Roy Jewell | Founder and CEO | Palma Ceia SemiDesign
Life often comes full circle. The same can be said of careers. And that’s been the case for Roy Jewell, founder and CEO of Palma Ceia SemiDesign, an up-and-coming startup to watch. He began his professional life at one of the world’s largest semiconductor companies and now, more than three decades later, he is back in semis, leading a startup on the verge of releasing its first chips. In between he left semis, moved to Silicon Valley and took two companies public on Nasdaq. In this interview Roy shares some of what his past taught him, the state of semiconductors today, and what he sees for the future. Spoiler alert: He’s optimistic.
You studied physics in college and grad school. Did you have your eye on the semiconductor industry at that point?
Mostly I had my eye on education. My father was a teacher and so of course encouraged his kids to go to college, and early on I was drawn to physics.
As I was finishing my master’s I was, like everyone finishing school, looking for employment. It was not a strong job market, not much hiring going on. And even though a physics background is actually pretty handy if you’re working in semis, it doesn’t come with any real specific skills that those companies look for. You’re more of a scientific generalist. I came across Texas Instruments in my job searching and applied, and after getting through the interview gauntlet I was offered a job in TI’s Design Automation Division. It turned out to be a great opportunity. I got into the semiconductor industry with the market leader and was able to learn from there.
And in fact they gave me a large amount of autonomy and responsibility. TI allowed me to build up a process modeling team, which is how I basically started my career.
Part of my job involved working with Stanford University, where a professor, Bob Dutton, was driving what was probably the world’s top research in TCAD, short for Technology CAD. That’s a segment of EDA, the electronic design automation software industry. TCAD software models semiconductor fabrication and semiconductor device operation, and it’s actually vital to the semiconductor industry. It involves an understanding of fundamental physics and chemistry, so my background actually made me ideally suited to liaising with Bob and his group, probably more so than most other people at TI. So having that physics education was sort of dumb luck, I guess.
But then you left the semi world.
Sort of. Bob was on the board of Technology Modeling Associates, a startup software company founded by some of his former students. They started out by offering consulting to commercial companies, one of which was TI, that had licensed the technology from Stanford, helping them use it. Then they licensed the Stanford technology themselves and productized it.
It was about this time that Bob introduced me to the TMA founders. After a while they convinced me to join them, so I moved to California and got into the startup life.
How did that experience compare to working at TI?
Different in many ways – probably not a surprise. It was of course a much smaller company, so we had to operate with more limited resources, and at a faster pace. It’s a different business, software versus hardware. The sales process is very different, more hands on in some ways. And for me personally it was an adjustment, since I was no longer in the business of making semiconductors but selling to the people who do.
At one point we — like small companies often do — found ourselves in a cash crunch. In those days some of the most successful chip makers were in Korea, so I started spending a lot of time there, courting business with the big companies. I decided it would be a good idea for me to learn Korean, which went a long way in forging strong relationships there. Eventually we closed some large deals and got back on solid ground financially.
From there we eventually grew the company beyond the original technology and were able to enter some new markets, giving us greater opportunity. We took it public on Nasdaq in 1996, and a few years later we sold TMA to Avanti, another EDA company. I stayed there for about a year and decided it was time to move on and do something else.
What was that something else?
Initially, I just wanted to take some time off. After a while I was contacted by people at Magma Design Automation, an early-stage next-generation EDA company founded by Rajeev Madhavan, the CEO. Rajeev and I got to know each other and I joined Magma as chief operating officer.
In those days we were competing directly with Synopsys and Cadence, trying to get a foothold in place & route, the design tools semi companies use to lay out chips. It was a pretty competitive situation.
Ultimately, we were very successful in landing customers that were working on application processors for wireless applications, companies like TI, Nvidia, Qualcomm and Broadcom. We took Magma through an IPO in 2001, and eventually grew it to the No. 4 EDA company, behind Synopsys, Cadence and Mentor, and Synopsys acquired us in 2012. I stayed with Synopsys for a short while to help with the integration, and then I decided I wanted to go out and start another company.
Another EDA company?
No, I knew I did not want to go into EDA again. I wanted to build on the relationships I had in EDA, but I wanted to move into something more closely related to chips.
Sounds like semiconductors were beckoning.
Again, sort of. That’s when we formed Palma Ceia, initially as an analog/RF IP company. Which generated a number of pieces of IP – high-performance data converters and transceivers for major semiconductor companies.
After a while we decided a better model was to move from delivering IP to delivering a chipset. So we made a pivot in early 2018.
You were finally getting back to chips. What drove that change?
A few things. We decided it would be easier to leverage all that IP into chipsets – just a better business model. And we saw changes in the market, much of which was related to the Internet of Things. We didn’t set out to be an IoT company, there were already lots of those.
There are a lot of Wi-Fi solutions for PCs and mobile handsets. But what’s driving the business now is the emergence of IoT as a viable market segment – you don’t need the same level of throughput and speed that you do for these other applications, which are predominantly focused on uses like streaming video for consumers. These are customers looking for additional sources besides the major suppliers, which have historically been focused on other segments, such as mobile handsets and laptops.
