This week, I’m talking to Revathi Advaithi, CEO of Flex. Now, Flex isn’t a household name, but it is definitely in your household: it’s the third-largest electronics manufacturing company in the world, making everything from hair dryers to the Mac Pro to autonomous driving systems for electric cars.
Revathi and I also talked about the rise of automation and how to think about the future of the manufacturing workforce worldwide. There are going to be fewer people managing more robots as time goes on — and we’ll have to train people differently and think about education and labor policy differently.
In fact, pay attention to how often this conversation comes back to incredibly complex global politics. A lot of times on this show, I talk to people who make software. They write code on their computers, put it on the internet, and they’re off and running. And yes, there are often important policy issues in the mix, but they are nothing compared to the issues of running a global manufacturing company at scale around the world. There’s a reason people say hardware is hard.
This transcript has been lightly edited for clarity.
Revathi Advaithi, the CEO of Flex, welcome to Decoder.
Thanks, Nilay. Thanks for having me.
Talking to the CEO of a big global manufacturing company at this moment in world history, I think, is a great opportunity. I just want to get started at the start.
Flex is one of those companies that is kind of in all of our homes. I ran through my house today looking for things that were manufactured by Flex, and it’s a long list, but you don’t have a huge public profile. You’re not a company that’s out in front in consumers’ minds. At the very beginning, tell us what Flex does and who your customers are.
When I took this role with Flex, I remember my kids telling me, “Mom, you go to all these companies nobody’s ever heard about.” I worked at Eaton before and then Flex. They said, “Why can’t you go to Coca-Cola or Apple or somewhere where we can get all these free products?” Then I get to Flex, and they’re so excited because we make so many different things. We make everything from something as complex as a full autonomous electronics package for the next generation of autonomous cars, to hair straighteners. Think of everything in the middle. We are in the medical space, we’re in the automotive space. We do a lot of industrial work like energy meters, inverters, things like that, power products.
We make a lot for the cloud data center space; of course, consumer businesses; and used to make a lot of things like cell phones, but that part of our business, we have slowly reduced. But think about something, we probably make it, and that’s how I like to explain it to people. Everything from vacuum cleaners, hair straighteners, coffee machines, to autonomous cars, to diabetes equipment, everything in the middle.
One thing that is complicated to understand, and I think people don’t see — let’s pick a hair straightener, for example. “We’ve made a new hair straightener.” It’s got all these features and they do the marketing. I always assume they must have a bunch of engineers and designers on staff at that brand that invented a new hair straightener.
That is not always the case, right? What is the split between how much of that work happens at the hair straightener company and how much is happening at Flex, that you’re then selling and repackaging for various companies?
Each segment is somewhat different, Nilay. In the hair straightener example, usually the aesthetic parts of it, the part that is kind of touched and felt by the consumer itself is dealt by the company that is taking it to the consumer itself. The behind-the-scenes, in terms of the touchscreens, how the electronics work, Bluetooth, wireless, any of those types of packaging, design usually is done by [companies] like ours. Sometimes we purely make things, sometimes we joint-design many things with the customers, and sometimes we just say, “okay, here’s a set array of components you can put into your product and redesign it.”
A lot of companies won’t make everything. Because it doesn’t make sense. You should only make what you’re really good at and everything else, somebody else should make who’s doing it better. I’d say it just varies by segment in terms of who actually designs what part of a product.
One of the sort of the sillier jokes in tech coverage is, very recently, the instinct to just put a Bluetooth chip or a Wi-Fi chip in something and call it smart, and now you’ve got a “smart air conditioner”; we see so many companies doing this stuff. Every year I go to the Consumer Electronics Show and I see more and more things with chips in them. I wonder whether the air conditioner company hired an engineering team or did they just find a reference design and slap it on the air conditioner? Who in the end is responsible for it? Because you have to care for computers in a very real way.
Where does Flex come in there? Do you say, we can just make your air conditioner smart? Do you say we’re going to do a ground-up redesign of air conditioners that multiple companies can resell? Or is it, we’re just going to wait and see where the market takes us?
I’d say most of the time, what we try to do is have a platform design? Whether it’s an air conditioner example, or I’ll use one that we do a lot of work around, is audio, like the headphones you’re wearing on your head. We do a lot of audio design work, which is base platform. We will do the wireless Bluetooth integration, which will keep up with the next generation of those. Now there’s a lot of work that’s being done around hearing aids. Becoming more integrated with your TV, with your phone and all of that. So, we’ll do base platform designs for audio that’ll integrate all those basic things and we’ll provide that to most customers.
And we’ll keep up with all the generations of those because they move every day, and you integrate that into your set of products, or we’ll integrate that for you into your headphone design. Most companies will do that. Frankly, putting wireless or Bluetooth, integrating that into your design is really not that hard, but keeping up with the next set of chipsets, how much more accurate is that integration? How is it picking up noise cancellation? What are the different ways that you are doing sound recording? Those kinds of things become more complex. We will keep up with all generations of that, and a lot of customers will do the packaging elements of it, which they’ll bring together.
