AI-Enhanced Chip Design: Pioneering the Future at DAC 2025

On July 9, 2025, a DACtv session illuminated the transformative role of artificial intelligence (AI) in chip design, as presented by Ankur Gupta of Siemens EDA in the YouTube video. The speaker explored how AI is revolutionizing electronic design automation (EDA), addressing the semiconductor industry’s challenges in managing escalating chip complexity, optimizing performance, and accelerating time-to-market. By integrating AI into design workflows, the industry is poised to meet the demands of next-generation technologies like AI accelerators, 5G, and IoT, while fostering sustainability and innovation.

Modern chip designs, with billions of transistors, push traditional EDA tools to their limits. The speaker highlighted AI’s ability to enhance critical design phases, including synthesis, place-and-route, verification, and power optimization. Machine learning (ML) algorithms, for instance, predict optimal circuit layouts, reducing design iterations by up to 25%. Reinforcement learning models further refine power-performance-area (PPA) metrics, achieving 10-15% improvements in energy efficiency and performance. These advancements are crucial for AI accelerators, which require high computational throughput and low power consumption to support generative AI and large language models (LLMs).

The presentation emphasized AI’s impact on verification, a bottleneck in chip design. Traditional methods, reliant on manual testbench creation, struggle with the scale of modern SoCs. AI-driven tools automate testbench generation and coverage analysis, cutting verification time by 20-30%. For example, LLMs can generate SystemVerilog assertions from natural language specifications, streamlining the process. The speaker cited industry adoption, noting that top semiconductor firms have reduced tape-out schedules by months using AI-enhanced EDA, aligning with the market’s projected growth to $1 trillion by 2030.

Cloud integration is a key enabler. AI-driven EDA platforms, like those from Synopsys and IBM, leverage cloud scalability to handle compute-intensive tasks. Hybrid workflows allow seamless bursting to public clouds, mitigating on-premises compute shortages. The speaker highlighted the importance of infrastructure-as-code (IaC) tools like Terraform, which streamline resource allocation for design workloads. This flexibility is vital for startups and smaller firms, enabling access to high-performance computing without massive capital investment, thus democratizing advanced chip design.

Sustainability was a central theme. AI data centers, consuming gigawatts, raise environmental concerns. AI-optimized chip designs, such as low-power microcontrollers for edge devices, reduce energy use by minimizing cloud data transfers. The speaker noted that AI-driven thermal and power modeling in 3D chiplet architectures can cut energy consumption by 15%, supporting greener technologies in automotive and IoT applications. Collaborative efforts with foundries ensure process design kits (PDKs) integrate seamlessly with AI tools, enhancing manufacturability and reducing waste.

An audience question on academic access to AI tools prompted discussion of federated learning, where institutions train models locally on limited datasets, aggregating insights without sharing sensitive data. This approach, inspired by healthcare’s AI-driven molecule discovery, could enable universities to contribute to chip design innovation while protecting intellectual property. The speaker advocated for industry-academia partnerships, like those with NYDesign, to cultivate talent and drive AI adoption.

The session underscored AI’s pivotal role in shaping chip design’s future. By automating complex tasks, optimizing resources, and prioritizing sustainability, AI empowers the industry to navigate the challenges of trillion-gate SoCs. The speaker urged attendees to embrace this transformative era, leveraging DAC’s collaborative environment to advance EDA tools and ensure the semiconductor industry meets the demands of an AI-driven world.

Also Read:

Synopsys FlexEDA: Revolutionizing Chip Design with Cloud and Pay-Per-Use

Perforce and Siemens: A Strategic Partnership for Digital Threads in EDA

AI-Driven Chip Design: Navigating the Future