Diverging Viewpoints

By Ed Sperling
The raw materials of semiconductor design include smart, well-trained people and money to fund good ideas from those people, whose backgrounds typically come from engineering, math, physics, computer science, materials science and sometimes even chemistry.

While many experts, executives, and industry groups have been sounding the alarm in recent years about everything from lack of students with the right backgrounds to a shortage of work visas to insufficient startup capital, that sense of worry is far from universal.

System-Level Design spoke with five of the leading EDA and IP industry leaders on these subjects. Their comments follow, along with some observations and future considerations.

1. Lip-Bu Tan, president and CEO of Cadence
Education: Being a VC, I’m always looking for top talent. It’s critical. Last year, we started aggressively on college recruiting. We need fresh minds and ideas. We’ve done pretty well worldwide, and we will continue to do that. But we also need to extend EDA over the next 20 to 30 years, and we need to reach out to do that. I’m very passionate about education. We do need more engineers and talent because we have a lot of innovation in the industry, and we need that to continue. At the college level, immigration and local schools have to address how to recruit the best and brightest to get into the semiconductor industry. How do we address that and make it sexy to be in the semiconductor industry? We need to get people excited about the end product, not another tool or PC board. That’s why we have to reach out to the top professors. I’m on the advisory board at UC Berkeley and MIT, and I’m a trustee at CMU (Carnegie Mellon University). They have to tell students the place to go isn’t Google or Facebook. It’s a company like Cadence.

Startups: If you add up all the cash at semiconductor companies, there is $400 billion. They’re all struggling to grow the top line, and anyone running these companies knows it’s hard to do disruptive innovation on the inside. They’re always looking outside to buy companies. But if the VC trend is not reversed, then 10 years from now they’ll have fewer choices to buy. It’s important to see more innovation and startups being funded. I’m encouraging some of my friends to go back and fund semis. Some successes will help, like Inphi, but we need more. In China and India, the governments are investing in startups. I would like to see a reversal of the trend that those countries have the most startup activity.

2. Aart de Geus, chairman and co-CEO of Synopsys
Education: The number of well-trained people available is not an issue if you look at this globally. But that’s the point—it’s global, not just local.

Startups: To say that we need more startups to continue innovating implies that large companies can’t innovate. They can and they do. What’s happened is that innovation has moved from points to systems. It’s now at the intersection of many things. That’s radically different than even two years ago. There are two dynamics working here. One involves the dynamics down to the physics level. Features get smaller even though some of the chips are getting very big. The other dynamic is up to function. Technologically and economically, that’s our job to make it all work. Startups can’t do that.

3. Wally Rhines, Chairman and CEO of Mentor Graphics
Education: We’re not producing nearly as many EDA-trained graduates as in the past. We need more emphasis on the university infrastructure. I oversee a board that has given $30 million to universities, and Mentor has its own programs. This is very important.

Startups: Funding for EDA research has not kept up. VC funding we can track, and last year there were about 38 VC-funded EDA startups. Mostly those were in the United States. That doesn’t count the EDA startups that are done with a couple people without using VC funds. Still, the number of startups in this area has fallen, and there certainly are not as many VC-funded fabless companies. There are more in Asia than in the United States.

4. Charles Janac, chairman and CEO of Arteris
Education: China already looks like Silicon Valley in the 1980s. There is a lot of innovation and dynamics. The world hates a vacuum. Someone always fills it.

Startups: The key to startups is that you have to have enough differentiation and value, and then still have enough money to sustain R&D. If not, the customers buy from the big players. So you have to be cheaper than free—meaning you add real value—and over the life of an SoC, you need to have 10 times payback on your business model. It takes 10 years to build a solution in IP, and that solution is very capital intensive to prove in silicon. IP markets are very concentrated. The value is from the best technical solution, and everything has to connect to one infrastructure.

5. Simon Segars, president of ARM (and CEO as of July 1)
Education: We’re not seeing any lack of talent at the moment, but it is coming from different places than in the past—China, India and to some extent Europe.

Startups: There is a clear lack of VC money going into startups, and it’s understandable. It’s so expensive to build a chip that not many companies can afford to do it successfully. But there is an enormous market for lower-cost devices in developing countries, like the $50 table being sold in Thailand for schools.

6. Naveed Sherwani, president and CEO of Open-Silicon
Education: The problem seems to be in United States. The reason is that students can see that there are better returns and acknowledgement of other areas and they are drawn towards them. The semi industry needs to run a grass-roots campaign to recruit students, arrange internships, recognize and support professors who teach and research, and work with local and federal agencies to promote semiconductor education, research and training. It is crucial for our future.

Startups: There are really smaller number of start ups. I believe that it is already having a very negative impact on innovation. There are many reasons why there is a decrease in startups, but low ROI is the primary one. Rather than pursuing true innovation, many folks went for quick incremental ideas, which led to low ROI. We need to get focused on ideas that will transform our business. A second reason is that we are in hardware-software innovation cycle, where ROI for software and system concepts is much higher than hardware.

Conclusions
Big companies always will be able to attract top talent. They have globally known brands and deep pockets. However, if the number of engineering and science students drops they may have to pay more for qualified people, which also means they will be able to afford fewer than in the past.

Startups, in contrast, will have to fight for their talent, leveraging the promise of either a sale or an IPO and a share in the future once a sale or public offering takes place. But startups also provide huge opportunities for engineers and scientists to reach into many areas that might be off limits to them inside large corporations, adding new perspective and drawing out new ideas through cross-pollination. While that theoretically is possible in large companies, departmentalization and sheer physical separation make it more difficult. And though it’s important for large companies to innovate, the record of acquisitions suggests that it’s faster to buy that innovation than to grow it.

Where there is money available, schools will train students, and students will flock to professions where they can be assured of a future and professional growth. Youth unemployment is high in many countries at the moment, which can be a key driver for industries where there is opportunity. In places such as India and China, the funding is available both for educational research, as well as to fund startups after students graduate. In the United States and Europe, most of that funding has come from private sources, but there is heavy lobbying to become more competitive everywhere and on every level.

With increasing emphasis on semiconductors, the rise of the Internet of Things, and a focus on solving some really difficult technology problems while keeping costs under control, there will be no shortage of opportunities and ideas. The big question is who will leverage these opportunities—and where.