Artificial intelligence is increasingly seen as a way to overhaul semiconductor design, from generating layouts to predicting verification outcomes and reusing intellectual property. Yet according to a new analysis from Semiconductor Engineering, the limiting factor is not the sophistication of AI models or available computing power. Instead, the obstacle is data: AI-driven EDA workflows require structured, contextualized, and verifiable information, and most current design environments simply do not provide it.
The report points to fragmented repositories, inconsistent metadata, and undocumented IP reuse as key problems. Without organization and traceability, the relationships that give design data meaning are lost, making AI automation unreliable and hard to implement. The authors argue that the path forward begins with data organization rather than algorithm selection—teams must first build a strong data foundation that connects every stage of the design lifecycle before optimizing processes.
To address this, the analysis highlights Keysight SOS as purpose-built infrastructure for achieving AI readiness. While the piece is effectively a vendor-oriented white paper, its central claim is concrete: for AI to deliver trustworthy insights in chip design, the industry must treat data infrastructure as the first and most critical step. No alternative approaches are compared, but the message is clear—clean, connected data is the real enabler of AI-driven semiconductor innovation.