Cadence and NVIDIA expand partnership to reinvent engineering for the era of AI and accelerated computing

At CadenceLIVE Silicon Valley 2026, Cadence announced an expanded partnership with NVIDIA. This delivers fast solutions across agentic AI, physics-based simulation, and digital twins to enable new levels of productivity and accelerate next-generation engineering design flows across semiconductor design, physical AI systems, and hyperscale AI factories.

Cadence and NVIDIA are accelerating Cadence EDA and SDA solutions using the Cadence® Millennium™ M2000 supercomputer with NVIDIA CUDA-X, AI physics, Omniverse libraries, and NVIDIA AI infrastructure.

By combining Cadence’s leadership in agentic AI-driven design, electronic design automation (EDA), and systems design and analysis (SDA) with NVIDIA CUDA-X, AI physics, and omniverse libraries for industrial digital twin solutions, the companies will redefine engineering productivity across three critical design areas and accelerate innovation at true agent speed.

“Agentic AI and digital twins are reshaping the entire engineering landscape, from semiconductor design to global AI systems,” said Anirudh Devgn, president and CEO of Cadence. “Our expanded collaboration with NVIDIA will accelerate the convergence of design and physical realization, connecting Cadence AgentStack, Physical AI Stack, and AI Factory digital twins with NVIDIA’s accelerated computing breakthroughs to deliver unprecedented speed, accuracy, and reliability in simulation and system development.

“We are at a tipping point in computing, where CUDA-accelerated computing and AI are reinventing the engineering process,” said Jensen Huang, founder and CEO of NVIDIA. “By building everything as a full-fidelity digital twin first, we are the first in the digital world to be able to explore, test, and optimize ideas at unprecedented speed and scale. Together, NVIDIA and Cadence will realize this vision and transform the way engineers design, build, and operate the world.”

Cadence tool acceleration for EDA and SDA

Cadence and NVIDIA are accelerating Cadence EDA and SDA solutions using NVIDIA CUDA-X, AI physics, Omniverse libraries, and Cadence.^® millennium^™ M2000 supercomputer with NVIDIA AI infrastructure. As part of this expanded collaboration, Cadence will accelerate solvers based on a wide range of principles and leverage AI physics models to speed up engineering workflows by up to 100x.

Cadence EDA and SDA customers and partners, including Ascendence, Argonne National Laboratory, Honda R&D, Samsung, and SK Hynix, are already leveraging Cadence solutions accelerated by NVIDIA to bring accelerated products to market faster.

AgentStack: Agentic AI for next-generation chip design

Cadence recently announced ChipStack^™ AI Super Agent applies agent AI combined with principle-based EDA tools to transform semiconductor RTL design and verification. Initial deployments with more than 10 major customers have already demonstrated productivity gains of up to 10x for design and verification tasks.

Building on this foundation, Cadence today announced AgentStack^™a head agent designed to coordinate all aspects of semiconductor and system design. AgentStack extends the ChipStack AI super agent mental model and super agent architecture beyond RTL and verification to physical design, custom/analog design, migration, and system-level design workflows. AgentStack connects Cadence agents to the Cadence EDA platform. The Cadence EDA platform leverages NVIDIA Nemotron and runs on NVIDIA accelerated computing to orchestrate long-running multi-agent workflows.

As an early partner, NVIDIA has adopted AgentStack flows for semiconductor and system design flows and provided real-world feedback to help Cadence enhance and expand AgentStack for adoption across a wide range of industries. This evolution marks a significant shift from traditional script- and GUI-driven flows to agent-driven flows that can infer design hierarchies, relationships, and protocols, significantly reducing iteration cycles from days to hours.

Also read: AiThority Interview with Glenn Jocher, Ultralytics Founder and CEO

Embedded agent AI for physics AI

Beyond semiconductor design, Cadence and NVIDIA will extend their collaboration to embedded agent AI for physical AI, combining Cadence’s physical AI stack with NVIDIA robotics simulation libraries and accelerated computing to help bridge the critical “sim-to-real” gap for robots and autonomous machines. By integrating and accelerating Cadence’s high-fidelity multiphysics simulation and AI workflows with the NVIDIA Isaac open-source simulation library and Cosmos open-world models, customers gain an end-to-end agent tuning workflow that links world model training, accurate physics, large-scale scenario testing, and continuous real-world feedback.

At a high level, the collaborative stack orchestrates AI agents throughout their lifecycle, from training orchestration, physical surrogate training, and policy optimization to validation and deployment feedback. This workflow spans virtual training in NVIDIA Isaac Sim and Isaac Lab, evaluation with detailed cadence physical models, and mission-scale scenario simulation in VTD (virtual test drive) and VTDx (enhanced high-fidelity simulation environments for complex real-world scenarios).

The results are then deployed to NVIDIA Jetson robotics and edge AI systems, where live virtual twins enable continuous monitoring and improvement. The Cadence-NVIDIA flow is designed to significantly accelerate experimentation while making physical AI systems safer and more reliable when deployed in the real world by incorporating accurate physics through training, validation, and inference.

Achieving lowest cost per token with AI Factory Digital Twin

This collaboration also extends to AI Factory, where Cadence will integrate NVIDIA Omniverse DSX Blueprints to deliver a next-generation AI Factory digital twin that helps customers design, simulate, and optimize large-scale Vera Rubin and Grace Blackwell AI factories for training and inference. These AI Factory digital twins focus on an important new metric for hyperscale AI: tokens per watt, or the number of model tokens processed per unit of power consumed.

Using Cadence system analysis and data center simulation tools in conjunction with the NVIDIA DSX Library and Omniverse DSX Blueprints, customers can consider trade-offs in GPU power settings, system configuration, and cooling architecture before deploying physical systems. The 10 Megawatt (MW) AI Factory joint use case demonstrated that modeling GPU operation at reduced power (MaxQ) can increase tokens per watt by up to 17% and billions of dollars in annual revenue per gigawatt in large-scale deployments, increasing net annual revenue and highlighting the value of simulation-driven design for AI factories.

The digital twin of the NVIDIA DSX-based AI Factory also demonstrated that MaxQ operation combined with warmer coolant could increase the production of tokens per watt by approximately 32%. By capturing the interactions between IT loads, cooling systems, airflow, and control logic in a high-fidelity digital twin, operators can safely drive their AI factories toward maximum tokens per watt while respecting power and thermal constraints.