From the Frontlines of Agentic AI EDA: Accelerated by the Arm Computing Era

A Report from CadenceLIVE Silicon Valley 2026

CadenceLIVE Silicon Valley 2026 opened with a high‑signal punch: a keynote fireside chat between Cadence CEO Anirudh Devgan and NVIDIA^® CEO Jensen Huang, followed immediately by Devgan's visionary address. Jensen framed the EDA industry's arc as a decisive move from generative AI to reasoning and now to agentic systems. A central theme was that agentic AI is fundamentally about tool use: these agents do not replace proven systems; they rely on them. Jensen's analogy of a humanoid robot in your home is not a microwave itself—but becomes powerful by using the microwave and every other kitchen tool efficiently. AI agents will operate 24x7 by orchestrating trusted EDA tools, not by passing them. Jensen was explicit that EDA is irreplaceable in this future because agents scale intelligence by leveraging known, validated engineering infrastructure.

Devgan's subsequent keynote built directly on that foundation, sharpening the narrative around agentic AI as Cadence's core direction. He introduced Cadence's ChipStack, VeriStack, and InnoStack AI Super Agents as a concrete abstraction layer for agent‑driven design, verification, and system innovation—positioning Cadence less as a collection of point tools and more as an always‑on, intent‑driven workflow platform. Cadence's long‑standing theme of "Design for AI and AI for Design" was reframed as an operational reality, with agents coordinating across implementation, verification, analysis, and signoff to collapse cycles and increase throughput.

Within that context, Devgan highlighted Arm^®-based compute as a critical enabler of the agentic future, tying it directly to Arm's latest announcement of its first Silicon AGI SoC as a signaling moment for where this architecture is headed. With Cadence's agentic EDA becoming a continuously running digital workflow—backed by digital twins and physics‑based simulation—compute efficiency, scalability, and heterogeneity become first‑order design constraints. The keynote clearly emphasized that Arm's efficiency makes it especially suitable for an always-on, agent-driven EDA layer, where success depends on cost, performance per watt, and overall system throughput.

Arm Leaders in the Spotlight at CadenceLIVE

From the Arm partnership perspective, CadenceLIVE 2026 was not just a keynote story—it was a full set of Arm‑anchored technical sessions, demos, and a featured panel presence:

Arm-Based Compute Kiosk (Demo / Meet‑ups)

Arm's Cloud AI team anchored the Arm compute presence on the show floor, creating a practical venue to discuss Arm‑based infrastructure for EDA workloads.

Closing the Final PPA Gap: AI‑Enabled Optimization for Arm's Highest Performance CPU

Nagaraj Kelageri (Sr Director, Arm) focused on the partnership's performance theme: using AI‑enabled optimization to close PPA gaps for top‑end CPU designs.

Shaping the Future of RTL Feedback Methodology

Giuliana Emmolo (Sr Engineer, Arm) highlighted methodology innovation—how faster, higher‑fidelity RTL feedback loops translate to better closure behavior downstream.

Enabling End‑to‑End EDA on Arm Compute for Lower TCO and Higher Infrastructure Flexibility

Jyotsna Chitri and Tim Thornton (Arm Infrastructure) addressed one of the most actionable partnership areas: running end‑to‑end EDA flows on Arm compute with a clear emphasis on infrastructure flexibility and improved total cost of ownership.

Featured Panel: "Closing the Sim‑to‑Real Gap in Physical AI"

Suraj Gajendra (VP, Arm) represented Arm in the event's public "spotlight" panel theme on bridging simulation and real‑world deployment in physical AI systems.

Cadence EDA Agentic AI Workflow Meets Arm-Scale Deployment

Put together, the public keynote narrative and the Arm session roster tell a coherent partnership story. The keynotes and track summaries describe an industry moving toward agentic orchestration—AI systems coordinating long, multi‑step workflows across established design infrastructure. Arm's CadenceLIVE presence then grounds that vision in deployment reality: PPA closure for leading CPU designs, faster RTL feedback methodologies, and end‑to‑end EDA on Arm compute aimed at lowering Total Cost Ownership (TCO) while expanding infrastructure choice.

From an "Arm + Cadence partnership report" standpoint, CadenceLIVE 2026 made the stakes clear: As agentic AI turns EDA into an always‑on, scalable "tool‑using" workflow engine, then where and how those workflows run becomes a strategic differentiator—and that is exactly where Arm‑enabled compute and Cadence's AI‑driven flows intersect most powerfully.

Learn more about Cadence's Arm solutions or contact us.