Bronco AI at DAC 2026, What a Difference a Year Makes

During my time at DAC 2026 I felt like I was attending two events. One was the traditional DAC where approaches to designing complex chips using the best tools and methodologies were the focus. There was another focus for the conference. This one was like the traditional DAC with an important difference. Here, the discussion centered on agentic methodologies that could work with designers to deliver results with far greater speed and quality. The conversation moved up a level from tools.

While tools were still very important, the way those tools are run and how the results are interpreted was the focus. Discussions around this up-leveled methodology occurred at traditional companies who have been at DAC for a long time as well as newcomers who recently arrived on the scene. In this post, I’ll focus on one of those new companies. Bronco AI was at DAC last year and again this year. The progress I saw in just one year was astounding. Here is some more detail behind Bronco AI at DAC 2026, and what a difference a year makes.

A View from the Top

I had the opportunity to speak with David Zhi LuoZhang, co-founder and CEO of Bronco AI. David has an interesting career path. Before starting Bronco, David built AI fighter jets at Shield AI. He was also on an Accellera panel that I chaired at DAC. You can learn more about that panel here. 

I discussed what happened at Bronco between DAC last year and the present day with David.  Like many things that are related to AI, there was a definite feeling of exponential change in our discussion. David explained that Bronco started with design verification/debug. This task represents a large amount of time and effort for any design project, so a good high-impact contribution.

At that time, Bronco was working with some large companies addressing very difficult debug problems. He explained that these problems were well beyond what the customer’s best engineer could solve, even when advanced LLMs were added. By tackling this problem first, it set up Bronco for success in the broader market. He went on to say that addressing these debug problems required a very deep and detailed understanding of the design.

Activities such as reading the code and reading the specification to achieve the required depth of knowledge were involved. Once that level of capability was achieved, activities such as testbench generation and adding new features to the design were within reach. Bronco achieved these capabilities with a combination of the best AI, the best tools for AI, and the best system to learn over time. The “best tools for AI” comment requires some more detail. More on that in a moment.

All of this set the stage for a new, more expanded and more powerful Bronco AI. There is one more ingredient to mention. Last year, use of AI agents was just beginning. The model was very task/transaction oriented. “Debug this problem” was a popular one. Today, the use model has expanded to goal-driven, proactive agents.

David went on to explain the impact of this new approach. Using the old approach, AI was used to dramatically accelerate the debug of a problem, once the problem was found. Weeks of debug could be collapsed to minutes/hours. But it still took months of simulating test vectors to find the problem to debug.

Today, Bronco can leverage its learning capabilities and autonomous orchestration to run continuous tests and find problems much sooner.  The tools autonomously run tests, analyze results, generate more tests and repeat the process. David doesn’t expect design teams to stop all the constrained random testing that is underway with the latest design.

But running Bronco alongside that activity will start to find many problems faster, some that may never have been found with current approaches. Over time, this will create a fundamental shift in the way design is done to deliver faster, higher quality results. Said another way, keep your current flow and add an additional flow that is ready out-of-the-box to provide fast help and deep insights. Keep the current flow but begin to reduce the schedule as you gain confidence in the new flow. Given the exponential trajectory of AI adoption, I expect next year’s DAC will be very different.

I revisited the “best tools for AI” comment with David. He explained that Bronco has developed various EDA tools that integrate well with Bronco’s AI. These tools don’t replace the signoff tools that are needed for tapeout. Instead, they facilitate massive exploration and analysis of the design without the need to check out massive numbers of licenses for the signoff tools. This provides a cost-effective, convergent design flow that also accelerates the tapeout process.

David described a design flow that includes autonomous AI agents that run tests, find and fix bugs, generate new code and new features, and apply architectural decisions in one cohesive loop. He pointed out that the impact of such a flow is greater than the impact of each individual task. Clearly this is a vision of the future, one that is backed up by a lot of achievements Bronco has already made.

David explained that last year, most customers were just starting with AI. Processes and infrastructure were not there yet. This year, customers are far more aggressive. Bronco is in production at very large chip companies. Its technology has been successfully deployed for designs that include data center scale AI accelerators at leading nodes, complex low power IoT SoCs, and high-speed interconnect. What a difference a year makes.

To Learn More

An AI design revolution is coming, and Bronco is one of the companies that is leading the way. You can learn more about Bronco AI on SemiWiki here, and dig deeper on the company’s website, where you can also request a demo.

There is also a webinar on Wednesday, September 9, 2026, from 9:00 AM – 10:00 AM PDT. The title is: Cutting Full-Chip SoC Debug from Days to Minutes with AI. David will walk through a live demonstration on a real design and take questions on how customers deploy, evaluate, and scale Bronco within their silicon projects. You can register for this important webinar here.

And that’s a summary of Bronco AI at DAC 2026, and what a difference a year makes.

Also Read:

Bronco AI Webinar: Full-Chip SoC Debug in 15 Minutes

Bronco Debug Stress Tested Measures Up

Verification Futures with Bronco AI Agents for DV Debug