Particularly the example around slapping wireless and Bluetooth on everything is the simplest thing to do today. Whether it’s your windows or your desk, everybody’s slapping wireless and Bluetooth on it, but we’re more excited about things like who’s working on the noise cancellation aspects of it? Who’s working on the integration of everything around it that works seamlessly? We’ll design all that, put it in a platform, [and] give it to customers. We’ll do some specifications for them that’s unique to them, and then they’ll probably embrace it and put it into their design.
The reason I ask is I’m really fascinated by the relationship between manufacturing innovation and consumer innovation. I can dream up all kinds of things, but you have to be able to manufacture it at scale, and that scale of manufacturing is what changes the world. I could have dreamed up the smartphone 20 years ago, and I probably did draw it in a notebook when I was 15 or something, but you had to figure out how to make it at scale and make it affordable. Where do you see the relationship between your investment into manufacturing innovation and how it enables the products that we see?
Do you spend more time thinking about whether to innovate on the core capabilities of the device? Or do you spend more time saying, we have to innovate in our manufacturing capability so that we can actually build all of the things that we’re inventing?
We have to do both. Let me give you an example that’ll bring it to life. One thing that’s kind of near and dear to our hearts is these drug delivery things that you put on your skin these days that are either sensing something in your body — or eventually, maybe it’s even transferring some product into your body — used a lot for things like diabetes. In an example like that, what we would say is, do we have to work on that sensor capability on that particular product that goes on your skin?
Do we have to work on the electronics or even, for example, the communication capability on that? Because it usually talks to your phone and says, “Hey, are you low on insulin? Should you be giving yourself a shot?” We will be working on all those aspects of it with the customer to design that particular product. That’s one aspect of it, but these days, it’s so complicated in terms of the level of automation, the requirements around even something like sterilization of some of the sensors that go into your body.
Designing the manufacturing is as important as designing the product itself
The sterilization process itself is so complicated — automated, but complicated — that designing that manufacturing process is even more complex. Someday, if you have time, you should go to one of our factories in Chicago, which is one of the most complicated drug delivery manufacturing setups that we have. Designing that manufacturing is as important as designing the product itself, because it’s quite critical to make the product with the least quality issues, keep redesigning it so you’re lowering costs, or [make] new design changes that are coming in, [that] changes the manufacturing itself.
We have to be able to do both, and they’re both quite different capabilities. We have different teams that work on manufacturing processes itself and what we have to do to make that adaptable to the product, and very specific product experts who are working on it with customers.
We just came out of the Trump administration, there was a huge push to bring manufacturing back into the United States. Even outside of the Trump context, manufacturing is permanently tied up in politics and policy around the world. There are trade issues, there are employment issues, or national security issues in some cases.
Where does Flex fit into that conversation? Because the focus is usually on big consumer companies like GM or Microsoft. Flex obviously plays a huge role in that. Where do you sit at the policy table?
We’ve been talking a lot, particularly to the new administration, and as the incoming new Commerce secretary came in, we spent a couple hours with her, in terms of really trying to take the noise out of manufacturing in the country. It seems to be a hot topic, but it’s also a very complicated topic.
The reason that manufacturing left the United States is for the simple reason that labor arbitrage was better in some other parts of the world, but those things have changed pretty significantly because labor costs have gone up in other countries. Equipment automation has played a significant role, so that kind of cost arbitrage is coming closer and closer together.
When you say cost arbitrage, explain it for us. What specifically do you mean?
If you think about making a product, there is the material costs; where do you buy the electronics, where do you buy the plastics that go around it, all the things that you buy to put a product together. Then you have to take it to some factory around the world, and it costs labor to put all that product together. You buy the material, [then] there is labor costs, and then of course testing and all that stuff, and eventually it becomes a product.
One of the biggest components of costs in many products, particularly consumer products, was the labor side of it. That’s why jobs moved to countries where there was cheaper labor.
The labor was cheaper in China, for example.
China, or Southeast Asia, or wherever. Those labor costs have gone up as those economies have developed. [As] those labor costs have gone up, the material costs have come down. Electronics costs have come down significantly. The cost of putting these two together has come down because we have new equipment, a lot of automation, [and] a lot of robots. Productivity is amazing. That has also brought the total cost down. You don’t need 100 people to make something anymore. You can do it with five, because we may have a fully automated machine, you may have a robot helping you, but you need five highly qualified people to do it.
That starts to bring the conversation back to countries like the US or Western Europe that have lost a lot of manufacturing jobs. [They] can play a role, but I think what we have to be realistic about is what type of manufacturing really fits with the United States. So, semiconductors, which is in the news a lot today, [and] making chips, which are complex, those definitely are a great fit; large things that you can’t transport around the world, like making cars, those are a good fit.
Complicated, highly automated manufacturing is a great fit for the US
Anything like medical, where it’s complicated, where it has high automation, where you need very complicated processes and a high-skilled workforce, those are a great fit. I think those are the things you have to focus on in terms of bringing manufacturing jobs to the US. Sometimes people focus a lot on [building] cellphones here, or laptops. You have to be realistic about [if] those things are really going to happen or not. Because those are still labor-intensive.
