MALAGA GPEF Keynote 2: Matt Crowley JABIL

Thanks, and thanks to the Optica team and Jose for having me. Our CEO, Mike Dastoor, had an unanticipated surgery a couple weeks ago, and so I have the honor of stepping in for him, which excites me in a couple different ways. One, I have a photon joke, and I've been telling this photon joke to my 9, 12, and 15-year-old boys for a while, and it gets the kind of reaction that you would imagine. So I'm hoping I finally got the right audience for this joke. So here goes: A photon checks into his hotel room. Bellhop comes up and says, 'Photon, can I take your bags up to your room for you?' He said, 'No, I don't have bags. I'm traveling light.' Yes, all right, well, okay. The boys will be happy to know it got about the same reaction that they had, so that's good. No, the other reason that I'm super excited to be here is that, you know, silicon photonics is a huge part of our strategy at Jabil, and so having the opportunity to talk to you all today about the role that scale plays in the industry is super important for us. I think it's super important in general if photonics is really going to have the impact that it has potentially. So, a little bit about Jabil. It's actually Jabil. We are a large enterprise, roughly $30 billion, $29 billion annually. We just reported earnings for fiscal '24 last year. We have 140,000 employees all over the world. We have factories in over 25 countries, and we've been around for 50 years. And so that is all to say that we understand scale and we understand longevity. Additionally, we have a ton of end markets and they're very broad. So in addition to the business that I'm responsible for, where we have our wireless, 5G, cloud, hyperscale business, we also have businesses in healthcare, in automotive, in defense and aerospace, in consumer. And so we have a very unique view of not only scale but of the marketplace because photonics plays a role in a number of these different areas, not just in technology in the hyperscale space. We really feel like photonics has the capability to change the industry, and so we kind of think about some core end markets that we participate in, specifically around cloud and data center, in telecom and mobility, in automotive and transportation where you have LiDAR, and healthcare where you have optics helping to precisely navigate the human body, and in consumer electronics around AR and VR. And so all of these end markets have the potential to take photonics to the next level, and it's a big level. So it's discreetly $35 billion from a component opportunity, but it supports a $2 trillion market. So everybody in this room has a massive opportunity to go and take advantage of what photonics means to the industry. So we're super excited about it. Now, when you think about scale, there are some requirements from the market that you have to be able to deliver, and so I'm going to touch on each one of these. Pricing is obviously a core component of it. There is a geographic diversity requirement, which I'll touch on briefly, but the ability to quickly scale and create volume is critical. So we'll go through one at a time. Now, for all the economists in the room, you're going to have to give me a little leeway here because these models are probably not accurate necessarily, but they're illustrative of a concept. And so traditionally, you've seen and are used to demand-supply curves where volume and pricing adjust, price points come in line, volume starts to take off. But what we see in the industry today is that's changing because you have mega customers, hyperscale customers, they expect something different. They don't adhere to the traditional models. In their mind, they want to see pricing and scale all at once. There is no curve. They want the price point now. They want the volume now. Delivering into that is a significant challenge, and so you can't go about thinking about it in the traditional way. So we've thought about it a little bit differently, and this is true generally of our portfolio. But historically, a lot of product companies will look at a product, the price, the materials, the markup, the interest, and they'll target a specific profit per widget, let's call it. And that traditionally works, but with the hyperscale space, you can't think about it individually. You have to think about your portfolio a little bit differently. So for us, we focus on return on invested capital because we can create velocity inside of that, return cash to our corporation, which ultimately is the earmark of a good business. And instead of focusing on just lower margin widget business, we focus on the economy that we create and the cash that we return to our enterprise. And so typically, where you would see inventory turns at two to three times per year, we're focused on turning inventory five to six times per year, and we're focused on return of invested capital in upwards of 90 to 100% back to the company. And so that creates the ability to manage a portfolio. So the focus on the traditional margin per unit is less impactful when you're returning more cash to the company. And we've done that successfully in our server and rack business. We are doing it now in our silicon photonics business, and it creates a really nice complement to the portfolio. Okay, so what does the market require? Lower price points, geographic diversity, and I'll just touch on this briefly. Everything happening in the world today has required the de-risking of locations for manufacturing. That's just table stakes. Specific to silicon photonics, there's a bit of a nuance in that you have parts that move all over the world because you don't have one centralized ability to take a silicon photonic part