Synopsys has announced the introduction of the industry's first autonomous Electronic Design Automation (EDA) workflows, developed in collaboration with Microsoft and currently being evaluated by AMD. These agentic AI workflows, powered by AgentEngineer™ technology, are available for evaluation on the Microsoft Discovery platform. By moving beyond simple task automation toward long-running engineering execution, this initiative aims to accelerate the entire chip design lifecycle from specification to silicon. This development represents a significant shift for semiconductor engineering teams facing unprecedented silicon complexity and the need for increased design velocity.
Synopsys Autonomous Workflows on Microsoft Discovery
The announcement introduces two distinct autonomous workflows designed to handle complex engineering tasks. The first is a fully-autonomous debug closure workflow that utilizes domain-specific and task-level agents to perform verification and root cause analysis (RCA). This workflow is orchestrated via Microsoft Discovery to identify design failures and automate debug tasks. Early evaluations of this specific workflow indicate a 25% to 40% reduction in debug cycle time, which can save engineering teams multiple weeks of effort.
The second workflow focuses on implementation and closure, leveraging Synopsys implementation agents and Fusion Compiler on Azure. This system utilizes agents to automate implementation quality-of-results (QoR) tuning and closure. Initial results from these implementation workflows demonstrate improved QoR. Together, these tools aim to combine large-scale reasoning with domain expertise and cloud-scale compute to support the development of next-generation AI infrastructure silicon.
AMD and Microsoft Collaboration on AI Silicon
The collaboration between Synopsys, AMD, and Microsoft focuses on addressing the scaling challenges of AI-powered systems and hyperscale computing. AMD is actively evaluating these autonomous workflows to accelerate the development of its next-generation products. Alex Starr, AMD Corporate Fellow, noted that AI-driven workflows leveraging Discovery and deep EDA domain knowledge represent a paradigm that can improve both design velocity and overall silicon quality.
Microsoft Discovery serves as the foundation for these applications, designed to accelerate scientific and engineering innovation. Aseem Datar, Microsoft Corporate Vice President, highlighted that chip design is an ideal application for Discovery due to its inherent complexity. This milestone marks the first EDA applications available for evaluation on the Microsoft Discovery platform, building upon previous work regarding agentic specification-to-RTL workflows.
Key Takeaways
- The Synopsys debug closure workflow has demonstrated a 25–40% reduction in debug cycle time during early evaluations.
- These autonomous EDA workflows are the first of their kind available for evaluation on the Microsoft Discovery platform.
- AMD is actively evaluating the application of these autonomous workflows to accelerate its next-generation product development.
TechInsyte's Take
In our view, this move by Synopsys signals a critical transition from "AI-assisted" tools to "AI-agentic" engineering. By integrating autonomous workflows directly into the Microsoft Discovery cloud environment, Synopsys is addressing the massive compute and reasoning requirements of modern silicon design. The fact that AMD is already evaluating these workflows suggests that the industry is moving past the hype phase and into practical, high-stakes implementation. For CTOs and hardware leaders, this represents a shift where the bottleneck in chip development may move from manual engineering execution to the orchestration of intelligent, autonomous agentic stacks.
Source: PRNEWSWIRE
