NVIDIA Supercharges Chipmaking with Blackwell and CUDA-X

Leading players in the semiconductor industry — TSMC, Cadence, Siemens, Synopsys, and KLA — are integrating the new NVIDIA Blackwell platform and CUDA-X libraries into their workflows.

These tools significantly accelerate simulations, data processing, and the development of complex chips. For instance, TSMC uses NVIDIA cuLitho to speed up photolithography, identifying and correcting issues up to 25x faster before chips even reach production.

Cadence introduced its Millennium M2000 platform, built entirely on Blackwell. It also leverages NVLink Fusion to handle AI workloads and visualize the entire chip design cycle.

Scaling and Precision

Siemens is enhancing design verification and optical correction in its Calibre platform through parallel processing on Grace Blackwell, which is critical for today's nanometer nodes.

Synopsys achieved impressive results: up to 30x faster simulations in PrimeSim and 20x in Proteus. Integration with Blackwell boosted performance 12–20x across key EDA tools.

KLA, known for manufacturing quality control, is also adopting Blackwell. They plan to use the RTX PRO 6000 Server Edition to accelerate real-time AI analytics during chip fabrication.

NVIDIA Blackwell and CUDA-X are becoming the foundation for the next generation of chips — faster, more innovative, and more precise. It's not just a platform; it's an infrastructure for semiconductor breakthroughs.