Anything that is significantly labor-intensive, I think, and without a high level of automation, will still struggle, but the beauty of what’s happening in the United States today is skilled labor is getting better and they’re getting more scale, like we can see in our factories in Chicago. Automation is unbelievable. You put those two together and we’re seeing some very complex manufacturing processes return back to the US.
Let me be painfully reductive and try to understand what you’re saying. You’re going to disagree with me, but I’m going to try. You’re saying, we’ve automated so much that we don’t need 50 low-skilled employees in China. We can get by with five highly paid managers of robots in the United States. Is that the right policy outcome? Sure, that’s what’s happening, but saying, “Oh, we solved our labor problem with automation” seems like a very different class of problem.
It is not as simple as that, because it’s not as though you’re just solving that with bringing 100 jobs from somewhere in Asia to bringing five jobs to the US. There is a significant skills gap and education gap to also make that available. The policy problem is skilling workforce, which is a huge part of the issue; are they getting trained? Coming out of high school, apprentice programs are nonexistent, how do you get people to train in technical education? Reskilling people for that is significant. If you think about the jobs I worked in as a factory supervisor in Shawnee, Oklahoma, or Hutchinson, Kansas, those were heavy machining, heavy manufacturing jobs, very manual. [They] required a whole different set of skills in programming your machine and being able to troubleshoot it, and then make sure the robot [is] working or make sure the program is working.
I think reskilling is a pretty significant issue. That’s one big policy issue that has to be top of mind. But the second one, obviously, is providing lower cost of capital. We can’t forget that. I know that’s a complicated issue, but one of the big reasons why many Asian countries do well is that the cost of capital is much lower. Providing small businesses, [and] medium businesses, cheaper cost of capital so they can set up these businesses, put the automation in, skill their labor: all that is important. It’s not as simple as just saying, “I can hire five people with the skills and then it’s all going to be great.”
You mention the cost of capital. What you’re talking about is the rates at which companies can get money to build things. That is also government policy. It seems like we started out talking about manufacturing and we are running into the policies of various governments that have a direct impact on manufacturing. How much time do you spend on, let’s call it, industrial policy, versus operating the actual manufacturing company that you run?
My main job is to run our company and keep our customers happy. But recently, with administration changes and with the huge conversation in most countries about, how we bring supply chains back to our country? What do we need to really do to train our workforce? What help do you need from capital’s standpoint? Recently, it’s been a lot of conversations with various governments, in terms of what does it take. I’d say it comes and goes, right? If it’s an election cycle, a little bit more. If it’s not, a little bit less.
But being one of the world’s largest manufacturing companies, it’s amazing, manufacturing is always a hot button when it comes to jobs and the election cycle. It always becomes a big topic about what you can really do to create manufacturing jobs. Then this whole thing around costs of shipping is so high, trade issues, now pandemic issues, you can’t move product from A to B. Everybody wants to make stuff close to where they are. That’s driving a lot more conversation around what policy we have to have to make things closer to where we are. I’d say, in the last kind of a year and a half, [it is] a lot more than what we should typically do.
Obviously COVID is still ongoing, I think it is still the biggest story in the world. You have something like 50,000 employees in China. The Chinese government was very aggressive in responding to COVID. I think that workforce and that supply chain came back a lot faster than anyone anticipated. What, from your standpoint, made that possible?
If the question is on China’s specifics, the Chinese government is very strong in being able to implement policies and make them stick. For example, this is always hard to envision, but [COVID] happened at the peak of Chinese New Year. We suddenly thought, “Oh, we have 50,000 people coming back to work, what are we going to do with this?” The Chinese government had policies in place about what was getting downloaded on people’s phones. We were able to track every one of those 50,000 people and tell exactly where they were, when they were going to come back to work, if they should come back to work based on whether they were in a hot spot.
The control mechanisms were so significant. You can’t do that in every country. Right? Think about doing that in India or in the US, nobody’s going to allow you to track anything. China is just good at managing large scale, any kind of tracking management, that’s a whole different thing than most countries can do. I think that helped China a lot, it continues to help China a lot. They can implement lockdowns and shutdowns and enforce it like no other country.
That edges into some dicey territory, right? China is able to enforce lockdowns, they’re able to enforce and, in some ways, monitor their citizens in ways that the United States would certainly theoretically not allow, and then practically, we have seen absolutely not even be capable of even suggesting. As you work with the Chinese government, is that something that you think about, that they veer towards autocracy more than democracy?
Over the years, having lived and worked in India, in China, in Europe, in many states in the US, I would say that we have to allow local governments to function the way they feel [is] the best fit for their needs. Do I have a point of view on democracy? Do I have a point of view on the parliamentary system or how elections are run? Absolutely, in almost every country I’ve lived in. But in terms of what works for that particular country, you have to allow the local processes to work. I’d say Chinese citizens are very proud of how well their country has done over the decades. I have many friends and colleagues there.