What do you see driving the semiconductor opportunity in this space?
First, we see very interesting applications springing up, enabled by IoT technology. In particular the Industrial IoT space – Industry 4.0 – is very interesting and offers a lot of opportunity. Smart factory is a term nobody was using until a few years ago and now everybody talks about it.
The surveillance and safety market is going to be huge. This has particular importance in China, where industrial accidents are not just a problem for the company, but for management. In fact, government and company officials are often held personally accountable for on-the-job accidents that injure or kill workers. They can be denied promotion for years and their career can be ruined by the accident. With that kind of personal risk, you can see why they’d want to use technology as much as they can to prevent accidents. I think agriscience is going to be an interesting growth driver, with connected sensors helping famers to know when to plant, irrigate, etc. As you can see there are a lot of applications coming to market and driving opportunities.
Second, gaps in technology have created opportunity. For instance, current connectivity capabilities are not sufficient to serve those market opportunities I was mentioning. We see a new segment like surveillance and safety opening up, but it requires that devices operate at ranges that traditional technology can’t meet. Power requirements exceed standard Wi-Fi capabilities, and IoT device infrastructure redesign or redeployment is impractical. And going back to the agriscience example, a gating factor is distance – farms can be big places. The new technologies that enable longer-range connectivity will be vital.
Any interesting examples you can tell us about?
Sure, let’s talk about an example in medical technology. With a heart monitor, a grandmother might be in a building, and her heart monitor is connected via Wi-Fi 6. But then she walks out of the building and after about 100 feet she’s too far from the router and the Wi-Fi 6 connection drops. Now her heart monitor needs to find another way to connect to the cloud, so it has to switch over to another connectivity technology — like Wi-Fi HaLow or NB-IoT — which can operate at longer range.
Semi companies are getting pretty creative in developing chips for IoT to address use cases like this. Combo chips that serve multiple standards need to be provided, simply because of the way people live. Providers might trade off bandwidth, data rates and propagation distance. Use cases are becoming more complex, stretching the capabilities of traditional Wi-Fi. And that creates opportunity.
I’m seeing some great use cases which until recently weren’t possible, and now that they’re here it will be hard to live without them. And delivering the technology to make them work is quite an opportunity.
These are great opportunities. What sets Palma Ceia apart?
Frankly we’re the only company that has solutions for Wi-Fi 6, Wi-Fi HaLow and NB-IoT. It positions us well for those applications.
What are major issues in the semiconductor industry generally today?
For the industry, fab capacity is tight right now. Getting things like substrates is a challenge. One of the major constraints we have right now is in operating capital — an issue I’m happy to say we have largely solved — and being able to get adequate fab access.
At Palma Ceia we minimize the risk in our business by working with the leading fab, TSMC, for all of our manufacturing today.
Any special impact from the Covid pandemic?
In the early phase it had everyone looking at their businesses and trying to figure out what had to change. But the lead times in semis are so long you can’t afford to wait very long. You’ll miss market opportunities, or lose your manufacturing slots at the foundries, that kind of thing. I think everybody figured out how to survive in the short term so they could keep operating for the long term.
It definitely hampered me as far as keeping in touch with customers and partners, because so much of those relationships depends on seeing people in person. And nobody could travel. I finally just decided, no matter how inconvenient, I had to go to China to meet with key partners and customers. I got the vaccine as soon as I could and flew right after getting my second shot. Then I had to be in quarantine for 2 weeks. It was pretty restrictive – I couldn’t leave the facility or see visitors. No wine or meat, and I drank plenty of water. It was like going to a spa for a cleanse. Just without the spa.
I was there for a few weeks after the quarantine. A long trip but it was necessary.
Anything in particular you have learned that you can share?
If you work long enough you learn a lot, so yes. For one thing, everything you do needs to start with the customer – if you aren’t doing something that helps a customer, you probably won’t be in business very long.
Too often you learn because you came too close to disaster. One story I have isn’t even mine, it’s from a company I know well, a really big company I won’t name. They say cash is king, and that’s certainly true for startups, but it’s also true for large companies. And nobody would ever have imagined this company having financial trouble. But a few years ago, they hit a slump, and they were sure they couldn’t make payroll. As a result, that company’s management, in its own words, went through a terrifying experience. They managed to scrape by and make payroll, but it was really close. And that company, again in its own words, is now “damn cheap” when it comes to spending because they don’t want to be in that situation again. So that’s an example where you can learn from someone else’s experience.
One lesson I learned on my own: speak and drink in the customer’s native language.
As we wrap up let me ask – what do you see in the future for semiconductors?
I wouldn’t have gotten back into the business if I wasn’t optimistic. Let’s go back to the opportunities I mentioned earlier — connected medical devices, industrial IoT, safety & surveillance, etc. They’re all driven by new product applications that are enabled by technology. And without the semiconductor industry that just wouldn’t happen.
This is an industry that goes through cycles on a pretty regular basis. I’m often surprised, and a little amused, at how the naysayers come out during the downturns and predict doom-and-gloom outcomes for the business. But it always comes back, usually stronger than before. I see no reason that will stop.