EDA
System
IP
SolvNetPlus
Training & Education
If you stop learning, you fall behind.
For semiconductor companies and professionals alike, this has never been more true. Artificial intelligence (AI) is rewriting the rules of the game.
I’ve spent many years in the semiconductor industry, and my career has covered nearly every facet of systems design — spanning hardware, software, chips, electronic design automation (EDA), and now AI. At every turn, I’ve had to seize the opportunity to expand my knowledge and adapt.
The only difference with AI is the speed at which it is evolving. I’ve never seen a technology progress so quickly.
Explore AI-Driven Design with Synopsys DSO.ai
Leverage Synopsys Ai to enhance power and performance. James Chuang describes how to achieve new productivity levels.
From automation to optimization
Just as continual learning and reinvention have shaped my and many other colleague’s careers, they are also woven into the fabric of Synopsys.
From the earliest days of EDA to the new frontier of agentic AI and beyond, we harness automation to help engineers adapt to perpetual change. And their most pressing challenges — growing complexity, a scarcity of talent, shrinking time-to-market windows — are more pronounced than ever before. Today, chip designers are dealing with sub-micron process technologies, 2.5D and 3D hardware integration, and significant growth in verification cycles, with less time and margin for error.
We recognized early on that AI could help, and we started infusing it into our EDA tools.
Our first foray was design space optimization, and we deployed DSO.ai in 2018. It was an industry first, allowing engineers to explore large design spaces, optimize multiple power, performance, and area (PPA) objectives simultaneously, assess many design recipe combinations, and automatically tune flows and tool settings.
Building on that success, we expanded AI integration across our product portfolio, including implementation, verification, and testing tools. AI became an engine quietly working behind the scenes, dramatically impacting what was achievable in chip design.
AI as assistant
In 2022, the launch of ChatGPT suddenly shifted the realm of possibility. Instead of AI being an optimizer buried deep within software, it could potentially interpret and directly interact with engineers.
At Synopsys, we quickly assembled a team to explore the merits of generative AI and how we could leverage it in our products. And we started working with our longtime partners, Microsoft and NVIDIA, to collaborate on solutions based on their models, infrastructure, and AI services.
It was another exciting phase of learning and reinvention. For the first time, we began envisioning AI as different forms of assistants.
We built a Knowledge Assistant that allows engineers to ask broad questions about our tools, technology, or the design they’re working on. We then expanded the concept with a Workflow Assistant that analyzes scripts and suggests improvements to reduce errors — all based on the design context, not generic snippets. Both of these AI assistants provided a massive productivity boost.
We have since moved into something truly generative: the ability to take a design specification and automatically generate the register-transfer level (RTL) and associated verification tests.
Each phase built upon its predecessors, putting a broad set of generative AI capabilities into the hands of our customers.
Raja Tabet (far right) at Mobile World Congress 2025
AI as colleague
Now we are stepping into yet another phase of reinvention: agentic AI. These AI systems set goals, reason about complex environments, plan, and take actions across tools and data to drive outcomes. Agents work alongside human engineers, who remain in charge of high‑value decisions around architecture, tradeoffs, and risk.
For organizations facing exploding semiconductor design complexity and a shortage of talent, agentic AI is a game changer.
We envision that within the next 12 to 24 months, organizations will begin creating an agentic AI workforce. Our customers will lease or buy from us “AgentEngineers” with different personas — think of a digital implementation agent, a verification agent, an analog agent. These specialized AI colleagues will run more experiments in parallel, generate and triage tests, propose fixes, and keep flows moving.
AgentEngineers will help organizations address their capacity limitations and time-to-market challenges, tame the complexities of chip design, and eliminate the engineering resource bottlenecks they have today.
Frankly, we see unprecedented opportunities available to our customers.
Reinventing Synopsys from the inside
My team at Synopsys (the Engineering Excellence Group) sits at the center of this transformation, building the core foundational technology and platforms for assistive, generative, and agentic AI solutions.
We also apply these same technologies internally, optimizing our software and hardware development processes and all of our corporate functions to deliver the efficiency and productivity we need.
Cultivating this experience is critical for establishing credibility. If customers are to trust AI‑driven systems in their mission‑critical design flows, we must first hold ourselves to the same standard. Along the way, we confront hard questions about robustness, transparency, and change management that inform the platforms and products we ship.
Of course, we continue to work closely with partners like Microsoft and NVIDIA. In an agentic world, an ecosystem approach is essential. No single model or vendor will prevail. Successful systems will combine multiple models, specialized tools, proprietary and open data, and robust orchestration frameworks — all underpinned by trust, transparency, and strong change management.
What comes next
I am proud of what we’ve accomplished over the past seven-plus years, but this AI journey has only just started. At Synopsys, we’re accelerating the pace of investment to stay at the forefront and continue what we’ve always done: look for opportunities to optimize and automate the work of engineers.
With more complex nodes, more stacked designs, and more ambitious applications, the pressures will only increase. And the gap between what human teams can do and what the market expects will continue to widen.
Closing that gap requires a step‑function improvement in automation and redefining how engineers spend their time. With agentic AI, engineers will hand off the toil of chip design and focus on higher-order, system-level thinking and decision making — the architectural work that truly differentiates their products.
As AI moves from tool to teammate, those of us leading engineering organizations must be prepared to rethink roles and workflows. To me, that prospect feels like the natural next chapter for engineering excellence.
The future of AI in chip design is the next big opportunity for all of us — as professionals and as an industry — to reinvent ourselves once again.
Continue Reading
Designing the Future: How 3DIC Compiler Is Powering Breakthroughs Across the MultiDie Design Landscape
Engineering for the Exponential Complexity of Physical AI