I’m very supportive of how policies are enforced in the country. It works for people who are living there and who are seeing the benefits of that over the last two or three decades. Do my personal opinions match with all of them? I would say, maybe not, but that applies to almost every country I’ve lived in. I’d say that we try to keep that separate from, are we creating jobs in those countries that we’re in? Do we do that within the culture and values that fits me and fits my company? Yeah, absolutely.
I would characterize that answer as what I heard from leaders of large global companies for a long time. “We’re going to do business around the world and we’re going to believe in the local governments. Maybe if we’re really good, we’re going to export American values by showing up and being good stewards. We’re going to give everybody rock-and roll records.” That was the answer until about five years ago, and I think the Trump administration kind of blew up that answer.
His approach to China was very aggressive. That has led to a broader rethink of globalization across the world, particularly for leaders that do business in China. Do you think that that answer holds? As the trade war has heated up, the cost of moving goods rising, as tariffs have changed, as Brexit has happened, the stock globalization answer has gotten a little more tenuous, I think for leaders of large global corporations. How has your view changed there?
I would say — and it’s not the stock globalization answer — I think you have to step back and say, what started the whole trade issue? Many administrations before the Trump administration [tried] to work on the question of, is our IP safe? Are there the right measures in China for IP security, and what can we do to make sure that is well-managed? That’s where the whole crux of the issue started, which all of us are highly supportive of. I have lived there, I worked there, I would say yes [it is a] very important topic.
Let’s make sure that that happens well. Do I believe in a free market economy, do I believe that the world is served best by making things where you can find the most productive, most efficient system, and that way everybody benefits? I would say, yeah, absolutely right. Let’s think about the semiconductor crisis for now. The easy answer would be why don’t you put every fab job in the US, if you want to make that work, and move it out from wherever it is in Southeast Asia today, and try to make all the fabrication for chips in the US.
The cost of that would be so prohibitive and would make the industry so uncompetitive. That doesn’t really make sense. You wouldn’t even be able to do it. But on the other hand, do you want to invest in certain parts of it, like more assembly and test capacity, or more memory making, which now doesn’t sit as much in the US, and then ramp up costs where it makes the most sense? Then maybe you spend $50 billion instead of $400 billion, for building an end-to-end semiconductor industry.
The traditional thesis of globalization, [from] when The World is Flat [by Thomas Friedman] was written, may not be exactly the same, but the ideas and the philosophy of, there is a free market economy, we should be using the best aspects of that free market economy, because that provides the most efficient system for the world, I think that piece is still is true.
You have a lot of decisions to make across a huge variety of disciplines. What is your decision-making process and how has it evolved?
It’s been two years since I’ve been in this role, but I’ve been in the C-suite for a while now. My process has always started with, what is the best for my colleagues? That’s where my process always starts. Sometimes what’s best for my colleagues may not be the best for my customers, may not be the best for other stakeholders in the mix, but that’s where it starts. I’ll give you an example of this, when Malaysia was going through a shutdown during the pandemic. We have a pretty large presence in Malaysia.
We decided, well before the government decided, that we were going to shut down our Malaysian factories, because we saw the pandemic actually rising pretty quickly. We decided it was the best for our employees. We couldn’t control it. We said, “Let’s just shut down. Everybody go home, we’ll take a pause.” And then the government announced a shut down right after that, maybe a week or so after that. My customers were calling me left, right, and center. They were [saying], “Hey, your competitors haven’t shut down. Other people haven’t shut down. Why are you doing this?”
We said, “Well, it’s the best thing. We can see [the] numbers. We can see what’s happening in our communities. It doesn’t make any sense.” That’s part of our culture. We took tremendous heat for doing that from our customers. But after that, three or four weeks [later] after we were able to bring back our employees safely, I got a call from one of the CEOs, and he said, “You know what? I know I called you and said it was a wrong decision, but now I’m calling you and telling you that it was the best decision you made.” Tied to that, what always comes first is, what’s the right thing to do for our employees and our colleagues?
When you have 165,000 employees, you have to think on a case-by-case basis, every country is different. Then of course, our customers and our shareholders. My decision-making process has never changed from that. Whether I was a shop floor supervisor in Shawnee, Oklahoma, or I was in Hammond, England, there were always colleagues around me. Because I come to work because I like the people I work with, that’s the most important thing, and we want to do what’s right for our colleagues around us.
Then, I’d say there’s been some level of thinking around, which types of customers do you want to support? And we have backed out of certain customers that don’t meet our purpose. We have done that. I think our decision-making has changed around that.
Give me an example of that. Usually, when I talk to platform CEOs, we end up talking about content moderation, but backing out of manufacturing because you don’t want to support a customer, I was not anticipating that to be a thread here. What kinds of things have you backed out of making?
It’s hard to give you examples because I can’t name any of my customers, it’s really hard, but I’d give you a simple one, a very specific defense customer, where we were making very specific products that we felt like we shouldn’t be involved in. We exited that and we stopped quoting those types of products in the defense space in our industrial vertical. That’s an example of something that just didn’t fit our purpose and what we thought we should have been involved in that we decided to back out of.
