Market Realities

Speculation about EDA’s demise or absorption into other industries is greatly exaggerated. Here are five reasons why.


The speculation about EDA’s future—will it consolidate, will it be incorporated into large IDMs or foundries—has surfaced again.

The reason this time is that EDA is in a retrenchment period as the semiconductor industry grapples with increasing complexity, multiple options ranging from multi-patterning to stacked die to more third-party IP, and the rising cost of complex SoCs at the most advanced process nodes. Unless the EDA industry does something very wrong, though, here are five reasons why it doesn’t have a whole lot to worry about.

1. The continued fragmentation of end markets means that big chip makers likely will be the ones consolidating, not EDA. The Internet of Things is a good example of how this will play out. There will be many, many things that are connected to the Internet. Cisco has put the number at 50 billion things, although that’s about as trustworthy as a weather forecast three years in the future. Still, the number is likely to be very, very large, and most of those won’t be made in quantities of 100 million. They may not even reach 10,000.

2. While the speculation is that software will overtake hardware, the reality is that the software industry is even more fragmented than the hardware industry. And aside from some core applications, such as Microsoft Office, Adobe Photoshop, and some browser and social networking software, as well as enterprise applications such as Oracle databases, none of this can work without semiconductors. In fact, semiconductors are the core building block of almost everything electronic, and increasingly they’re replacing mechanical parts in automobiles, appliances and health care. The more likely scenario is for hardware companies to develop much more efficient software, with significant improvements in both performance and power, rather than the other way around.

3. EDA companies continue to branch out into a variety of new markets ranging from RTOSes and thermal analysis software (Mentor Graphics) to IP (Synopsys and Cadence), and their tools are being used well beyond just the semiconductor. In fact, it’s hard to actually come up with a label for what these companies will evolve into as they transform themselves to take advantage of new opportunities in much larger systems.

4. Consolidation typically happens when markets shrink, going from horizontal to nearly vertical, or when they broaden from vertical to horizontal and economies of scale are required. These shifts frequently occur when over-enthusiastic funding for new technology and opportunity collides with market realities, followed by a price war and commoditization that leaves only a handful of players still standing. EDA already has gone through a number of these exercises, and its growth has now stabilized to levels consistent with a mature industry. The drop in the number of lawsuits between companies is a good indication of just how far this maturation has progressed.

5. One of the biggest threats to technology is always smaller, faster and cheaper. In EDA, the technology has become so complex, and the learning curve so steep, that developing tools and re-learning how to use them are huge barriers to entry for would-be competitors. On top of that, this isn’t a one-tool-solves-all market anymore. Point tools are important, but integrated suites are vital. It’s much more difficult for startups to make inroads in core EDA markets, and increasingly it’s difficult to even stake out a significant piece of the IP world.

On top of all of these factors, EDA is a tough industry to survive in. It’s complex, the returns are often less than you’d expect from something this difficult, and the payoff isn’t always within the time frame anyone would expect. Moreover, the value is to reach many customers, not just one. Given all of these factors, EDA will remain an interesting business to work in, but a bad risk for massive consolidation and acquisition.

—Ed Sperling

Leave a Reply

(Note: This name will be displayed publicly)