One-On-One: Lip-Bu Tan

Lip-Bu Tan, CEO of Cadence, sat down with Semiconductor Engineering to talk about the impact of massive increases in data across a variety of industries, the growing need for computational software, and the potential implications of U.S.-China relations. What follows are excerpts of that discussion.

SE: What do you see as the biggest change for the chip industry?

Tan: We’re in our fifth generation of technology. There is 5G, AI/ML, the hyperscale cloud, autonomous driving, and a huge industrial 4.0 digital transformation. But even more important right now, it’s all about data. We have entered a data-centric era, where 90% of the data was generated in the last two years. Of that, 80% is unstructured data—that’s video and audio—and its growth is explosive. Less than 2% of all that data is being analyzed. So the next big thing is going to be about data analytics and how you can use data effectively to drive decision-making in different vertical industries. It’s a huge transformation.

SE: If only 2% of the data is analyzed, is there still 98% of the data that has value to it? Or is 90% of that just garbage?

Tan: The problem now is how can you automate that data labeling so it becomes useful? It’s not that 98% of the data isn’t useful. But we don’t quite know how to use it because it’s not being labeled correctly. Another part of this is AI/ML. How do you leverage this in some of the vertical markets? One of the areas I’m really, really excited about is using this for drug discovery. We’re just at the beginning of using data for improving the medical industry. Also, with Industrial 4.0, there is so much data that can be used to drive smart energy management and industrial automation. How do you make all of this more useful? We’re just at the beginning stages. What’s really exciting is that semiconductors are at the core, and design software will play a huge role here.

SE: Another area where data analysis is critical is in the automotive sector. Where do you see autonomous driving heading? We seem to be a very long way from telling a car to take us somewhere at the push of a button.

Tan: There are quite a few challenges in terms of the technology required to get there. Driver assistance is happening, and on the freeway where the lanes are clearly defined, that’s okay, though you still need to be cautious. Autonomous parking is working, too. But in terms of a fully automated vehicle where you tell it your destination and then the car takes you there without any intervention so you can just focus on your work—I don’t think that is happening anytime soon.

SE: How do you see things shaking out with China?

Tan: China is a very big market, and they’re determined to be self-sufficient, which is why there are so many IPOs happening there. A lot of money is pouring in, and the semiconductor industry there is one of the hottest for investments. The U.S. needs to invest in semiconductor R&D and manufacturing, and the Biden administration’s plan to invest $50B in the semi industry is a welcome move in that direction. Also, because there’s a big local market, China seems to be building out an alternative supply chain to reduce its dependence on U.S. companies, which can lead to big opportunity losses for U.S. companies. In the long run, hopefully the two nations will find common ground where we can compete and still collaborate to solve some of the world’s problems, rather than each region trying to figure things out in our own ways. That would be good for everybody.

SE: That also is increasing the demand for talent. What impact are you seeing?

Tan: It’s a real challenge for all high-tech companies to recruit talent. We have a university program where we engage with professors and thought leaders, and we try to recruit the best students. It’s not easy because you have to compete with Facebook, Google, Snowflake, and others. People tell me that we should start looking for talent at the high school level now. We have to recruit talent globally. In India, China, and Europe, there are some really good schools. We have a very active university program. Last year, we had endowed professors at Berkeley, Stanford, and CMU (Carnegie Mellon), and recently MIT. We also have a big effort around diversity and inclusion. In addition, we have a very active internship program. Internships are the best way for us to recruit talent, because we get to know these individuals, and they work on important projects with us. That allows us to select the very best amongst them. They also become our ambassadors in their universities as they share the good news with their peers and friends.

SE: So, does the cost per employee now go up because you have to be more competitive and do more training than you did in the past?

Tan: Yes. And we have a very active program to train our workforce. Training is an absolute must. Over the last year, we have been doing that virtually, and we try to get people involved in specific projects so they can have hands-on experiences. The most important thing is to create an environment for people so they can quickly ramp to be productive and effective.

SE: One of the big issues everyone is wrestling with now is security. The SolarWinds attack is just one example. Is there an opportunity for tooling for security, building it into the verification and debugging process, and making it a core part of chip development?

