In this interview, Heidrick & Struggles’ Scott Bae speaks to Dr. Roawen Chen, the chief supply chain and operations officer of Qualcomm Technologies. Dr. Chen first defines the concept of a “fab” [fabrication facility] versus “fabless” semiconductor company, and explains how the management of the supply chain is nuanced between the two. He then shares what he calls the “semiconductor trilemma”—how leaders have to balance three competing KPIs: pursuing the most advanced technology, having the lowest cost, and having unlimited supply—and discusses the leadership capabilities he believes are critical to manage that trilemma. Throughout the conversation, he also offers his advice for leaders who are managing the current uncertain and volatile market dynamics as it pertains to the chip space, how he has seen the role of the supply chain and operations leader evolve, particularly since the Covid-19 pandemic, and how he approaches the balance between developing talent internally and recruiting talent externally, particularly when conditions are changing so rapidly.

Below is a full transcript of the episode, which has been lightly edited for clarity.

Welcome to The Heidrick & Struggles Leadership Podcast. Heidrick is the premier global provider of diversified solutions across senior-level executive search, leadership assessment and development, team and organizational effectiveness, and culture shaping. Every day, we speak with leaders around the world about how they’re meeting rising expectations and managing through volatile times, thinking about individual leaders, teams, organizations, and society. Thank you for joining the conversation.

Scott Bae: Hi, I'm Scott Bae, a principal in Heidrick & Struggles’ Industrial Goods & Technology and global semiconductor practices. Today I'm excited to be joined by Dr. Roawen Chen, the chief supply chain and operations officer of Qualcomm Technologies. Dr. Chen is responsible for Qualcomm's worldwide semiconductor operations and supply chain, as well as all manufacturing engineering functions.

Prior to Qualcomm, Dr. Chen was vice president of manufacturing operations at Marvell Semiconductor, and previously he held a variety of leadership roles, including vice president and general manager of the Mobile Communication and the Connectivity Business Unit during his tenure at Marvell and technical and business positions at Intel and TSMC. Dr. Chen, thank you so much for joining us today.

Roawen Chen: Thank you for inviting me.

Scott Bae: To kick us off, for the benefit of our listeners who may not be as familiar with this industry, can you define the concept of a “fab” [fabrication facility] versus “fabless” semiconductor company, and how the management of the supply chain is nuanced between the two?

Roawen Chen: First, let me tell you that fab is a terminology in our industry that represents a factory. So, when you talk about fabless, it means a company without the manufacturing facility. There's another term you're going to hear a lot. One is a fabless industry. Another one is called a vertical IDM industry and IDM stands for integrated device manufacturing. And the reason they are called vertical is because those are the companies that have design and they also do have manufacturing at one place.

I can tell you probably 30 years ago, 40 years ago, almost 100% of the semiconductor companies were IDM, but nowadays I would say 99%, at least from the market value point of view, are probably fabless companies. You know, if you are a smart engineer in Silicon Valley—and that really unleashes a lot of talent for the people—they don't need to spend a hundred million dollars thinking about how to build a factory. They just focus on the design [of microchips].

Many people around that time in the eighties or nineties, very few people believed in the foundry business model or believed in the fabless model. I can tell you the fundamental difference between fabless and the vertical IDM [is that in a fabless model], not only do I not need to spend a lot of money to build a factory, but I focus my entire energy trying to provide the best product. Then you just take out the product to TSMC or other foundry [for manufacturing].

So, I gave you kind of a high-level view, but there's also a very important thing, which is a foundry—the reason TSMC is very successful is because their business model is not trying to compete with their customers. They are only focused on manufacturing. They provide the best yield, the best cost structure for their customers, but their business model is not trying to compete with their customers. We call this a pure-play foundry. And as I mentioned to you previously, a foundry business is an extremely capital-intensive business. They have to spend hundreds, billions of dollars to build capacity for next one or two or three years for the entire industry, based on everybody's forecast. This is an interesting business model; I’d never see any business model like this one. But today you probably look at the tech industry, the fabless industry, including Nvidia, including Broadcom, including Marvell, including Qualcomm, even Apple, and [these companies] probably represent the most vibrant semiconductor segment.

This is the background of our industry. 

Scott Bae: Thank you for the great summary, and this is usually where I start when educating even my colleagues at Heidrick on the fundamentals of the semiconductor value chain, so this is a great way to kick off our discussion. Now when we previously spoke, you've talked about the concept of the semiconductor trilemma, which I thought was a simple but brilliant framework describing the challenge that chip leaders face.

