Modern semiconductor development is approaching a new constraint: engineering capacity. As chips become more complex, companies must manage smaller geometries, higher performance, lower power consumption, and greater integration while maintaining increasingly aggressive development schedules, tells Synopsys.
A leading-edge chip can require hundreds or thousands of engineers and more than a year of development before reaching manufacturing. At the same time, many semiconductor companies are expected to introduce new generations of silicon annually. Finding enough specialists is difficult because expertise in architecture, verification, implementation, and analysis takes years to develop.
Synopsys argues that AI could help close this capacity gap. AI tools are already producing productivity gains of two to five times across specific chip design tasks. However, the company sees a larger role for AI beyond accelerating individual activities. It could change the way engineering work itself is executed.
Complexity affects the entire chip development flow. Engineers must translate design intent into architecture and RTL, verify correctness and corner cases, and resolve timing, power, congestion, and design-rule constraints. Analog design, simulation, and analysis add further workloads as teams explore trade-offs and diagnose problems.
Many essential tasks also consume expert time without fully using engineers’ specialized knowledge. Setting up runs, debugging failures, identifying root causes, and repeatedly adjusting implementation parameters can slow projects and introduce human error.
Agentic AI could shift more of this work to intelligent systems. Rather than simply assisting engineers or recommending optimizations, AI agents could execute workflows toward defined engineering objectives. Engineers could then concentrate on judgment, creative problem-solving, and system-level decisions.
Synopsys presents this approach as a way to scale engineering capacity without requiring the semiconductor industry’s limited pool of specialists to expand at the same rate as chip complexity. As design demands increase, intelligent systems could therefore become an important part of keeping semiconductor development schedules manageable.