Tan: Security is a very important topic, and we pay a lot of attention to that. But with security, there is no endpoint. There is always some new hacking technique that emerges. We have to be constantly prepared. The recent attack is just the tip of the iceberg. You always have to keep track of new vulnerabilities, and we are always seeking additional ways to protect our source code and the software we develop. We put a lot of effort into this, but it’s a never-ending process.

SE: This is basically shifting right, where it becomes part of silicon lifecycle management, right? Everything is now being connected, and all of it is expected to last longer.

Tan: Yes, and this is an issue that I discuss with my executive team quite often. For an automotive or data center component, whatever the design or whatever the field, if it fails, you need to shut down that system. In the data center, this could cost billions of dollars in lost revenue. So, whatever the silicon design is, they need to have enough agents to really know what’s going on so they can do something from a preventive perspective. Even on the design side, they can do more provisioning to address this, as it’s a huge concern. So, should Cadence do something to address the issue and provide a solution? Absolutely. And it’s broader than just protecting the device. For example, we have a partnership with Green Hills Software, as well as an investment in them, and they focus on making more secure embedded software and devices.

SE: Given all of that, what is Cadence these days. Is it an EDA company? An IP company?

Tan: We are a computational software company that enables Intelligent System Design. The foundation of our strategy is basically core EDA and IP, which we tie to design excellence. We’re incorporating parallelism, AI, machine learning in our tools and are enabling the move to the cloud with cloud-native tools. System analysis has become very important in terms of EM and power. Anything we can do to help to get the power down and utilize some of the new techniques involving system analysis of physical effects is necessary. We’re quietly building that up, just like our IP business, along with security. So, we call it intelligent systems. The next level is basically pervasive intelligence, applying AI/machine learning to some of the verticals, like the data center/cloud and autonomous driving. We’re also starting to look into the medical field, using some of the AI/ML expertise to address that. We’re expanding more into the system design realm.

SE: That also requires moving up a couple levels of abstraction, right? You can’t do it everything at once, so now you have to examine the profit centers for your business and place your bets.

Tan: We have an opportunity, and we listen to our customers about the market requirements. And accordingly, we prioritize, and then over time, make the changes. We have a culture of innovation and a deep pipeline now, but we cannot do everything organically. We also have a very proactive acquisition strategy, headed by Nimish [Modi, senior vice president of marketing and business development], to drive the right acquisitions and accelerate our strategy. The market has changed a lot. It’s now shifting to the cloud. So, how do you create products in the cloud? At the same time, e-commerce is taking off, and more and more people are buying things online instead of the traditional way. These will drive new infrastructure requirements. We need to identify those trends and help our customers build solutions to address them. Some of the bottlenecks and challenges are shifting, and being good engineers, we need to figure out how to solve those problems. This is why we’ve made investments in system analysis to develop products leveraging our computational software expertise.

SE: We’re seeing the build-out just beginning on the edge, which is where a lot of this computational software will be required. How fast do you think that is going to start ramping up?

Tan: It’s one step at a time, and we have a lot to offer in terms of computational software. At the end of the day, the customer is looking for performance, and we have shown up to a 10X performance improvement in many of our newer products. If we can provide them with a better solution to design better products, they will want us to help them. That’s a good validation for what we offer, and we’ve been seeing several repeat orders on our recently launched system analysis products. We have endless opportunity in terms of continuing to innovate and meet the customers’ requirements. When you are humble and adaptive, the customer will point you to where they want you to be, and they’re willing to pay you for the value you provide. That’s the most exciting way to launch into a new area.

SE: So what are the hot new markets for you?

Tan: I have close to 10 drivers that I’m pursuing, and some of these were accelerated by the pandemic. E-commerce is accelerating, as is cloud infrastructure, which is a sea change that some of the new semiconductors require. There’s some exciting infrastructure scale-out in storage, the network, processing the data, and then analyzing and using the data. This is a huge area of interest for all of the hyperscale companies. From our perspective, we need to understand what the market and customer need is, and then we can figure out how we can leverage computational software to provide the solutions.

SE: How about the scale-out of the infrastructure to support the electrification of cars?

Tan: Yes, you only have to look at how much data one car is generating, and we have hundreds of millions of them. I’m very bullish about the automotive industry, and you’re going to see a lot of new advances. The foundation of that is the silicon, computing, and computational software. And we will continue pushing to address the opportunities that come up along the way.