Can you share with our listeners what that leadership trilemma is in this industry, and what do you feel are critical traits for leaders to effectively manage it? 

Roawen Chen: The semiconductor trilemma means that every semiconductor company, particularly if you are running operations and you are pursuing the three very important KPIs: one, you want to pursue the most advanced technology, [that is] number one. Number two, you want to have the lowest cost, and then number three, you want to have unlimited supply. So, the semiconductor operation trilemma means at a given time, is it possible you can you get all three? That means you have to constantly optimize all three from time to time. This is what I call the semiconductor trilemma. 

Scott Bae: Yeah, that's very insightful. I think we've seen companies in the semiconductor industry struggle who've tried to have it all as it pertains to the trilemma, rather than focusing on how to best optimize these to determine what the competitive edge of their operation will be for the situation.

The next question here: the chip industry certainly had a spotlight in different iterations throughout its existence, but perhaps none compares to the last five years, where our reliance on microelectronics was exposed during Covid, and now with the boom of generative AI and a volatile supply chain with geopolitical tensions. On that note, what learnings or advice do you have for leaders who are managing the current uncertain and volatile market dynamics as it pertains to the chip space? 

Roawen Chen: My advice is you always need to find a way to enable flexibility, including supply chain diversification. Also holding the strategic inventory, and making sure that you are also aligning your inventory strategy to multiple scenarios, including tariff scenarios like semiconductor tariffs invoiced by either China or the United States. And also you need to factor in the export country scenario. Some of the suppliers, for example, in China [or] if they are on the Entity List, you need to be running a scenario that either builds more inventory or you have to diversify in a different country. So those are the keys for a code that enables flexibility. 

But you know, somebody also would say that we can build a perfect planning system to deal with uncertainty. And my view is when we deal with the supply chain and the geopolitical uncertainty, none of the systems will be perfect. My experience has always shown that relying on the best people is far more effective than depending solely on the best system. So, being able to recruit the best and the most dynamic people is probably the best insurance you can have. 

Scott Bae: We're going to touch on the topic of talent acquisition a little bit. But first of all, taking a step back, how have you seen the role of the supply chain and operations leader evolve, particularly since the Covid-19 pandemic?

Roawen Chen: I think there's a big difference. In hindsight, we learned a lot. I think the main difference is that we started to put much more weight into the supply diversification instead of just unconditionally pursuing cost efficiency. That’s number one. And we also learned, you know, frankly not only from Covid, probably even before that, we started to understand that the customer forecast is always considered as a leading indicator.

So, there's a lot of human factors into the forecast number. For example, in the seller market or in buyer market, you are driving a different customer behavior. I think you just need to find the balance and how to handle that one. 

Scott Bae: And this is a good segue to the next question. Talent acquisition, at every level, is among the biggest challenges that keeps our semiconductor clients up at night. And it's a challenge that's somewhat unique to this industry, given how technical and niche it is with significant educational requirements in most of the functions. On that note, what attraction and development strategies have you found the most success with, and how do you approach the balance between developing talent internally and recruiting talent externally, particularly when conditions are changing so rapidly?

Roawen Chen: There's several questions you have to ask yourself. For example, What do you try to drive in your organization? Do you think you want to drive a more homogeneous organization? Or is your goal to try to drive a more heterogeneous organization? Each organization has a different pro and con, but if you believe that a heterogeneous organization is a better organization, then the ideal is you have to mix with some internal and also external talents. In my organization, I believe in a heterogeneous organization because I think it's the most powerful organization, but then the overhead is you probably will spend more time to converge the cultural elements, number one. Number two is, What do you believe? Do you want to leverage the strengths of the people you have, or do you want to spend a lot of energy trying to improve their shortcomings?

It depends on if you think you can spend time and energy and the resources to provide enough training to improve the capability of your people. Or is a quicker way to hire people and exactly leverage the strengths they have? That is also a different approach. And also, certain skillsets, critical skillsets, can be developed internally, or you have no choice but to go outside.

We do have a lot of experience before, and we found out that we look around and maybe the only way to have this unique skill is that you only can go outside. And also, a rotation program, I think, is good training. I found out that rotation is very necessary. You want to incubate the future leader. But if you don't have an internal rotation [program], what I call external rotation is equally good. I do have some examples—some of my guys, they quit a Qualcomm job and joined another company, but after a couple of years they decided to come back. And I will say, it was pretty good because they brought outside knowledge back to Qualcomm. I call that external rotation.

