Presentations at DAC 2026 demonstrate how AI co-pilots and on-premises LLMs are being integrated into chip verification flows, with Samsung engineers using an on-premises LLM for pre-silicon Android software testing on emulators. This approach helps triage kernel panics and identify genuine software bugs while respecting data confidentiality, validated on a flagship SoC project. Additionally, STMicroelectronics presented an automated Python-scripted system to manage the bidirectional handoff between simulation and FPGA prototyping, treating them as fluid views of a continuum.
AI is shifting chip verification left, reducing debug cycles and accelerating time-to-market for complex SoCs, directly impacting compute hardware development.