SMTC's CEO Ed Smith Discusses Obsolescence & End-of-Life Challenges

Podcast Transcript

Bill: Hello everybody. I'm your host, Bill Bradford. Thank you for joining this episode of Flipping the Channel. I'm very excited to have as our first guest on this podcast Ed Smith. Ed is the President, CEO, and board member, of SMTC. Ed also spent a significant portion of his career at Avnet Electronics in various executive roles. And before that, he was the CEO of Semtech. So Ed, thanks for being here on the show.

Ed: Thanks for having me, Bill. I appreciate it.

Bill: Let's start to maybe fill in the gaps a little bit about your background and how you got into this space. What's the short story of Ed Smith?

Ed: Sure. I started in the distribution business very early in my career. After about ten years, I decided to go into the EMS space because I said, "Hey, if I haven't found something hard enough, let me find something even harder." I went from Avnet to Semtech. After about three years, we sold that business to CTS, who ultimately sold Benchmark's business. So today, it's part of the Benchmark family. I then went back to Avnet. It was nice to be invited back to my old company. I then got into the electronics distribution field in a big way. I spent the next 12 years first as the Senior Vice President of Sales and Operations, then ultimately, for my last eight years, as President of Avnet. Then I left Avnet and wrote a book. As I was deciding what to do with that book, SMTC called me, and I started working there. The perspective of being both on the customer side and the distribution side, I think, really helps me understand what's going on in the marketplace.

Bill: So what do you see going on in the marketplace now that you brought it up?

Ed: You know, it's changing as quickly as I've ever seen in my career right now. Things are getting tighter with wafers and the capacity of parts. Technology is changing, getting smaller and smaller, down to eight nanometers, now three nanometers, and continuing. As the technology changes, guess what happens? They "end of life" parts. Older technologies go away. They need that space and capacity for the newer technologies. So you have the constraint of lead times jumping out. You have technology changing, and you have some end-of-life issues going on in the marketplace. This is probably the fastest I've ever seen a change in our marketplace. We had COVID, and when that started, factories closed down, and things got tight with lots of constraints. Then it opened up, and things became very normalized. And then, all of a sudden, demand just went off the charts. It was mostly driven by electric vehicles and the new electronics in cars. And clearly, the defense-aero being strong in some of the older technologies has created quite the constraint and lead time problem.

Bill: So specifically then around some of the obsolescence and end-of-life issues, what has this market kind of done to that space?

Ed: It's exacerbated an issue that already happens, right? Normally companies say, "I'm going to have an end-of-life issue." They have a year concerning timing. You try to figure out how many they need. Then you try to guess the market. Sometimes that works well. Sometimes it doesn't work well. But when the market moves as quickly as it's moving now, sometimes they run out of wafers before you even get the opportunity to buy for the end of life. Or they want to stop building it, and they have extra wafers, and they can use different companies like Flip to sell you the wafer, and you can then package it and build it later. But most companies don't want to tie their cash up, particularly when the COVID-19 issues are going on. Nobody wants to tie cash up into end-of-life products and obsolescence. They'd rather somebody else do that and then use the cash at a later date, even if there's a little bit of a higher cost.

Bill: So, how have you been dealing with the challenges of the obsolescence parts? What are some of the things you've done at SMTC to mitigate the risks and issues there?

Ed: So we'll go into the distribution market like Flip, and we'll say, "Hey, here's the guess we have for the next year or two." Then we try to work out a deal where they hold them for our customers for a period of time. Sometimes we'll say, "Here's what we think we need," and get them to hold the wafer. So it's less costly for them. And then they can realize a finished product later on. The last, but the most costly, is we'll try to ask the customer to redesign their product. I've had customers for a $20 part spend a million dollars redesigning a product just because they can't buy $20 parts. You never think of that end of the life cycle until you get there. The reality is they've made a commitment to build something, but they can't get there without redesigning and being able to get the parts. That's the worst outcome that possibly can happen. Holding the wafers is probably the best outcome. And then trying to guess the quantity is probably the middle outcome in tying up working capital.

Bill: We've also seen challenges in this constrained environment, too, with mainstream products being tightly allocated. In some cases, customers who can't buy material to build their new product sometimes go back and build some of the old product if they can get components. Sometimes they're going back to the well on designs that are four or five years old or older because that's what they can procure components for and book and sell since the new stuff is so constrained. It's kind of the converse of the normal situation where they've advanced because they can no longer buy at the end of life. But I don't know if you've seen those types of examples.