So that is a part of how to balance internal talents or external. I think generally speaking, finding an internal candidate is always a better approach than recruiting from outside. As a large company, it always takes a long learning curve for an outsider to fit into the larger company culture. Unless you are running out of options and then you have to go outside.

Scott Bae: Building on that, can you share with us your perspective on building a healthy organizational culture within your function and across the organization as a whole, and what does a healthy organizational culture look like at Qualcomm? 

Roawen Chen: I think Qualcomm is known for our consensus-driven culture, but there some misperceptions about consensus. At Qualcomm, we are not unconditionally pursuing consensus, because there is always a two-sided equation. Consensus-driven culture will certainly enrich the teamwork. However, without knowing the balance, the consensus building also will create a very inefficient decision process, so everything will be very slow.

I think that in the culture we try to incubate, it is tricky to find the balance. It is all about putting the right people in at the right time. So that means it’s not necessary that you need to invite everybody to be a [part of the] decision process. And also, you have to assign a different weight applied to different people, and you always need to be mindful that you invite people's opinion, but that doesn’t mean you are asking for their approval. Ultimately, you are the leader and the final decision maker, and you should not outsource the decision power.

I think that the worst nightmare I could have is something called groupthink. And I saw a lot of large companies eventually follow groupthink. I don’t know if you know this, but what I mean is basically you join the meeting, and you bring a lot of people, just to make sure everybody will cover your blind spot. But if everybody is thinking about somebody else—you probably know the issue—frankly, I have no clue what they are talking about, and then eventually you are going to make a decision based on an assumption that might not be the best outcome.

So, Qualcomm does have a consensus-driven culture, but we also constantly trying to find a balance.

Scott Bae: Excellent. And onto our final question of the podcast. The chip industry is evolving so rapidly and is affected by so many external factors like the AI boom and geopolitics as we previously discussed. What is your prediction for where the semiconductor industry will be in five years, and how will leaders need to evolve to manage through this change?

Roawen Chen: Honestly, I don't think anyone in our industry knows how to predict five years from now from previous experience. Frankly, I do not think we, in the semiconductor industry, have the capability to even predict one year from now. There's a very famous saying that the only thing we can predict is the fact that we cannot forecast.

OK. Put the joke aside. There are a few megatrends we can foresee. The first one is AI. Similar to a PC movement, internet, or mobile, I think that AI will have a profound impact on humankind, right? But the uncertainty is how the society will evolve to accept AI and find a way to coexist with AI. There is always a risk and opportunity, but you know, my personal view is AI will bring more benefit to mankind than doing harm. This is my personal view.

The other megatrend is deglobalization. I would say the golden era of globalization is forever gone. Frankly, I do not think anyone in our industry is prepared for the worst-case scenario. The worst-case scenario is the following: What if we are requested by our own government to decouple the market to each other? For example, American company, you should decouple your market from China, or Chinese company [you should] decouple the market in the United States. Or the other extreme cases is, What if China takes over Taiwan? As I mentioned to you, TSMC controls 95% of the leading node capacity.

For those kinds of scenarios, I don't think anybody can be prepared. I personally believe the most likely scenario, what we can do, is we need to find a way to decouple supply chains to a certain degree. We have to do more making China for China and then making US for the US. What that means is, if you want to serve the China market, then you should start using more China supply chain, and I think that each government probably would demand more. The same thing applies to the United States. If anybody wants to do business in the United States, there will be a push for you to use the supply chain which already [includes a] factory in the United States. The next 5, 10 years, there'll be a making trend. And I don't think any semiconductor company can afford to decouple itself from the China market, for example. It’s just unthinkable. 

Scott Bae: Well, we'll have to circle back in five years to keep score on these predictions. Dr. Chen, that wraps up our interview today. Thank you so much for your time and insightful dialogue. I very much enjoyed the discussion. 

Roawen Chen: OK, great. Thank you.

Thanks for listening to The Heidrick & Struggles Leadership Podcast. To make sure you don't miss the next conversation, please subscribe to our channel on your preferred podcast app, and if you're listening via LinkedIn or YouTube, why not share this with your connections? Until next time.

About the interviewer

Scott Bae (sbae@heidrick.com) is a principal in Heidrick & Struggles’ Calgary office and a member of the Industrial Goods & Technology Sector.