Reticore Labs provides an autonomous design intelligence platform that speeds up physical design for AI hardware across all levels—from individual blocks to full-chip SoCs and 3DICs. The system leverages multimodal AI agents to automate and optimize the design workflow, enabling faster time‑to‑market for complex semiconductor products. It is positioned as a single, scalable solution for designers building next‑generation AI chips.
Funding
Funding not disclosed
Founders
Product
Problem
Designing AI hardware involves complex physical design steps across multiple hierarchy levels, from individual blocks to full-chip SoCs and 3DICs, leading to long development cycles and delayed time‑to‑market for AI workloads.
Solution
Reticore offers an autonomous physical design platform that applies AI‑driven intelligence to automate layout, placement, routing, and optimization tasks across all design scales. The system integrates a unified workflow that handles block‑level, full‑chip, and 3DIC designs within a single environment, reducing manual effort and iteration time. By leveraging domain‑specific models for agent‑based and multimodal AI workloads, the platform accelerates design convergence and improves performance‑power trade‑offs. Designers can generate production‑ready layouts faster, enabling quicker silicon tape‑out for complex AI accelerators.
Target Audience
Primary customers are semiconductor design teams and AI hardware companies developing custom accelerators, SoCs, or 3D integrated circuits for AI applications.
Features
- End‑to‑end automation of placement, routing, and timing closure for block, SoC, and 3DIC designs
- AI‑powered optimization engine tuned for agent‑based and multimodal AI workloads
- Single unified platform that eliminates data transfer between disparate tools and stages
- Built‑in support for 3D integration, including through‑silicon via (TSV) planning and thermal analysis
- Real‑time design rule checking and performance feedback to guide rapid design iterations