Insights from the 2026 CadenceLIVE in Silicon Valley

Last week we attended CadenceLIVE in Silicon Valley, where some excellent keynotes and breakout presentations took place. These sessions made one thing clear: the EDA industry is going all-in on AI. As Anirudh Devgan (President and CEO of Cadence) noted in his keynote presentation, the growth of the semiconductor industry is being driven largely by AI, with the industry predicted to be worth over $4 trillion by 2035.

Anirudh laid out what he sees as the three critical layers of AI integration: accelerated compute, simulation and optimization, and AI agents. He made the argument that these layers are interdependent, and effective AI integration cannot be achieved without all three. In turn, this is driving unprecedented demand for both engineers and compute capacity, creating a clear opening for companies that can accelerate any of these layers.

The keynote presentation from Anirudh and the CEO of NVIDIA, Jensen Huang, centered on agentic AI, but they also took the time to discuss the expected exponential token demand growth through AI tool integration, and touched on the challenges of optimizing analog designs within their design space. With the rollout expected to start in digital design verification, including optimization of verification workflows, Mach42 is primed as the unique player in enabling this change. This was also evidenced in the conversations we had at our booth.

A resonant theme from the panel ‘Closing the sim-to-real gap in physical AI’ was Design for AI ↔ AI for Design. That is, not only can we design systems to leverage AI, we can then use the AI to design better systems. This means using AI to more rapidly and easily create rich simulation environments, with agentic AI likely to be transformative for this purpose. The panel closed with the following predictions for the coming year:

• AI agents will completely transform and overhaul physical AI simulation environments

• This transformation will lead to previously unknown simulation possibilities

• Neural-network-powered physics modelling will become more prevalent

• There will be a richer asset library of real-world environment models

It’s clear that AI is still the main topic on everyone’s mind. While most of the discussions revolved around agentic and LLMs, Mach42’s focus on analog circuit modelling is evidently leading the way for the next generation of design and innovation.