Ed: Oh yeah. There are all kinds of different things that go on. In aerospace and defense, we do a lot of repair for aircraft. We do a lot of aircraft on ground (AOG), and we're FAA certified. When you've got to repair something that may not have been built in the last 10 years, last 20 years, all of a sudden, you're starting to try to find something that that is not so easily found. Companies like Flip help us out every day. Sometimes we have to rebuild it. Sometimes we have to redesign it. The reality is redesigning is very expensive and timely. You're not getting an aircraft back off the ground if you're redesigning aboard. So you need to find it or have an alternative solution to get there for sure.

Bill: What's the craziest story you've had trying to obtain and find old, obsolete, end-of-life parts to support a customer?

Ed: We had a gentleman who rebuilt braking systems for the 737 and the 787. To do those repairs, he would have every design going back 40 to 50 years. He just retired after 60 years with the company. I would walk into his office, and there would be blueprints stacked up to the corners and all over the place. If you said to him, "Hey, there's a 737 serial number such and such down," he would go and find some stack and look through it and finally find the blueprint and say, "Okay, here's what we need to do this - here's what needs to happen." And then if you said, "Well, we can't get those parts anymore," he would say, "Let me go look at the blueprint and see if there's a newer one or an older one that we can switch to." I would say he had no less than 50,000 blueprints in his office, stacked up all over the place. I always used to joke with him that if they ever fell, he'd be pinned underneath there, and we may not find him for a week or two underneath all of those different blueprints. It's very complicated stuff and very costly. You don't think about the cost of obsolescence or not finding a part until you have to do a design that could cost millions of dollars and months and months worth of work.

Bill: IHS tracks end-of-life notifications, as well as PCNs, and they've been noticing a trend, that they've not just been increasing year on year but starting to accelerate. It makes me think of Moore's Law. The types of consumer products that are driving lines today keep those cycles shorter and shorter. It gets tougher and tougher to support the applications that need a platform that's five years plus - in some cases, decades-old platforms. What steps do you think the industry could take to try to mitigate this? I mean, it seems like we're on this train that's just exacerbating the problem every year with faster acceleration and quicker end of life of the older parts.

Ed: I've spent time on the subject, pretty extensively. One thing that would be nice is that as technologies change, that they always have backward compatibility on the new parts. I know they're faster. I know they're different. You design an 8-bit, 16-bit, or 32-bit micro-controller. When you have the next generation released, it would be nice to work with the same software and the same configurations. At least my customers would have to do the least amount of work to upgrade to the newest technology.

It's the same as if you look at FPGA. Those are really expensive parts that, when they go end of life, you have no way to upgrade. You have to redesign from beginning to end completely, and it's expensive. So I would think backward compatibility would help as they make new parts and new technologies come out. And if you think about it, it should be capable of doing that. Our computers worked for five or ten years, even though the technology changes every year, right? Wi-Fi is changing every year. It gets faster. It gets better. But they make it backward compatible so that if your computer is five years old or ten years old, you don't need a new computer just to operate. The first would be backward compatibility.

Second, where I think the industry gets itself messed up is there is no consistent way companies put out the end-of-life notifications and the PCNs. Some companies give a year. Some companies give half a year. Some companies give three months. Since it's not consistent, what happens is customers become very flustered when they get it. They don't know if they have 3 months, 6 months, 12 months to figure out what they're going to do and how they're going to change, or what they need to buy. We just had a part that had been end of life for a year. What we did, interestingly enough, was we found enough customers that wanted to buy this part. Then we convinced the manufacturer to build it for us to provide to customers. We built up enough demand that it'd be worth building one more set of products. But that's the exception. The rule is normally, if it was backward compatible and you got enough notification, you can come up with a solution.

The last I think would be nice is if suppliers would build extra wafers of every set of parts that they built at the end of life, and they froze those parts in nitrogen and held them. Then they would take a company like Flip and say, "Okay, here's the end of life of this set of parts, but here are the wafers for that set of parts." So if customers need them in five years, three years, ten years from now, you can certify those parts and recreate those parts. If you think about it, they do owe the end customer some responsibility. The customer is saying, "Hey, I designed your part in, and now I can't make my product because you end of life the part with not enough notification or with no way to be backward compatible." So those are just a couple of those thoughts. There are more, but if you did those three things, I think that would solve 95% of the problems.

Bill: Interesting. Your first suggestion on backward compatibility - do you see any manufacturers or types of components where that's done today or where they have made attempts to do that effectively?

