Silvaco Group (technology company) has announced a collaboration with NVIDIA (technology company) to develop next-generation digital twins for semiconductor design and manufacturing, combining Silvaco's physics-based simulation software with NVIDIA's accelerated computing infrastructure and artificial intelligence platforms.

What the Collaboration Covers

The partnership, announced, brings together Silvaco's TCAD, EDA software, and semiconductor IP portfolio with NVIDIA's CUDA-X libraries, PhysicsNeMo framework, Omniverse libraries, Nemotron open models, and NVIDIA Cosmos.

The stated goal is to help semiconductor customers build, train, and deploy high-fidelity digital twins capable of predicting, optimizing, and validating complex semiconductor systems with greater speed and accuracy than current approaches allow.

Silvaco, headquartered in Santa Clara, California, and listed on Nasdaq under the ticker SVCO, has organized the collaboration around four principal focus areas: GPU-accelerated physics simulation, AI-driven surrogate modeling, digital twin visualization and collaboration, and scaled cloud-native engineering workflows.

GPU Acceleration and a Key Early Benchmark

On the simulation side, Silvaco intends to use NVIDIA accelerated computing and CUDA-X libraries to speed up its semiconductor device, process, photonics, and multiphysics simulation tools.

The company cited one completed proof point to illustrate the potential: a fully scaled three-dimensional finite-difference time-domain simulation of a photonic edge coupler, run across 32 NVIDIA GPUs connected with NVLink, using 3.2 billion mesh nodes.

That simulation completed in under four hours and achieved less than 0.15 dB difference between measurement and simulation results. According to Silvaco, the same workload did not converge when run on CPUs.

The companies said GPU acceleration is expected to produce dramatic reductions in simulation runtimes and increased design productivity more broadly across Silvaco's simulation product lines.

AI Surrogate Models and Physics-Informed Workflows

A second pillar of the collaboration involves Silvaco's use of NVIDIA PhysicsNeMo to develop AI surrogate models. These models are described as customizable tools that complement high-fidelity physics simulation and accelerate the exploration of design alternatives.

Rather than replacing physics-based simulation, the surrogate models are positioned as a complementary layer that allows engineers to move more quickly through design space before committing to full simulation runs.

This approach reflects a broader industry shift toward hybrid workflows that combine the accuracy of physics simulation with the speed of machine learning inference.

Silvaco has indicated it will offer these surrogate models as configurable components that customers can adapt to their specific design and manufacturing contexts.

Visualization, Collaboration, and Factory Applications

Beyond simulation and modeling, the partnership extends into visualization and real-time collaboration. Silvaco plans to connect its digital twin environment with NVIDIA Omniverse libraries and NVIDIA Cosmos to create collaborative simulation environments that span semiconductor fabs, manufacturing systems, robotics platforms, and infrastructure applications. The intention is to give distributed engineering teams interactive access to simulation data and visualization tools simultaneously.

The companies identified factory optimization and predictive manufacturing as specific application areas for these capabilities, alongside semiconductor process, device, packaging, and photonics simulation, and AI-assisted development of next-generation chips and advanced nodes.

Silvaco also stated plans to build cloud-native workflows designed to support design, testing, and validation across distributed teams and compute environments, which the company described as scaled engineering workflows suited to global engineering organizations.

What Customers Can Expect

According to the announcement, the combined technologies are expected to help customers reduce simulation cycles from weeks to days through GPU-accelerated simulation and AI-driven modeling, enabling faster design iterations and reduced time-to-market.

The companies also said the high-fidelity digital twins will provide deeper visibility into system performance for more precise validation and optimization, and that cloud-based visualization and AI-driven workflows will enable global teams to collaborate more efficiently on complex simulations.

Executive Statements

Walden C. Rhines, President and Chief Executive Officer of Silvaco, framed the collaboration in terms of a broader industry transformation. "The convergence of physics-based simulation, accelerated computing, and artificial intelligence is transforming design and manufacturing," Rhines said.

"By combining Silvaco's deep expertise in semiconductor and multiphysics digital twins with NVIDIA's industry-leading computing and AI platforms, we can help customers model increasingly complex systems with greater speed, fidelity, and confidence."

Da Yang, senior director of product for semiconductor and EDA at NVIDIA, described digital twins as increasingly essential tools for engineering and manufacturing.

"By using NVIDIA AI, open models, libraries and accelerated computing, Silvaco is connecting high-fidelity simulation, helping customers move faster from modeling to insight across semiconductor design and manufacturing," Yang said.

Company Background

Silvaco describes itself as a provider of AI-enabled TCAD and EDA solutions and semiconductor IP solutions used across display, power devices, automotive, memory, high-performance compute, foundry, photonics, IoT, and 5G and 6G mobile markets.

The company has offices across North America, Europe, Brazil, China, Egypt, Japan, Korea, Singapore, Taiwan, and Vietnam, in addition to its Santa Clara headquarters.

Tracking the World's Most Complex Semiconductor Pipelines Just Got Simpler

Semiconductor fabrication facilities, advanced packaging plants, and R&D campuses are rising across markets worldwide at a pace that makes manual intelligence-gathering nearly impossible. If your team is still piecing together project data from fragmented sources, you are already behind the curve in one of the most capital-intensive industries on the planet.

The Global Project Tracking (GPT) platform by Blackridge Research was built to give semiconductor manufacturers, equipment suppliers, construction contractors, and investment analysts a single, continuously updated source of truth for facility development activity across every major region.

Whether you are monitoring fab expansions in East Asia, greenfield wafer plants in Europe, or advanced chipmaking investments emerging across markets worldwide, the GPT platform delivers the structured, actionable intelligence your business development and strategy teams depend on.

• Upcoming Projects

• Tender Notices

• Contract Awards

• Projects Under Construction

• Completed Projects

See how the Global Project Tracking (GPT) platform by Blackridge Research can sharpen your competitive edge in the semiconductor sector. Book a Free Demo with our team today.