Abstractly, take me through that decision. Did you get an email from someone saying, “Hey, we make this thing and we shouldn’t,” Was there a meeting? How does that work?
It works very, very elegantly. It works like this. One of my leaders who’s working for me says, “Hey, listen, we’ve recast our vision, our mission. We’re thinking about the kinds of products we should be in. We’re quoting this particular product in this country. It’s a customer we already make product for. What do you think we should do?” I’d say, “Hey, listen, does it fit with our new purpose? Where do we think we are, as we have defined our new purpose?” If the answer is no, I’d say, “Well, okay, so I guess that makes your decision for you.”
Then we have to talk about, how do you elegantly manage that exit? How do you talk to the customer and do all of that behind the scenes? But what’s helpful is people can just read what we’re saying about our culture and what we are trying to do and reach their own conclusions. It doesn’t take a lot of convincing to make that happen, but it takes some amount of work to manage the exit.
Let’s talk about this chip shortage. It is having huge effects across every industry. Ford just announced it’s going to cut production by 50 percent this quarter. Apple, which has been so far seemingly unaffected, said the chip shortage will affect them in the future. How is it affecting Flex, and where do you think it’s coming from?
It’s definitely affecting us, like everybody else, because everything we make has some form of a chip in it. Almost every business, automotive, medical to some extent, industrial, consumer — think of a business, it affects all of them quite significantly. We’re in the middle of this with customers, with suppliers, with foundries. I think it all begins for the simple reason that most companies are not great at giving forward-looking forecasts. The idea that everything has more electronics has been happening for a while.
The idea that chip content is going up is understood, but nobody’s good at forecasting these things. Automotive customers, if you take them as an example, cut down production last year at the peak of the pandemic, somewhere between 50 percent to 80 percent. That happened like last year in Q1, [and] moreso in Q2. Then you think, okay, fine, business as usual should be coming back after that. Well, no, not only does it come back, but it comes back to a very significant ramp of plus 80 percent. That didn’t happen overnight, the fact that electric vehicles are more present and have more electronic content.
The demand for chips has only gone up, day by day
The fact that you’re just going to see automakers need more chip content than they had before, autonomous vehicles the same way, should have been seen and it should have been well predicted. But forecasting is a highly inaccurate model and nobody gets it right. Now, you combine that with the fact that consumer products were going through the roof in the pandemic. Whether it’s phones, whether it’s iPads, whether it is everything you use in your home, they all have some kind of chip incorporated in it.
[We had] consumer products going through the roof, China getting concerned that they’re not going to be able to buy enough chips, and so they start to stock up as the whole issue is going on with Chinese customers. So, the combination of all of that meant, as the return from [the] pandemic and demand came up, it just wasn’t enough in built foundry capacity to manage it. There’s the foundries, then there is the people who assemble, test, package, and put it together. Everyone was running out of capacity. Any little hiccup on top of that; power goes out, there’s a fire or something like that, makes it worse, and that’s what has added up to this.
Again, to be reductive, it sounds like what you’re saying is demand fell off a cliff for a minute, and no one forecasted how much it would come back. At the same time, foundry capacity supply did not increase. There’s not a shortage of raw materials or anything. It is literally that the manufacturing capacity does not meet demand right now?
The railroads and all that goes into chip production, there’s still enough [capacity] available. It mainly comes down to foundry capacity and then packaging capacity not being enough.
Do you think that demand is going to stay high? Do you think it’s going to come down? How do you think that plays out?
I think our view is that demand continues to stay strong, just because everything is electronics and everything’s going to need chips in it. That is just going up significantly day by day. The requirement for chips is just going to go up, so that demand is going to stay strong, but it takes a while for foundries to ramp up capacity, so that’s going to take well into next year, or maybe even further, to get that type of capacity ramped up. I think we’re going to have to wait and see how that plays out.
One thing that has struck me is the conversation about the chip shortage. We keep talking about chips like they’re a commodity, that you can just get some computer chips and you’re off and running. That is not true, right? There are many kinds of chips that do many kinds of things. They operate at many different process nodes and sizes. Where is the most acute shortage?
Today, the shortage is mainly in what’s used in the automotive industry, because the automotive industry makes up like 8 to 10 percent of the chip usage in the world. It has not kept up the way consumer products have, which have had many iterations and many changes. The automotive semiconductors have been pretty outdated.
What a surprise.
Yeah, the equipment is outdated, the designs themselves are outdated. That’s where I would say the biggest bottleneck is today. For that, not only do automakers have to redesign the things that they are doing and have to bring it more to current standard, but then foundries have to invest in equipment that is not transferable from what is being used for consumer products. That means you have to go buy equipment, which is billions of dollars. [Then] you have to put it in place and ramp up production, so it’s very, very specific to the automakers’ needs. I would say that’s probably the biggest gap today. Foundries have to decide whether they want to invest in that, whether that is where their strategy fits, which is also part of the equation, and we’re talking about three or four foundries in the world, of which two are the most critical.
