Astera Labs reduced chip design simulation times by 3.5X using NVIDIA B200 GPU-accelerated EC2 instances running Synopsys PrimeSim, creating what industry observers call a self-reinforcing ecosystem where AI chips are designed using AI-accelerated tools.
NVIDIA announced strategic partnerships with Synopsys and Applied Materials at GTC 2026 on March 16, targeting GPU acceleration across electronic design automation (EDA) workflows. The collaboration extends to quantum chemistry applications for semiconductor manufacturing processes.
Jitendra Mohan from Astera Labs stated that NVIDIA B200 GPU-accelerated computing on AWS has significantly reduced simulation times, a bottleneck in chip development cycles. Traditional CPU-based EDA tools require weeks for complex simulations that GPU-accelerated systems complete in days.
Synopsys launched AgentEngineer L4, an agentic workflow system announced March 16, 2026, combining its EDA tools with AI acceleration. The platform represents the semiconductor industry's shift toward AI-driven design automation, where machine learning models optimize circuit layouts and predict performance issues before physical prototyping.
The Astera Labs, Synopsys, NVIDIA and AWS partnership creates a complete stack for accelerated chip design. Cloud-based GPU instances eliminate the capital expense of on-premises compute clusters while providing elastic scaling for simulation workloads.
GPU acceleration addresses the growing complexity of modern chip designs, where billions of transistors require extensive simulation across power, timing and signal integrity domains. Each design iteration on traditional infrastructure delays time-to-market by weeks, a costly bottleneck when semiconductor products have 2-3 year development cycles.The chip design industry's adoption of GPU acceleration mirrors the broader AI infrastructure trend, where training workloads drove initial GPU deployment and inference applications followed.
This creates a feedback loop: faster EDA tools accelerate development of next-generation AI chips, which deliver better performance for the design tools themselves. The ecosystem locks in NVIDIA's architecture as the dominant platform for semiconductor design infrastructure.