Cadence Design Systems has taken a major leap in semiconductor design innovation with the launch of its ChipStack AI Super Agent, an advanced agentic AI solution developed to automate front-end silicon design and verification workflows.
The company describes the ChipStack AI Super Agent as the world’s first agentic AI framework for chip design automation delivering up to 10X productivity improvements across design coding, testbench creation, regression testing, debugging, and automated issue resolution.
Intelligent Design Through Agentic AI
The ChipStack AI Super Agent builds on Cadence’s long-term Intelligent System Design approach, integrating AI orchestration, accelerated computing, and simulation in one cohesive platform. It coordinates virtual “AI engineers” that work with Cadence’s established EDA platforms such as Verisium™, Cerebrus, and the JedAI data and AI framework.
Together, these technologies are designed to automate complex design and verification processes while maintaining accuracy and functional coverage.
ChipStack supports both cloud-based and on-premises deployment models. It is also compatible with advanced AI architectures, including NVIDIA Nemotron foundation models and OpenAI GPT-based cloud tools, with further customization possible through NVIDIA NeMo.
Early Adoption and Industry Collaboration
The new AI workflow is already being deployed by several high-profile chipmakers and system companies, including Altera, NVIDIA, Qualcomm, and Tenstorrent.
Anirudh Devgan, president and CEO of Cadence, called the launch a defining moment in the company’s AI strategy. “ChipStack represents a major leap in our design-for-AI and AI-for-design vision,” he said. “By leveraging intelligent agents that can autonomously operate our tools, we’re enabling dramatic productivity gains and freeing engineers to focus on next-generation innovation.”
Arvind Vidyarthi, senior director of engineering at Altera, shared early results from deployment: “The ChipStack AI Super Agent has reduced our verification effort by around 10X in certain areas, helping us achieve closure faster and more accurately. Combining an engineer-in-the-loop workflow with AI-driven verification gives us a significant productivity boost.”
From NVIDIA’s perspective, the collaboration underscores AI’s growing centrality to chip design. “As semiconductor complexity scales, AI has become essential to designing the next generation of chips,” said Timothy Costa, GM of Industrial and Computational Engineering at NVIDIA. “Integrations like ChipStack, which combine reasoning-based AI techniques with NVIDIA’s accelerated computing, unlock new levels of efficiency and design insight.”
Paul Penzes, vice president of engineering at Qualcomm, noted that the company’s evaluations also show promising results. “Early findings indicate strong performance enhancements, and we look forward to leveraging the productivity gains across a broader user base,” he said.
Tenstorrent’s principal engineer of RISC-V cores, Daniel Cummings, highlighted the tool’s performance benefits in formal verification. “Over a three-month evaluation on three major design blocks, ChipStack reduced verification time by up to 4X. Running it on Tenstorrent hardware also showed our system’s potential for high-performance on-prem inference in LLM workloads,” he remarked.
Paul Cunningham, vice president and general manager of R&D at Cadence, summarized the motivation behind the product: “The ChipStack AI Super Agent is a game changer for design and verification productivity, helping customers overcome the global shortage of senior engineering talent.”
Currently available in early access, the Cadence ChipStack AI Super Agent marks a pivotal step toward fully autonomous chip design, bridging design intelligence with scalable, enterprise-grade AI capabilities.
