Synopsys, in collaboration with NVIDIA, has unveiled an AI-based engineering environment that connects semiconductor design to system verification. Amidst increasing design complexity and shrinking development times, the concept is to combine simulation, digital twins, and agent AI to verify performance and identify issues prior to the production of physical prototypes.

Synopsys announced related achievements at NVIDIA GTC 2026, held in San Jose, USA, on March 16 (local time). This announcement aligns with the trend of Synopsys fully integrating mechanical, thermal, and fluid analysis capabilities into semiconductor design software following the acquisition of Ansys, which was completed last year.

The core of the technology lies in combining semiconductor design tools, multiphysics simulation, GPU acceleration, and agent AI into a single work environment. Synopsys has proposed a direction to reduce computationally intensive analysis times and automate complex chip design tasks using a multi-agent approach by leveraging NVIDIA's CUDA-X, Omniverse, and AI infrastructure. An L4-level agent EDA workflow was also demonstrated at the event.

Specific application examples were also presented. Applied Materials increased the speed of quantum chemistry simulations by up to 30 times, and Honda confirmed that with GPU acceleration using Ansys Fluent, computations were 34 times faster and cost savings were 38 times greater compared to 1,920 cloud CPU cores using only four GB200 GPUs. Astera Labs increased verification speed by 3.5 times by utilizing B200 GPU-based PrimeSIM on AWS.

In the field of robotics, ADI is building digital twins of tactile sensors and robot motion benchmarks using Synopsys software in the NVIDIA Omniverse and Isaac Sim environments. This is interpreted as an attempt to improve the accuracy of physical AI development by reducing the difference between results in virtual environments and real environments.

As the semiconductor industry reshapes around chiplets and high-density packaging, there is a growing need to examine not only electrical performance but also thermal, mechanical deformation, and system behavior during the design phase. This announcement can be seen as an example demonstrating that AI is expanding beyond simple design assistance to integrate simulation, verification, and virtual prototyping.