Thursday, August 6
Business · Technology · Leadership

Agentrys Raises $24.5M for AI-Powered Chip Design

Agentrys Raises $24.5M for AI-Powered Chip Design
Photo Credit: Unsplash.com

San Jose-based Agentrys has raised $24.5 million to develop AI agents for semiconductor design, combining a $19.1 million seed round with a $5.4 million pre-seed round. The startup plans to use the funding for hiring, agent-native tools and customer engagements involving chip verification and physical design.

Key Takeaways

  • Agentrys announced $24.5 million in combined seed and pre-seed funding.
  • Etna Labs led the $19.1 million seed round.
  • MediaTek led the $5.4 million pre-seed round.
  • The startup is developing AI agents for semiconductor design workflows.
  • Funding will support hiring, tooling and additional customer engagements.

Agentrys Secures $24.5 Million in Combined Funding

Agentrys announced on August 26 that it had raised $24.5 million across a seed round and an earlier pre-seed round. The financing consists of an oversubscribed $19.1 million seed round led by Etna Labs and a $5.4 million pre-seed round led by MediaTek.

The San Jose-based startup was founded by Mark Ren, who has experience in electronic design automation and artificial intelligence research and development. Ren’s background includes work at NVIDIA Research and IBM Research, as well as leadership of ChipNeMo, an industrial large language model project for chip design.

The new capital is earmarked for recruiting, development of agent-native tooling and expanded customer engagements. Agentrys said those customer engagements will cover semiconductor verification and physical design.

The funding gives Agentrys additional resources to develop its platform while expanding its work with semiconductor engineering teams. The company’s stated focus is automating and continuously improving chip-design workflows through AI agents.

Seed and Pre-Seed Funding Structure

Etna Labs led the larger seed financing with $19.1 million. MediaTek led the earlier $5.4 million pre-seed round and was identified as Agentrys’ first strategic backer.

The two rounds bring the company’s disclosed financing to $24.5 million. The funding structure combines financial backing from Etna Labs with strategic participation from a major semiconductor company.

The financing approach has similarities with other recent AI startups that have raised capital around specialized business applications, including AI infrastructure startup funding.

AI Agents Target Semiconductor Design Workflows

Agentrys is developing what it calls Agentic Design Automation, or ADA, for semiconductor engineering. The platform is designed to let engineering teams build and use AI agents within their existing chip-design workflows rather than relying only on fixed automation for individual tasks.

The company’s platform uses AI agents across semiconductor design processes and is intended to work with commercial electronic design automation software, proprietary engineering tools and existing computing infrastructure. Agentrys describes the system as an open platform that allows engineering teams to create agents suited to their internal workflows.

The system also includes a design intelligence layer that uses customer data, platform usage and evaluation signals. Agentrys said specialized tools and custom models are used alongside general-purpose AI agents for semiconductor engineering tasks.

The approach is centered on evaluating the results produced during engineering workflows. Chip-design tasks can generate measurable outputs that can be assessed against engineering tools and performance metrics. Agentrys uses those evaluations as feedback for its agents.

The company’s stated objective is to enable semiconductor engineering teams to build and maintain AI agents that can improve through repeated workflow execution and evaluation.

The focus on specialized AI applications also parallels other startup efforts to build artificial intelligence products around defined business workflows, such as enterprise AI software platforms.

Funding Supports Hiring and Agent-Native Tools

Agentrys plans to use the new funding to recruit employees and develop additional tools designed specifically for AI agents working on chip-design tasks. The company also plans to expand its customer engagements in verification and physical design.

Agentrys Raises $24.5M for AI-Powered Chip Design

Photo Credit: Unsplash.com

The hiring effort follows the company’s development of a team with experience across AI, electronic design automation and semiconductor engineering. Agentrys said its team includes researchers and engineers with backgrounds at companies including NVIDIA, Meta, AMD, Samsung, Google and Siemens.

Agent-native tooling is a separate focus of the financing. Agentrys is building tools intended to extend the capabilities of AI agents within semiconductor engineering workflows rather than treating the agents as standalone general-purpose software.

The company also plans to put its platform into more customer workflows. Its current engagements cover full digital and analog design flows, while system design is included on its roadmap.

Agentrys said each workflow execution can produce engineering observations that are evaluated and used as feedback for subsequent system performance. That process forms part of the company’s approach to improving its agents through repeated engineering tasks.

Customer Engagements Focus on Chip Verification and Physical Design

Agentrys said it is already working with several major fabless semiconductor companies, a leading global foundry and multiple chip startups. Its customer work spans digital and analog design processes.

Chip verification is one of the areas targeted by the platform. Verification involves checking semiconductor designs against defined requirements before production, making it a specific engineering workflow for AI-assisted automation. Agentrys is also targeting physical design, another stage of the chip-development process.

The company has reported an autonomous multi-agent workflow that took a 32-bit CPU from specification through a sign-off-clean GDS layout without human intervention. It also said its system exceeded 90% accuracy on NVIDIA’s public CVDP verification benchmark.

Agentrys reported that performance on the benchmark improved through successive iterations using its agent-evolution techniques. The company said the results are part of its work to apply AI agents to measurable semiconductor engineering tasks.

The reported workflow and benchmark results provide specific examples of the technical applications Agentrys is pursuing. The company is applying its platform to tasks where engineering outputs can be evaluated against established tools and metrics.

Agentrys Expands Its Platform for Chipmakers

Agentrys is headquartered in San Jose and has offices in Austin and Taiwan. The company’s leadership and engineering team includes experience in AI for chip design, electronic design automation, large language models and silicon infrastructure.

Ren founded the company after nearly three decades of work spanning electronic design automation and AI research and development. His previous work included NVIDIA Research and IBM Research, along with leadership of ChipNeMo.

MediaTek’s participation in the earlier financing gives Agentrys a strategic connection to a semiconductor company. MediaTek’s managing director of its Innovation Fund said the startup’s approach is designed to let semiconductor engineering teams customize and improve AI agents for chip-design workflows.

Agentrys plans to extend its technology across more stages of semiconductor design using the new financing. The company’s roadmap includes system design in addition to its current work across digital and analog design flows.

The company said its platform is designed to allow engineering teams to create agents tailored to their own workflows and operate those agents across existing engineering infrastructure. The new funding will support the development of that platform and its expansion to additional customer engagements.

Frequently Asked Questions

How much funding did Agentrys raise?

Agentrys raised $24.5 million through a $19.1 million seed round and a $5.4 million pre-seed round. Etna Labs led the seed financing, while MediaTek led the pre-seed round.

What does Agentrys use AI agents for?

Agentrys is developing AI agents for semiconductor engineering workflows, including chip verification and physical design. Its platform is designed to operate across existing electronic design automation software and other engineering tools.

Who led Agentrys’ funding rounds?

Etna Labs led Agentrys’ $19.1 million seed round. MediaTek led the earlier $5.4 million pre-seed round and was the company’s first strategic backer.

What semiconductor workflows is Agentrys targeting?

Agentrys is working across digital and analog design flows, with current customer engagements covering verification and physical design. System design is also included on the company’s roadmap.

How will Agentrys use its new funding?

Agentrys said the funding will support recruiting, agent-native tooling and expanded customer engagements across verification and physical design.

Kivo Daily

Your source for thought-provoking articles, personal development, and success stories.

latest posts kivo daily