Ed: Probably the biggest one is more of the overall technology like Wi-Fi. Even though they're going to 5G, you'll still have LTE. You'll still have some of the older technology where things will work with multiple technologies for a period of time. But then they'll come to a point and say, "Okay, when 6G or something else comes out, we're going to get rid of LTE," or "We're going to get rid of this or that." It'd be nice if they lasted longer. I would say memory, that normally works pretty well in the memory space.

Bill: Not the same pinout, just with a double density, and the customer doesn't have to use the excess density, for example.

Ed: Exactly. You make some tweaks to the firmware. You can do this and that, but you're not relaying out a board. They're not relaying out parts and doing a lot of changes. So there are some technologies. And then you have what I would call the bigger technologies, or the bigger chips where, when they make a change, you're redoing your board, right? You have the MPUs, FPGAs, a lot of our big micros, and the graphics chips. When they make a change in those chips, it makes a dramatic change. And then, for a while, it wasn't even just the hardware. Sometimes they would make a change on the BIOS that would make you redesign your board. So, if they could keep the backward compatibility going or had more than one technology built into a chip for a while, that would help a lot.

Bill: Very good. What are some of your other thoughts on industries? You said that you had been kind of a spokesperson for our industry and very involved in industry associations and things for many years. What other burning thoughts do you have on kind of the state of the industry right now?

Ed: I think the industry is going through a change, and I think it's going to be interesting where it all winds up, right? The industry has gone through this whole distribution consolidation. And for us, it's not been positive for the end customer. We have fewer choices of where to buy things. We have less opportunity to negotiate inventory levels and combines. It's made the situation more difficult. And I'm sure for the distributors, based on looking at the financials, it hasn't been financially positive for them. At some point, the question becomes, "Where does it go from here?" It'll be interesting to see, but we try to buy more and more parts. We're a global company. We buy more and more globally from different parts of the world, and we view things differently. Sometimes it's advantageous to us. Clearly, with the margins significantly down over the last couple of years, the distribution channels have been trimmed. Any time you have less choice to buy things and lower margins, we will have fewer choices to get things accomplished. I'm hoping the industry changes that. Hopefully, it goes back to having three or four distributors per product. But I'm not hopeful that it'll happen today, tomorrow.

Bill: And how is SMTC doing? Any news there you'd like to share?

Ed: We just signed an agreement with a company called HIG. We have a definitive agreement to have them invest a whole bunch of money into us and take us private. That's always exciting. And then we've had a good run. I couldn't be happier with the team. We came in here, and the company was $130 million three years ago. We'll do $450 million this year. So it's been a nice growth pattern for us. I think what we do differently than most of our other EMS guys that are our size is that we focus on the customer. What do they need? How do they need it? How do we be more flexible? How do we be high quality at the same time? And we've used the supply chain as a weapon, a positive weapon. We do more combines with our customers than probably anybody. When we look at on-time delivery, it's either there or not there. And that's the way we view things. We probably do things a little bit differently, and that's paid off so far. So I appreciate you asking. Thank you.

Bill: Congratulations on that. That's a great run so far.

Ed: It's been a good run. Let's hope it goes on for a couple more years until I can retire.

Bill: I can't let you off without returning to where you mentioned your book. You have to tell us a little about the book. You've got a chance to make a plug there.

Ed: It's not published yet, but it will be. And it's called An Ordinary Man in Extraordinary Times. And it talks about the journey of finding your way to giving back more than you take from this world. When I was a young executive, all I thought about is how to make more money and how to get promoted. As my wife had passed away and things had happened in my life, it made me reassess that.

Interestingly enough, the more I gave back, the more charity I did, the more I tried to help people one-on-one, the more I got back, and the more successful I was. It taught me not to just worry about myself, but positive things will happen if you worry about the good of the world. As you spoke earlier about being involved with the industry, to me, everybody in our industry should be involved. It's our industry. It's where we make our money, and we should give back. That's been one of the things that I've tried to instill in my children, and that's what the book's really about. It's about changing the cultures of the companies to being more giving than taking from this world.

Bill: I look forward to seeing it. We might have to have you back as a guest to specifically plug the book when you release it.

Ed: You got it. I'll give you one, and I'll even sign it for you for free.

Bill: All right, Ed. Well, I appreciate you being on the show. I hope we can get you back on a future episode as the industry dynamics are just changing every day. What we talk about today is going to be very different 6, 9, 12 months from now. So we'd love to have you back on the show at a later date.

Ed: I look forward to it. Thanks, Bill, and you have a great one.

Bill: I appreciate it. Thank you for joining this episode of Flipping the Channel. I hope you'll join us in our next episode when our guest is Greg Wood, Director of Product Management with IHS. We'll be discussing some of the trends in end of life in process and product change notifications.