Fuse EDA AI Agent supports a comprehensive range of Siemens EDA tools including Catapult for RTL coding, Questa One Agentic Toolkit for digital verification, Aprisa for physical implementation, Solido for custom IC design and verification, Veloce for hardware-assisted verification, and Calibre for physical verification sign-off. It also supports PCB-specific tools like Xpedition and Hyperlynx, along with Tessent for design-for-test workflows and Calibre OPC for manufacturing readiness.

Yes, Fuse EDA AI Agent is designed to enable seamless integration with third-party tools through its scalable MCP architecture and open framework for third-party integrations. This true multi-vendor tool solution allows customers to integrate their own workflows and models for enterprise-scale deployment.

Fuse EDA AI Agent is built on our deep understanding of semiconductor and PCB system workflows, specialized tool relationships, and the ability to interpret dense, physics-based EDA data. It includes specialized parsers for EDA file formats, built-in validation and domain-specific guardrails to prevent errors and hallucinations across multi-tool and multi-agent workflows.

Agent Skills are executable playbooks that enable accurate completion of complex, multi-step tasks across EDA workflows by integrating domain-specific knowledge. Designers can seamlessly customize these playbooks to meet their specific requirements.

Fuse EDA AI Agent offers flexible deployment through two options: a fully air-gapped, on-premise setup using local GPUs, or a hybrid approach that keeps data on-premises while routing the LLM through major cloud providers like Azure, AWS or Google Cloud.

We have expanded our strategic partnership with NVIDIA to bring advanced AI validation capabilities to Fuse EDA AI Agent. This combines NVIDIA AI infrastructure, including NeMo Gym for optimizing Agent Skills and MCP tools, Nemotron models and Switchyard for advanced reasoning with reduced token cost, OpenShell for secure orchestration and sandboxing and GPU acceleration of our physics-based EDA engines. The result is long-running and self-verifying AI agents that don't just automate tasks, but validate every decision with precision, improving result quality, reliability and token efficiency across engineering workloads.