At the 2026 DAC Chips to Systems Conference, Synopsys unveiled new agentic AI capabilities developed in collaboration with NVIDIA, bringing fully autonomous AI agents to chip design, verification, analog design, and engineering simulation. The companies said the technology transforms labor-intensive engineering workflows into automated processes that improve productivity and accelerate product development.
The new capabilities combine Synopsys’ expertise in electronic design automation (EDA) and computer-aided engineering (CAE) with NVIDIA’s accelerated computing platform, NVIDIA Nemotron AI models, and OpenShell runtime. The result is a long-running, autonomous AI framework that can reason, plan, execute complex engineering workflows, and validate its own work.
“AI is fundamentally reshaping engineering, and Synopsys is at the forefront of this transformation, enabling fully autonomous agents across every stage of silicon and systems development,” said Ravi Subramanian, chief product management officer at Synopsys.
【Resource】The Definitive Resource for Everything FPGA
NVIDIA highlighted the shift from AI assistants to autonomous engineering agents capable of handling end-to-end development tasks.
“The future of engineering is agentic, where AI agents reason, plan, execute complex workflows and verify their own work across the entire product development lifecycle,” said Tim Costa, vice president and general manager for computational engineering at NVIDIA.
One of the key demonstrations at DAC focused on design verification (DV), where coverage closure remains a major bottleneck in chip development. Synopsys introduced an autonomous verification workflow built on its AgentEngineer platform and NVIDIA’s Agent Toolkit and Nemotron 3 Ultra model. A central orchestrator AI agent coordinates specialized agents throughout the verification process, from test-plan generation to coverage closure and debug.
According to Synopsys, the workflow reduces verification closure from weeks to hours, delivering up to 50× faster time-to-validated RTL while improving coverage by an additional 20%.
The companies also extended agentic AI to analog and mixed-signal (AMS) development. Using AI-powered Custom Compiler Layout Synthesis together with AgentEngineer technology, engineers can describe design goals in natural language while autonomous agents manage design creation, SPICE simulation, implementation, and verification. Synopsys said the approach can improve engineering productivity by up to three times.
Beyond chip design, Synopsys introduced an autonomous CAE workflow for electronics thermal analysis. Built using NVIDIA Agent Toolkit, CUDA-X libraries, Ansys Icepak, and the open-source PyAEDT libraries, the workflow automates simulation setup as well as pre- and post-processing, significantly reducing engineering effort.
Synopsys also expanded GPU acceleration across its EDA and multiphysics portfolio, which now includes more than 20 GPU-enabled products. Recent enhancements include up to 18× faster PrimeSim SPICE simulations on NVIDIA GPUs, up to 50× acceleration for QuantumATK Gaussian-basis quantum chemistry simulations, up to 200× faster machine-learned force-field simulations on NVIDIA Blackwell GPUs, and a 10× speedup for Ansys Lumerical FDTD 3D electromagnetic simulations when used with Synopsys’ Multiphysics Fusion solution.
The collaboration also continues to leverage NVIDIA CUDA-X libraries, including cuLitho, cuDSS, cuEST, and the newly announced cuISS library, to further accelerate engineering simulation and design workflows.