You mentioned you’re investing in the systems that run electric cars. Is that just an opportunity to do a paradigm shift in the technology that runs cars?
Companies like Ford or GM, need older technology that people aren’t investing in to make pickup trucks, but you can also just design an electric car from scratch from new technology, design it to be iterated faster than any carmaker has to date, save Tesla, and basically create a new market and a new approach to cars. Is that how you see the opportunity?
That’s how we see the opportunity, but every carmaker is doing it differently. There are some who are just building off of existing platforms and trying to make a variation of it. At least using the base chassis and that being similar across their platform. Then there are others who are kind of saying, “Hey, listen, I’m going to build it from scratch,” like, Porsche is a good example there. Each carmaker is doing it differently, but I’d say, in some way, you have to play into that opportunity, whether you’re going to do it from scratch or not, or whether you’re going to take a current design and redesign it, in some way, you have to do it.
Because it’s just a basic expectation that we’re going to use less of traditional forms of fuel, and there’s no reason for us, as cost points are coming down for electric vehicles, that you don’t incorporate it in some way, but each one is doing it their own way.
This brings us back to what we were talking about at the very beginning, which is you’re investing in the platforms, the core technologies of cars. You’re not making a car.
You’re investing in something that’s almost a car. You’re going to design a bunch of stuff that could almost be a car and then a car company can come to you and say, we want this platform to actually design and then build a consumer product on. Where do you see the stopping point?
I never see myself putting together the body of the car and the seat of a car, and I’ve spent too much time on these industries to know that, that’s not a place that we’re good at, or that we want to be in, but where you do see it is, can you tie more electronic products together? If you’re making inverters, can you tie in chargers? Can you tie in that with your communication platform, for example? Today, if you think about automotive communication; wireless, Bluetooth, everything that sits in an automaker’s central console, it’s still pretty outdated in terms of how it’s manufactured. So can we design all that? Can we put all the electronic platforms together, which is an inverter, converter, charging unit, a battery management system, can we put the whole platform together?
Yes, absolutely. Will we do something in terms of the overall chassis design and all that? Probably not. You’ll stop at a point where you think you have the best opportunity, but someday, can you see doing something with battery design? Because if you’re doing the battery management system, do you want to do something in battery design? Maybe. I think that’s kind of how we see it evolve. But I don’t see us doing many of the mechanical things. I think those we’ll leave to the traditional people who are good at it, and we definitely don’t see ourselves putting together the full car.
The Biden administration has made chip production in the United States a huge part of its infrastructure package. There’s a similar plan in the EU. You mentioned there’s only three companies in the world that really do chips, two of which are the most important. The third is Intel, which is in the United States. Do you think that all of this investment will work on a timeline that actually affects the current shortage, or is that a decade away?
I would say what helps the current shortage is immediate foundry capability. You’ve heard those announcements come out from TSMC, from GlobalFoundries, from Intel, and those have to get ramped up over the next 12 to 24 months to help the current chip shortage. That’s where I think we would see some immediate effect. What the Biden administration is talking about is something more than that. They’re talking about longer term, how do you bring a larger percent of the foundry capacity back into the United States? Also, how do you maybe bring back more memory production into the United States? Those kinds of things, I think, will take a longer amount of time because a lot of that capability sits in Southeast Asia.
I’d say the infrastructure package the Biden administration is talking about will help some in the short term, but it mainly helps in the longer term because bringing back memory production to the US is going to take a lot of time and effort, and building that capacity is not going to be easy. Assembly and packaging almost doesn’t happen here at all. It all happens outside of the US. Bringing that back is going to take a significant amount of work, training labor, which we don’t have today for that kind of work.
I would say there’ll be some short-term help as Intel puts in additional capacity, but most of it, I see it through the next five to 10 years.
I keep asking about chips because you’re a manufacturing company and everything has a chip in it now. Is that just the rate limiter for you right now? Do you need a supply of chips to grow and build more products?
I’d say the other way to think about it is, yeah, if we didn’t have enough chips, we wouldn’t be able to grow fast enough. Right now, are we not able to keep up with demand because of chip shortages? Absolutely. The answer is yes. I think that applies for us and that applies for almost everybody else, so absolutely we see it that way.
There are some other limiters in the world; labor is one, climate is another, there was just a big power shutdown in Austin, Texas. Samsung told us it lost a month of production [in] a chip factory there. Flex has facilities in Texas. Were you affected by the grid failure there?
We were, but we were affected for a lot shorter time because our manufacturing doesn’t work like a foundry manufacturer. In a foundry, when you have an electrical shutdown of some kind and you don’t have enough backup power to run critical parts of your foundry, bringing those back up takes a significant amount of time, and that’s not how our production works. When we lose two or three days, we can start up again and almost pick up where we left off.
