Nvidia purchases $2b stake in Synopsys

Nvidia and Synopsys have expanded their partnership to collaborate on AI and accelerated computing for engineering and design.

Nvidia has purchased US$2 billion worth of Synopsys common stock at US$414.8 per share.

Synopsys, based in California, provides software tools for semiconductor and electronics design.

The companies plan to integrate Nvidia’s GPU, AI, and digital twin technologies with Synopsys’ engineering software to speed up design, simulation, and verification processes across sectors such as semiconductors, aerospace, and automotive.

The partnership will also focus on cloud-based solutions, digital twins, and autonomous design capabilities for electronic design automation.

Nvidia and Synopsys said they will work together on joint ad and sales initiatives to promote these solutions.

The partnership is non-exclusive, with both companies continuing to work with other players in the semiconductor and electronic design industries.

🔗 Source: Nvidia

🧠 Food for thought

Implications, context, and why it matters.

Synopsys physics solvers need CUDA (Nvidia’s parallel computing platform for parallel programming) rewrites for gains beyond 20+ apps

Hitting “up to 500x speedups” seen with Ansys Fluent 1 takes algorithm rewrites, not a port. Multiphysics (simulations that couple multiple physical phenomena) and electromagnetic workflows (high-fidelity electromagnetics simulations) need high-precision double-precision floating-point compute 2.
The migration will take years 2. Work runs through 2026, 2027 with no firm end dates 2. Customers seeking big gains may wait.
Digital twins (virtual replicas of physical systems) using NVIDIA Omniverse (a simulation and collaboration platform) and Cosmos 3 add validation. Physics-based solvers set accuracy 2, so AI surrogate models (machine-learned approximations of physics simulations) only assist when workflows need strict precision.

Cloud providers, systems integrators can profit from GPU-accelerated Electronic Design Automation (EDA) for mid-market teams

Making GPU-accelerated engineering available through the cloud for teams of all sizes 4 opens room for AWS, Azure, Google Cloud. They can bundle Synopsys tools with tuned GPU instances for mid-market teams that lack on-premise gear.
Systems integrators (firms that implement, integrate complex software and hardware) can refit workflows during the multi-year migration 2. Focus areas include aerospace, automotive and energy where Synopsys seeks new total addressable market 2.
Cost-optimization tooling vendors can build orchestration layers (software that schedules, routes jobs) that shift work between CPU (central processing unit) and GPU. Synopsys tools will support both 2, so users can skip GPUs when a CPU run works.