Copyright SiliconANGLE News
Santa Barbara-based ChipAgents.ai today announced it raised $21M in early funding to fuel growth for its agentic artificial intelligence platform for chip design and verification. Bessemer Venture Partners led the Series A round, with strategic backing from Micron Technology Inc., MediaTek Inc., Ericsson and additional semiconductor giants. Existing investors ScOp Venture Capital and Amino Capital also participated in the round. The round brings the total funding to date for the company to $24 million. Professor William Wang, founder and chief executive of ChipAgents, told SiliconANGLE in an interview that formal chip design began in the 1980s, when electrical engineers still worked on paper to develop gate architectures. But the advent of AI chips and new system-on-chip designs has pushed the number of logic gates to staggering new levels of complexity. “Chips became so complex that it used to be millions of logic gates, and right now it’s billions, or tens of billions, even trillions of logic gates,” Wang said. “So, when you have to design complexity at this particular level, nobody understands the code anymore.” Engineers now use sophisticated tools to visualize and construct chip architectures, but legacy software is falling behind. To address this challenge, ChipAgents has developed an agentic AI platform that automates routine design and verification tasks, which permits engineers to focus on innovation rather than tedious, line-by-line code work. ChipAgents operates on the front end of the chip design and verification process, before fabrication begins. Engineers use it from the earliest stages when they’re defining chip specifications, which can produce PDFs and data sheets hundreds of pages long. Wang pointed to companies such as Cadence Design Systems Inc. and Synopsys Inc. as early pioneers of simulation-based chip design and electronic design automation, or EDA — processes that ChipAgents now aims to augment with agentic AI-driven tools. Using AI, the platform can assist engineers in identifying inconsistencies or flaws by cross-checking specs across multiple documents, generate Register-Transfer Level code and automatically produce documentation. RTL is a hardware design abstraction that describes how digital circuits transfer data and how logic is executed at the hardware level, typically written in hardware description languages such as Verilog or VHDL. Wang said using the generative AI capabilities of ChipAgents, teams across architecture, design and verification can use the tool to streamline workflows and shorten time to market. One particular use case that dominates the company’s platform is verification. Unlike in software engineering, where if there’s a flaw a developer can just fix the bug on the fly, hardware verification demands near-perfection before a chip goes into production because mistakes can cost millions. “In the industry, people spend more time verifying the functional correctness of the chip, because it’s easy to write code, but how do you know that the code you’ve written is actually correct?” Wang noted. He added that for every design engineer, there are often two to three verification engineers. ChipAgents makes their work faster by automatically generating test benches, rules and assertions — tasks that could take weeks but now can be done in minutes. This helps teams quickly confirm that a chip’s implementation matches its specifications. Although Wang couldn’t name customers, he said many of the top 20 semiconductor firms are using ChipAgents. Founded in 2024, the company has seen explosive growth, usage has surged 60 times in the first half of 2025, off an undisclosed base. “We’re seeing a lot of usage every day and new use cases each day,” Wang said. “I think we’re witnessing the transformation of the semiconductor industry into agentic AI solutions for design verification.” With the fresh funding, ChipAgents plans to build on its research and development and focus on additional customer support. This will help improve on-the-ground support for semiconductor clients, which often manage multimillion-dollar chip projects. Wang explained that a major step will be opening a new headquarters in Santa Clara to put the company near the heart of Silicon Valley, expected to be operational within weeks. The company’s current base in Santa Barbara will remain a major R&D center. “ChipAgents’ agentic approach to AI chip design integrates seamlessly across the entire chip design workflow,” said Lance Co Ting Keh, a venture partner at Bessemer Venture Partners. “Bringing together disparate EDA tools from spec ingestion to waveform analysis, which we believe is the right way to tackle this complex, multistep process.” Image: SiliconANGLE/Microsoft Designer
