AI for chip design

 This symbolic tools doing the heavy lifting beyond LLMs post made em wonder how it is applicable to EDA (Electronic Design Automation).

Just like the Generative AI tech stack and resources for software, I wanted one for HW. 

Baed on research from chatgpt-

GenEDA, NetTAG, and DeepRTL2 build new bridges across LLMs and structural graph models.

NetTAG: A Multimodal RTL-and-Layout-Aligned Netlist Foundation Model via Text-Attributed Graph

Combines LLM-based text encoding of gate semantics with graph transformers to align both RTL and layout representations for netlist modeling

A Survey of Research in Large Language Models for Electronic Design Automation

Comprehensive overview of LLM applications in EDA, covering architecture, fine-tuning, and model-size implications.

The Dawn of AI-Native EDA: Promises and Challenges of Large Circuit Models

Insightful position paper advocating multimodal “circuit models” that align RTL, netlist, and layout representations.

DeepRTL2: A Versatile Model for RTL-Related Tasks

Introduces a unified model addressing both generation (Verilog synthesis) and embedding-based tasks (code search, equivalence checking).

GenEDA: Unleashing Generative Reasoning on Netlist via Multimodal Encoder-Decoder Aligned Foundation Model

Aligns graph-based netlist encoders with LLM-style text decoders in a shared latent space.

Enables high-level functional generation from low-level netlists. Improves generative reasoning

AiEDA: Agentic AI Design Framework for Digital ASIC System Design

LLM aided design

LLM4EDA

NVIDIA prefixRL

Designing arithmetic circuits with deep reinforcement learning

concurrent training and data collection

Analog-centric AI: surveys and papers on using GNNs, GANs, Bayesian tuning for analog layout generation—listed in the AI4EDA repository MDPI+5AI for EDA+5ResearchGate+5.

OpenROAD Project: DARPA-backed, open-source RTL-to-GDSII toolchain with ML-augmented flows and RL-based placement research ar5iv+5The OpenROAD Project+5Wikipedia+5.

RL floorplanning: Google's deep-RL chip placement (Nature 2021, Mirhoseini et al.) — early validation of AI outperforming human experts arXiv+2MDPI+2Wikipedia+2.

AI in real world chip design workflows

Transformers for NLP