What you’ll learn:
- EDA and semiconductor intellectual property (SIP) providers invest heavily in R&D to innovate and create or update products.
- EDA companies are adopting AI to improve the quality and performance of their products.
- AI EDA is an emerging area within EDA that leverages AI models as their foundational elements.
It’s likely that anyone reading Electronic Design is using some form of electronic design automation (EDA) tools, whether it’s design tools or verification software. After all, electronics begins with design.
The 40-year-old EDA product portfolio has expanded as well. It now includes the important and continuously growing category known as semiconductor IP, blocks of logic that can be used to design chips.
Still, myths persist about EDA’s value and its viability as the semiconductor industry moves through the era of AI and the slowdown of Moore’s Law.
1. Design is separate from manufacturing.
It’s true that an artificial wall was erected years ago that separated design from manufacturing, a result perhaps of siloed groups within the electronics supply chain. Engineers threw their designs over the wall, then went on to their next project, leaving manufacturing groups to sort out getting the design into full production.
Industry organizations such as SEMI are trying to change that perception. The Electronic System Design (ESD) Alliance, formerly the EDA Consortium (EDAC), was folded into SEMI with the goal of bringing design and manufacturing closer together.
2. EDA tools aren’t keeping pace with the demands of today’s complex chips.
Not true. In fact, it’s quite the contrary. Design and verification complexity continues to become more and more challenging.
Advanced processes, while allowing for increased functionality on each chip, also create new concerns for designers. Localized heating can cause localized performance differences, while reduced operating voltages reduce design margin. Heterogenous integration adds many new design concerns. The EDA industry continues to enhance tools as well as develop new tools to address concerns like these.
3. The innovation in EDA stopped long ago.
Not true. EDA companies, both large and small, invest heavily in R&D to innovate and create or update products. In reality, the percentage of revenue that’s invested by EDA companies is one of the highest in the technology industries—often higher than 30%.
EDA and SIP companies continue to invest heavily in R&D to address new challenges and opportunities related to both advanced processes and new applications. In the case of new process technology introductions, EDA companies play a critical role by working closely with the foundries to be able to deliver the tools that can deliver manufacturable designs in the new processes.
Other R&D drivers include advanced automotive and medical applications and complex new system packaging requirements.
4. Investors abandoned EDA.
In recent years, investors have funded startup and entrepreneurial companies in the EDA sector. As an example, a new category known as AI EDA is emerging with funding from venture capital firms that leverage AI models as their foundational elements. Their products range from functional verification and chip design to code and embedded development.
With more available funding, the industry should expect to see other innovative approaches announced over time.
5. It’s impossible to start an EDA company. The segment has less startup activity than it had five years ago.
The number of EDA and SIP company startups, located all over world in all facets of design and verification, continues to grow. Many are bootstrapping their company with consulting services while they build their products.
The global push to build out the semiconductor supply chain is providing opportunities for entrepreneurs to develop automation tools and SIP based on local demand and/or specialized requirements. The nature of the EDA business is that it’s constantly chasing moving targets in terms of process technology, advanced packaging, and verification of complex systems, including heterogeneous integration. This leads to a constant stream of new challenges.
6. EDA hasn’t had a significant breakthrough in many years and is thwarted by the emergence of chiplets and heterogenous design.
Leading-edge designs that now include chiplets and heterogenous integration have always challenged designers. While heterogenous integration introduces many new design concerns, the fact that many new complex products successfully include it suggests the EDA industry is keeping up with the challenges.
7. Verification problems are outpacing the speed of the tools, placing an increasing burden on verification methodologies.
Verification is a huge challenge, but one that the EDA companies are tackling. Hardware emulation, for example, is now a foundational system design tool in the verification flow for hardware-software co-design and co-verification and prototyping and software bring-up. That category is now called hardware-assisted verification for its ability to validate and verify more than 40 billion+ gate designs.
8. EDA is missing the AI tsunami.
Jay Vleeschhouwer, Managing Director of Griffin Securities, was quoted as saying:
“The answer must be no. While difficult to quantify, the contribution to the EDA companies is emblematic of this phenomenon for both machine learning and AI. Perhaps ML is the more relevant, having more to do with pattern recognition. EDA tools deal with massively complex patterns that lend themselves to massive computation. Clearly semiconductor design lends itself to these kinds of techniques.”
EDA companies are adopting AI to improve the quality and performance of their products. Announcements from EDA companies large and small about support for AI chips for their software and hardware tools are a regular phenomenon. While the EDA industry is not a provider of AI large language models (LLMs), it employs them extensively within their tools and flows to improve performance and accuracy. (See Myth #4 for additional information.)
9. Cloud-based design tools are used sparingly or not at all. Developers don’t want their tools or designs in the cloud.
In past years, there was a reluctance to migrate designs to the cloud, but that has changed. Increasingly, advanced EDA tools are available in the cloud, and the ability to rapidly adjust compute resources as design needs change has made cloud-based designs increasingly preferable.
For example, verification is one of the most resource-intensive parts of the chip design flow. Cloud-based access to tools gives engineering groups instant access to ramp up or down the amount of verification compute resources required at any given time.
10. EDA is quickly aging out.
Like many industries, long-time employees, innovators, and leaders are retiring after kick-starting and growing the industry into the critical industry it is today. This has opened many new opportunities for new thinkers and leaders.
STEM programs are inspiring high-school and middle-school students to learn about the exciting and challenging projects engineers and mathematicians work on in their curriculum.
Universities with Electrical Engineering and Computer Science departments are reevaluating their course curriculum to attract more talented students interested in the semiconductor field. Results are encouraging and ongoing.
11. EDA is just a small part of the overall semiconductor industry that’s headed toward $1 trillion.
EDA may be small in terms of yearly revenue—about $20 billion. When measured in terms of its impact on the huge semiconductor industry, it’s a critical segment. Without automation tools for the design and verification of today’s complex semiconductor chips and systems, there would be no new advanced processes, no new leading-edge designs and product innovation would be slowed and industry growth would turn negative.
This is why the tagline for the ESD Alliance—"Electronics begins with design”—is more than a tagline. It’s the reality of the role that the industry plays in leading the industry to the $1 trillion mark.