Because we don’t have those types of critical foundry manufacturing set-up that Samsung has or other companies have. That’s why you heard most of the packaging people, NXPs, TIs, Samsungs of the world say they lost a lot of time with the weather issue, because when they shut down, it’s not as easy as starting back up. They have to go through a process one at a time, it takes almost a month to start your ramp back up, and [it’s] very different for us. We lost time, but we were able to make up most of it. But we obviously were impacted by the shortages.
As we talk about where factories are located, how much do you think about things like grid resiliency and climate preparedness? Is that up there with skilled labor trade policy?
I would say we definitely think about it, particularly when we’re thinking about building a new factory somewhere. We’re thinking a lot about availability of power; do we have to have our own power setup? Do we have to have solar installed or do we have to have some other alternate power installed? What does the government support in terms of energy policy? Those all go into our consideration, particularly when we are doing a greenfield or a new site. And they’re very important considerations. Because we have large factories, they use a lot of energy, and if you have inconsistent and unreliable power, it’s very, very hard to run factories like ours.
We don’t want to be putting that much battery backup power, and that’s the reason it plays a huge role for us in terms of decision-making and how we think about where our future factories will go.
You have a lot of sensitive customers. We just saw a ransomware attack hit Quanta, which is one of your competitors and also [in] manufacturing. For those who don’t know: hackers attacked the system, they leaked a bunch of Apple’s schematics, and they said they would leak more unless there was a ransom paid.
That seems like an entire other surface for you to worry about. Is that kind of attack increasing? Is it stable? Is it just that this one happened to be very famous? Is that something that you have to spend a lot of time worrying about?
We do worry about that a lot. We just did a long deep-dive with our board on this topic, not just in terms of it being able to enter our factories or affect our customers’ products, but the implications of that, right? For example, if it happened in a medical product, that would be pretty significant for us. Of course, if it impacts large customers and it’s in the news, then it’s a PR issue. But in general, just the issue that it could affect our customers, it could affect consumers who are using our products, is of significant concern to us.
We have an incredible IT team, right? We have such a complex system and it’s everywhere in the world, but they are really good. Even when all the large attacks happened last year, we really had managed to minimize the effect of it. I’m spending time on it. My teams are spending time on it. Our board is spending time on it. We definitely worry about this a lot, and I would say that we have really increased our focus and our investments in this space to make sure we’re not affected, or our customers are not affected.
A couple of days ago, you tweeted, “We need to act quickly to implement solutions that bring every aspect of the global economy to net-zero carbon.” You are a huge consumer of power. Are you thinking that Flex will be net-zero carbon or are you just trying to buy power that’s net-zero carbon?
I think we want to, ourselves, be net-zero carbon. That’s what our aspiration is. We’ve already committed to pretty significant reductions through the next decades coming on. We’re installing a lot of our own power sources where we need to install as it becomes more reliable, and we don’t want to just be buying power that’s net-zero. If we can do it ourselves, I think that would be our ideal place to be. I think most companies should think about it that way, but I would say, it doesn’t take away that if you can buy net-zero, that’s fine too. But if you can do it yourself, even better.
The real thing is, can you commit to getting it done? Even for our teams, when tech companies, I don’t want to take away from large companies, software companies, that are making these commitments of being net-zero, but like, you’re not making anything at the end of the day. That’s a whole different ball game that when I commit to cutting emissions and being net-zero, I’d really have to do it because I have 130-plus facilities that are using a lot of energy, and I have to do it with my supply base, which is even more significant, so it’s a lot of work for us to make a commitment like that.
Do you have a timeline?
Yeah, we’ve given a 50 percent reduction for 2030, and we have actually worked through our ability to get that done, and that’s our first major hurdle that we’re going to overcome.
That’s Microsoft’s goal, for example, to be 100 percent net zero. You’re saying, well, they make bits and we make atoms, and that’s just a different world.
You mentioned that you’ve seen our coverage of Foxconn. We’ve spent a lot of time covering whatever it is they’re doing in Wisconsin. Now they’re intimating they might build cars there, so there’s a lot of threads to pull on.
One of the stories there is the huge tax incentive package from the state. As you think about locating factories, the political climate that you’re in, do you think those incentive packages work?
I’ll give you my personal view, Nilay, because I’ve built enough factories in enough countries to know that they work to some extent, if you need something as a gap filler, but will you make a decision to put a factory in a place just for the tax incentive when you don’t have a good supply base surrounding it and you don’t have the right labor or the skills available? The answer would be no. Does the tax incentive give some additional impetus to choose a place, all things being equal? Yeah, maybe yes, but there’s no way that you can make a decision based on a tax incentive if you don’t have the right supply base surrounding, that pays for the parts you need; you don’t have the right skill sets, then it doesn’t work.
For me, the tax incentive always is the last point of the decision, all things being equal, and it’s a very small portion of the decision. It really isn’t the major reason why one would make a decision to locate their manufacturing factory anywhere in the world.
One of the other stories with that facility was that it got very political. President Trump was there with a golden shovel. Apple reopened its Mac Pro facility in Austin. That’s a facility run by Flex. Trump was there. Everyone just sort of let him have his moment. What was your reaction to that sort of optical political moment?
