EDA Agents move from concept to production use

DAC 2026, the world's largest EDA conference and exhibition, opened today in Long Beach, CA. This year's theme is "AI For EDA," and almost all the leading EDA vendors are using the show to demonstrate their latest EDA Agent applications in chip-to-system design. Xpeedic unveiled the newest results from their EDA Agent partnership with Lenovo.

Xpeedic and Lenovo started a strategic partnership on the EDA Agent in November 2025, aiming to bring AI into the full EDA design flow and open up new, AI-driven approaches to designing smart devices and systems. The partnership has moved past the strategic-alignment stage and is now producing R&D results that can be verified and demonstrated.

A Closed AI Design Loop That Runs the Whole PCB Flow

The centerpiece of the release is an AI design loop that closes the gap between PCB design and simulation, covering the full process end to end.

On the design side, an EDA Agent handles three key steps: library building, layout, and design rule checking. The AI Library Builder generates and maintains component libraries automatically; AI Auto-Placement produces the layout; and AI Checklist runs the design-rule checks. On the simulation side, the EDA Agent speeds up iterative simulation for DDR and high-speed signals — the CCT (Channel Check Tool) module checks DDR channels automatically, and the High-Speed Channel AI Optimization module tunes parameters for high-speed channels. Together, the five modules form a single closed loop: library building, layout, rule checking, simulation optimization, and back again.

The result has already been validated on the PCB for Lenovo's AI PC motherboard. On the design side, the EDA Agent cut down on manual engineering work and lifted the efficiency of automated schematic-symbol and PCB-footprint library building by more than 50%. On the simulation side, it cut the number of simulation iterations needed and improved SerDes full-channel optimization efficiency by more than 80%.

Xpeedic: From AI-Accelerated Design to Physical AI Across the Whole Chip-to-System Chain

Beyond this PCB milestone, Xpeedic's EDA Agent reflects a broader company positioning: multi-physics AI agent collaboration paired with chip-to-system end-to-end autonomous optimization. Built around multi-physics simulation, it extends AI agents into physical-layer stages like signal integrity (SI), power integrity (PI), and electromagnetic-thermal coupling, spanning chip, package, board, and system design. Xpeedic isn't dropping AI into one point in the flow — it has built a closed physical-design loop that runs continuously from chip to system.