Table of Contents
The duo continues to progress through new collaborations aimed at embedding AI across semiconductor design and engineering workflows
The Nvidia GTC event has always been known to be more than just an industry gala. For years, it has served industry pundits, trends observers, and the audiences a glimpse into the next big things in tech. So, at the GTC-DC event last week, predictably, the highlights were AI, quantum, and 6G — the trio dominating headlines in 2025.
Synopsys, an industry bellwether in electronic design automation (EDA) and silicon IP, attended the event showcasing its engineering solutions powered by Nvidia — and announced yet another collaboration with the chip giant. According to the announcement, Synopsys is using the Nvidia NeMo Agent Toolkit and Nemotron open models for AgentEngineer, Synopsys’ agentic AI technology (released in May) for streamlining and improving engineering workflows.
“From our perspective, AI is really the right solution at the right time because we are seeing a tremendous increase in chip design complexity generation-over-generation, and for Nvidia to build these amazing chips on a yearly cycle, the only way to really have that kind of schedule and deliver that kind of complexity is to have AI all over the design flow,” said Shankar Krishnamoorthy, chief product development officer at Synopsys, during a panel discussion at the event.
Now in it’s third generation of AI solutions, Synopsys has been decisively incorporating reinforcement learning into its validation, test, and analog designs for years. The goal, as CEO Sassine Ghazi has emphasized in the recent past, is to embed AI — especially AI agents — in chip design and simulation to enable human engineers to offload the laborious steps to the agents, while achieving higher efficiency and productivity themselves.
The recent collaboration builds on this same effort and deepens integration of AI tools into Synopsys’ chip design flows for enhanced engineering productivity, design quality, and time-to-market.
In an interview with Reuters, Krishnamoorthy had said, “AI plays a huge role, because your R&D capacity is not growing… You’ve got a certain team, you’re not going to just double it, triple it, quadruple it. So you have to increase this R&D capacity.”
The integration of Nvidia AI toolkit into Synopsys’ technology has resulted in breakthrough accelerations, according to both companies. “Simulation software providers like Ansys, part of Synopsys, are using NVIDIA PhysicsNeMo to achieve up to 500x speedups in computational engineering,” writes Timothy Costa, senior director of CAE, quantum and CUDA-X at Nvidia, in a blog.
By leveraging Nvidia for Fluent, Ansys’ fluid simulation software, it is able to manage workloads much faster. Running on eight Nvidia Blackwell GPUs, Fluent demonstrated 50x speed gains compared to 258 CPU cores.
That acceleration is further compounded by the incorporation of Nvidia PhysicsNeMo AI physics framework and Nvidia DoMINO NIM for initializing simulations. The result was 10x more GPU time gains — and 500x faster speeds compared to traditional methods, said Nvidia and Synopsys.
“These are not incremental optimizations—they’re architectural shifts that redefine engineering velocity and fidelity,” Prith Banerjee, senior VP of Simulation & Analysis Incubation Group at Synopsys, wrote in a LinkedIn post.
Simulation tools are vital for catching bugs in chip designs early on. The new framework marks the beginning of advanced fluid simulations where a highly accurate initial state eliminates the need for costly computational tasks in the early phases. The combined effect of GPU performance gains and higher accuracy of AI physics leads to simulations being completed in 40 mins — compared to what took 2 weeks before, the companies said.
The collaboration is however not limited to AI. Synopsys is also leveraging Nvidia tools to accelerate computational materials simulation. Its QuantumATK atomistic simulation software now uses the Blackwell architecture and Nvidia CUDA-X libraries, achieving “15x improvement in time to results” for quantum-mechanical methods, resulting in faster development of solutions.
Despite their continued collaborations through the years, not many people know about the history of Synopsys and Nvidia’s alliance. It in fact goes back to when Nvidia’s corporate office used to be a one-room office. The alliance recently started to create buzz when last year Ghazi and Jensen Huang, CEO of Nvidia, together took to the stage at a 2024 GTC event. Huang revealed that Nvidia has been relying on Synopsys’ simulation and design tools to catch errors in its chip designs and tape-outs for years.
As computational power becomes a strong currency in the AI economy, the partnership of these two juggernauts will be a force-multiplier in the AI chip design, validation and automation space, as they lean on each other to capitalize on this huge market opportunity in AI-driven engineering.
As Hiếu Trần Trung, senior manager, SoC Engineering at Synopsys, wrote in a recent LinkedIn post, “The Synopsys–NVIDIA relationship positions both companies as central players in the shift toward AI-powered engineering. Their joint efforts are laying the foundation for a future where digital twins, real-time simulation, and AI physics models become standard tools across industries.”