I’d say my reaction to almost any of these situations is that in our factories, Nilay, we make what our customers are wanting us to make and what we have committed to make for them. Our factory is their factory. It’s an extension of their manufacturing capacity. We’re providing expertise, people, all of that. If it benefits one of our customers, and that’s what they want to do, then we’re here to support them, and that’s how we see it. It’s not our role to say no to a customer, unless it doesn’t fit our values, unless it doesn’t fit our philosophy. Typically, we don’t do that.
We host a lot of government visitors in many of our factories across the world, either by ourselves or with our customers, and we feel like if it benefits them and it serves a purpose, we’re happy to support it as long as it doesn’t go against our purpose and our philosophy.
That relationship to your customers is really interesting there in the context of factories being the ultimate political photo op, right? Your customers have some political viewpoints, you have some political viewpoints, then every politician wants to wear a hardhat and ring a bell.
It strikes me, when you say you make things, that it is a difficult job with lots of people involved. It’s energy-intensive. That is a big swirl of complication. I think you’ve done a good job saying, “We want to stay out of politics,” but here, at the end of this conversation, I’m wondering, how much can you really stay out of politics?
I’d say the way I see we stay out of politics is, my personal politics has nothing to do with what I do with my customers, what I do with my employees, and how we run our organization. We understand that we have a role to play in politics, in terms of taking a stand where we see that it’s right for our company, for our employees, do we take a stand there? Absolutely. But are there places where we don’t have a stand to take? Yeah, we don’t. On trade, did we have a stand to take? We did. Did we take a stand on giving our employees a day off for voting? It didn’t meet certain political parties’ criteria, but it didn’t matter to us. We felt it was the right thing to do. We said, “Okay, you get time off to go vote.”
I think we take a stand based on what we think is right for our organization and our customers. Does it play into the politics of individual governments? I think that’s not important to us at all. Whether that plays out today in India or in China or in Malaysia or in the US, I think we have to keep personal politics aside. We have to make sure we have great government relations, and then we mix that up with saying, “What’s the right thing to do?” Voting time off is important. It’s a right that we think personally is good for every employee to have. [The] US doesn’t give it, we say, we’re going to support that, and we make that happen. Doesn’t matter if any political party doesn’t like that agenda of ours.
But that one was relatively safe. Trade is a little dicier, right? How do you split your personal politics from the trade policies that would benefit your colleagues and your shareholders and your company?
This is not something anybody has made up. It started years and years and decades ago [with] trade policy, competing with free markets, competing with tariffs. We’ve all seen that in the last 30 years of our career, right? They come and go many, many, many times. The pandemic has changed it. Intellectual property security has changed it. Costs of logistics have changed it. If it is done for the primary purpose of politics, then I feel differently about it. If done very clearly because there is an agenda that helps a certain country, then I personally feel different about that.
But at the end of the day, my big support is, less trade and tariffs, and open free market policy is good for the world. The less I see of it, the better off it is, is my personal stand on this. Does it happen everywhere in the world? Absolutely.
I want to end with an expansive question on timelines. I often talk to software company CEOs. Their timelines from decision to execution can be as short as minutes. Maybe it’s more like a year sometimes, but it can be as short as minutes.
The button was blue. Now it’s going to be red. They push the change, it’s red, and off we go. Your timelines, with atoms, seem much longer. How should people think about the decision-making timeline for you, a manufacturing CEO?
Not as long as starting to make a chip in a foundry — which is years — but it is long. So, if we are starting to ramp up a program, let’s say in Austin, or in Columbia, or Memphis, by the time you buy the equipment, it’ll take, let’s say, 30, 40 weeks, the equipment comes, gets landed, gets qualified, you ramp up production. I’d say the minimum, best-case scenario if the stars lined up, would be 12 months. In most cases, it would be somewhere in the 12 to 24 months. When we make a decision, you have to first understand that demand is still going to be there 24 months down the line.
Then second is that, in that 24 months, nothing changes, and all the scenarios that you planned, whether it is government, or tariffs, or competency of the workforce, or any of that. We definitely have to make it in that timeframe. So, yeah, very different from, let me change the line in the software code and here I go.
What do you think people should be looking for next from Flex?
I’d say the most important thing, well, what we’ve talked a lot about is, we’re very excited about investing in businesses like health and automotive. Particularly because there’s so much happening in that space, from a technology change, that we think we can bring a unique perspective to, so we are most excited about that, and we want to participate in that. We want to be part of the change that’s going on in EV. We want to be part of the change that’s going on in autonomous. We love what’s happening in the healthcare space, in the amount of work that’s going on with our teams to rethink the health care space and help our customers do drug delivery better, or things like that better.
We are very excited about that. That kind of technology, that is changing the world in some way, either it’s helping people or helping with climate change because you’re using EV or autonomous, those are the things we’re the most excited about. What you want to see from Flex is more and more and more investment and more and more conversations around that. It’s just a personal passion, but it’s also what I think we can be good at.
Decoder with Nilay Patel /
A new podcast from The Verge about big ideas and other problems.