CadenceLIVE Silicon Valley opened against the backdrop of an industry under rapid transformation, where AI is no longer a downstream application but a core driver of semiconductor scale, complexity, and opportunity. Across design, infrastructure, and system architecture, the boundaries between chip development, software intelligence, and physical systems are dissolving—reshaping how the world builds compute at scale. The event set out to explore this inflection point directly, bringing together leaders across the ecosystem to examine how AI is redefining not just what is possible, but what is now required.

How Agentic AI Is Rewiring Chips, Systems, and the Future of Design

The fireside chat opened the conference on a lighter yet strategically charged note, with Anirudh Devgan, president and CEO of Cadence, welcoming Jensen Huang, founder and CEO of NVIDIA, to the stage and framing the session as both a celebration and a continuation of a long-standing technical partnership. The tone quickly shifted between humor and deep architectural insight as the two leaders revisited the evolution of computing and the accelerating role of AI. Jensen emphasized the shift from generative AI to reasoning and then to agentic systems, describing how modern AI now “perceives, reasons, and executes plans,” and stressing that “agents are just all over our company. Our token use is going exponential.” The conversation anchored on a central idea: AI systems are becoming tool users at scale, and the value lies in how effectively they orchestrate existing engineering systems rather than replace them.

The dialogue then turned sharply toward electronic design automation, where Jensen underscored Cadence’s role in enabling “fundamental tools” that remain grounded in mature, highly optimized algorithms. He noted that in this new paradigm, “tool use, Agentic systems is defined by the word, ‘tool use’,” arguing that agents will amplify—not bypass—EDA systems like ChipStack, ViraStack, and backend automation flows. Anirudh and Jensen also explored how agentic systems will scale across thousands of virtual engineers, dramatically expanding design capacity beyond traditional constraints.

The discussion culminated in a shared vision of expansion: Cadence moving from chips into full-stack system and physical AI world design, and NVIDIA enabling the compute foundation for this shift. As Jensen summarized, Cadence’s transformation is unlocking “an infinite number of more chip designers… and now Cadence is gonna help every company design everything that they make.”

Anirudh Devgan’s keynote began with gratitude to customers for their long-standing support while setting the tone for a conference focused on feedback and iteration. What followed was a candid acknowledgment of how fast the industry is moving, and how much of that momentum is being shaped by users themselves.

From there, the keynote expanded into a sweeping view of the semiconductor industry’s acceleration, driven heavily by AI. Anirudh traced how projections have rapidly shifted—from a trillion-dollar industry by 2030 to potentially $2 trillion within the same timeframe—calling the pace “remarkable” and unprecedented. He pointed directly to leaders like Jensen Huang and companies such as NVIDIA and Broadcom as key drivers, alongside hyperscalers pushing demand at scale. But the opportunity comes with strain: “every iteration of the design… it requires like, two times more engineers. This is not sustainable.” That tension—between exponential demand and limited human scalability—set up the core thesis: “AI for design and design for AI.”

A major highlight of Anirudh’s keynote was the introduction of the Cadence SuperAgent portfolio, marking a shift to a new level of automation in electronic design. The ChipStack AI SuperAgent was positioned as a breakthrough for front-end design, bringing automation to RTL generation and verification planning, areas that historically lacked algorithmic solutions. By leveraging a “Mental Model” to capture design intent, the ChipStack AI SuperAgent enables reasoning, generation, and iteration at unprecedented speed.

Cadence AgentStack extends this capability across the design lifecycle, providing a unified environment where designers can interact with multiple agents through a common interface and natural language workflows.

ViraStack AI Super Agent tackles one of the hardest problems in EDA: analog design, combining prior designs, IP catalogs, embedded expertise, and optimization engines to enable autonomous workflows with reported productivity gains of 3X to 10X.

InnoStack AI SuperAgent addresses backend design, enabling autonomous flows from synthesis through signoff, allowing engineers to define goals while the system orchestrates execution, even running multiple block-level designs in parallel.

All of these agentic workflows run on Cadence’s JedAI platform, designed to give customers flexibility across models and deployment choices, whether cloud-based or on-premises. At the same time, Anirudh emphasized that AI amplifies strong tools—it does not replace them—reinforcing continued investment in core EDA engines across placement, routing, simulation, verification, and 3D-IC signoff, alongside acceleration across GPU and heterogeneous compute platforms.

Anirudh then connected this moment to a broader, multi-phase evolution of AI, from infrastructure AI to the emerging frontier of physical AI across cars, drones, and robotics, and eventually into life sciences. He framed Cadence’s long-term strategy through this lens, emphasizing computational software, the “three-layer cake” of compute, physics-based algorithms, and AI, and the rise of agentic systems transforming design workflows. The keynote built toward a clear inflection point: automation is no longer incremental—it’s foundational. 

Following Anirudh Devgan’s keynote, Charlie Kawwas, president of Broadcom, took the stage to offer a silicon and infrastructure view of the AI revolution. He unpacked how explosive AI growth is translating directly into unprecedented demand across compute, networking, and memory, with AI-driven capital spending now surging into the trillions. Yet the more striking dynamic, he noted, is the widening gap between investment and demand—while dollars may double, compute requirements are expanding closer to fivefold year over year, a trajectory that can only be sustained through continuous semiconductor innovation and system-level advances.

In a moment that drew attention across the room, Charlie also shared a surprise slide for Anirudh and Cadence—a Broadcom interpretation of Cadence's “three-layer cake,” where the middle layer was reimagined as orchestrating tools, underscoring the growing importance of coordination and integration across increasingly complex AI systems.

Together, Anirudh Devgan, Jensen Huang, and Charlie Kawwas reinforced a shared vision: AI’s future depends on open, scalable, and power-efficient infrastructure, tightly coupled with intelligent design automation and system-level insight. As CadenceLIVE attendees heard firsthand, the industry is entering a new phase—one defined not just by faster chips, but by autonomous design, purpose-built systems, and collaboration across the entire technology stack. Cadence’s keynote made one thing clear: that the future is already being built today, alongside its customers and partners.

Taken together, the keynotes and fireside chat traced a clear arc: from exponential semiconductor demand, to the reinvention of design through agentic systems, to the infrastructure required to sustain the next decade of AI growth. Across perspectives—from Anirudh's vision of autonomous design systems, to Charlie's infrastructure-scale lens, to Jensen's framing of AI as a tool-using, agentic workforce—the message converged on a single shift: intelligence is becoming embedded across every layer of compute.

As CadenceLIVE closed its opening program, one signal stood out clearly. The industry is moving beyond incremental acceleration into structural reinvention—where design itself becomes autonomous, systems become adaptive, and collaboration across silicon, software, and physics becomes the foundation of innovation. At the center of that transition, Cadence is positioning itself not just as a participant but as a core enabler of how the next generation of technologies will be designed, simulated, and brought to life. 

Stay Tuned for More LIVE Updates from CadenceLIVE Silicon Valley!