolated tasks, Agentrys aims to create intelligent systems capable of managing broader engineering workflows, learning from previous experiences and improving their performance over time.
Addressing Challenges in Chip Design
Semiconductor design has become increasingly important as demand for advanced processors grows across artificial intelligence, data centers, automotive systems, consumer electronics and industrial applications. At the same time, developing modern chips remains a highly complex and time-consuming process.
Chip designers often depend on highly specialized engineers and thousands of hours of manual work. Engineers must manage numerous tools, analyze large volumes of data and repeatedly evaluate design decisions before a chip can move toward manufacturing.
Agentrys was founded to address this challenge by embedding AI-powered agents directly into engineering workflows. The company wants to enable semiconductor teams to create a digital workforce that can perform complex design activities while working within existing engineering environments.
The company was founded by Mark Ren, who brings nearly 30 years of experience in electronic design automation and artificial intelligence research. His career includes work at NVIDIA Research and IBM Research, where he led ChipNeMo, described by Agentrys as the first industrial large language model developed specifically for chip design.
Ren’s experience in AI and EDA forms a key part of Agentrys’ strategy to combine large-scale AI capabilities with the specialized requirements of semiconductor engineering.
Introducing Agentic Design Automation
Agentrys is positioning Agentic Design Automation as a new category within the semiconductor design industry. Traditional EDA software generally focuses on specific functions within the chip-design process. While these tools are essential, engineers still need to coordinate multiple tools, interpret results and make decisions across the broader workflow.
ADA aims to address this gap by introducing AI agents that can operate across different stages of the design process.
Agentrys envisions systems that can automate engineering tasks, learn from outcomes and continually improve their capabilities. The company’s longer-term objective is to create self-improving AI systems capable of handling increasingly sophisticated chip-design challenges.
This approach is particularly relevant to the semiconductor industry because engineering outcomes can be measured using objective tools and metrics. Design verification, performance targets and physical implementation results can all provide measurable feedback that AI systems can use to improve.
An Open Platform for Semiconductor Teams
Agentrys differentiates its platform from other emerging agentic EDA approaches in two key ways.
First, the company describes its technology as an open platform rather than a fixed collection of vendor-controlled AI agents. This allows customers to build and manage their own agentic engineering workforce according to their specific requirements.
The platform is designed to work across commercial EDA tools, internal engineering tools and existing infrastructure. This flexibility is intended to allow semiconductor companies to incorporate AI into their current development environments rather than completely replacing established systems.
Second, Agentrys is developing a design intelligence layer designed to learn continuously from customer data, usage patterns and evaluation results.
The company also combines agent-native tools with specialized AI models to extend what general-purpose AI systems can accomplish in semiconductor engineering.
The goal is to create a system in which each engineering workflow contributes new information that can be used to improve future performance.
Building Self-Improving Engineering Systems
According to Mark Ren, building reliable AI agents capable of performing real engineering work is a major technical challenge. Agentrys is focused on turning manual engineering processes into automated workflows that can deliver results more quickly while becoming increasingly capable over time.
The company sees this continuous improvement as an important difference between simply using a general-purpose AI service and building an AI capability that becomes a long-term organizational asset.
Etna Labs, which led the company’s seed financing, also sees chip design as a promising environment for developing self-improving AI systems. Because semiconductor engineering results can be evaluated through rigorous tools and measurable metrics, AI agents can receive objective feedback about whether their work meets required standards.
Etna Labs has invested in more than 20 AI companies and focuses on technologies involving foundation models, scientific applications, robotics and other advanced areas.
Support From MediaTek
MediaTek was Agentrys’ first strategic investor and led the company’s $5.4 million pre-seed round.
Brian Hsu, managing director of the MediaTek Innovation Fund, said semiconductor research and development has specialized requirements that make it different from many other applications of agentic AI.
MediaTek sees potential in Agentrys’ platform because it could allow semiconductor engineering organizations to customize and improve AI agents for their specific workflows. By capturing and reusing specialized engineering knowledge, companies could potentially scale AI development more systematically across semiconductor R&D operations.
The partnership also highlights growing interest among major chip companies in using AI to improve the efficiency and productivity of semiconductor development.
Demonstrating Autonomous Chip Design
Agentrys says it has already demonstrated its technology through an autonomous multi-agent workflow capable of taking a 32-bit CPU from its initial specification through to a sign-off-clean GDS layout without human intervention.
The company has also reported strong results on NVIDIA’s public CVDP verification benchmark, where it says its system was among the first to achieve more than 90% accuracy.
Agentrys uses iterative agent-evolution techniques to improve performance across successive evaluations. The company’s approach relies on feedback from engineering workflows to identify weaknesses and improve future agent performance.
The company is also working with several major fabless semiconductor companies, a leading global foundry and multiple chip startups. Its current customer engagements span digital and analog design workflows, while system design capabilities are planned for the future.
Expanding Across the Design Flow
The latest funding will allow Agentrys to expand its technology across a wider portion of the semiconductor development lifecycle.
The company believes each execution of an engineering workflow generates useful observations. Those observations can then be evaluated against engineering requirements, creating feedback that can improve the underlying AI system.
This creates a cycle in which the platform becomes more capable as it handles additional engineering tasks.
Agentrys plans to use its new funding to extend this capability across the full chip-design flow and make its technology available to a larger number of engineering teams.
The company’s vision is ultimately to provide semiconductor organizations with an AI-powered engineering workforce that can operate alongside existing tools, learn from real-world design experience and continuously improve.
Building the Future of Chip Design
Agentrys’ $24.5 million financing comes as semiconductor companies face increasing pressure to develop more complex chips while managing development costs and engineering resources.
By combining AI agents, specialized models, engineering tools and continuous learning, the company aims to address some of the most time-intensive challenges in modern chip development.
With a team that includes researchers and engineers with backgrounds at NVIDIA, Meta, AMD, Samsung, Google, Siemens and other major technology companies, Agentrys brings together expertise in AI, EDA, large language models and silicon infrastructure.
Headquartered in San Jose with offices in Austin and Taiwan, the company is now positioned to expand its technology and customer base. Its goal is to transform chip design from a collection of manually coordinated engineering tasks into an intelligent, adaptive workflow that becomes more efficient and capable with every design.
If successful, Agentrys’ Agentic Design Automation approach could help semiconductor companies accelerate development, capture specialized engineering knowledge and tackle increasingly complex chip designs with greater speed and efficiency.
Source Link: https://www.businesswire.com/