through OSAT in one place. And I'll touch more on that in the scale piece, but you've got to have the ability to de-risk dependency on one region, and that goes back to our footprint all over the world where we have 100-plus factories in 25 countries. So it fits well into our model. When you think about scale and velocity, so there are three challenges, and we'll go through each one for photonics OSAT specifically. The ecosystem is fragmented, we'll talk more about that. Automation has not been deployed at scale. It needs to be in order for this business to be economically viable. We have to create that capability. And the capital intensiveness of this business is significant. So how do we address that and still adhere to the return on invested capital tenet that I just talked about? We'll go through each one. So the fragmented value chain today, you'll have a part that goes to a number of different touch points all over the world. Couple problems with that. One, a lot of times you may have fiber hanging off of these parts. The risk of scrap is significant, and customers don't like a 50% loss on parts that in some cases cost up to $25,000. So how do you move the parts around the OSAT process without the risk of scrap? Additionally, each time these parts move, it's hitting a different supplier potentially, which creates a margin stack, and that margin stack is not something that customers are super interested in either. So figuring out how to address that, finding synergies, finding ways to consolidate, finding ways to do more of this process in one place versus all over the world is critical to the ability to scale. Automation. So today, automation requires a whole lot of parts to be fed through it. We don't really yet have the right alignment cycle times. So if you think about a machine that can do 30 parts per hour, you would have to have a whole lot of those machines in order to serve a hyperscale customer. That creates a capital intensity that becomes problematic to the return on invested capital model. On the other side of the automation piece, you have bottlenecks that are human beings. And so while you have some automation that you can deploy, you still have people that are feeding that automation. And so solving the automation problem becomes core to our ability to create scale in the space. And capital intensity. So when you have these machines that are massively cost, they are massive cost, you have to deploy a ton to be able to hit the scale that customers want. And in order to have the right line for silicon photonics, you're talking about $250 million potentially for 150,000 wafers per year. And so how do you create an economic model that makes sense in this environment? It's a challenge, and there's probably a big part of this story that a lot of the folks in the room may disagree with. That's okay. But we have to find some alignment across some core verticals. So technology owners in the room, and I know nobody wants to hear this, but if you're innovating around technology from day one, in order for photonics to really scale and really impact the industry to the degree that it has potential, you've got to be thinking about what does it take to succeed at scale with volume. It can't just be about the technology because the technology can be absolutely amazing and differentiated, but if you can't get to the hyperscale volume and the hyperscale expectations, the technology is not going to get into their fleets. For manufacturers, we've got to think about known good workflows. We have to find leverage from one generation to the next. We have to be able to reuse the things that we build and create as we go through the process from generation to generation. And then we have to consolidate the value chain. We have to be able to eliminate the risk of scrap in the OSAT process, and we have to find those synergies in order to make this economically viable. And then finally, on the end-user side, we've got to get to the point where there is some level of standardization because if every single customer is insistent on a snowflake-like process and/or product, we're never going to be able to get to scale. Now, the truth is in the hyperscale space, probably only takes one. Not going to name them, probably some folks in the room from them, but if you get that one, you can make it work. But if you find some level of standardization across all of these processes, then you have the ability to make this incredibly economically viable. You can focus on the return on invested capital, and you can focus on turning your inventory and returning cash to your company more frequently. And so that's what we do at Jabil. It's what we're really good at, and it's been very successful for us to date. And so we think there's a ton of opportunity here. I think the folks in the room, if we can find ways to address some of these issues, then photonics will be a massive impact to the industry. But we've got a lot of work to do to get there, and hopefully, we'll have conversations for the rest of the week to talk about how we do it. That's all I got.

Well, look, I'm not the most technical person, but what I would tell you as a concept is if photonics is going to accelerate and take over the industry, then we absolutely have to find ways to get to price points that exceed copper. So how we do that, I think our next panel is probably better suited to speak to technically. From a business perspective, is that how silicon photonics wins in the space? Yes, absolutely.

Well, we think about it in the context of is there a way to get not from, call it, 20 hops today all over the world, but to get to three? Because if you can do that, you can eliminate the margin stacking, you can eliminate a lot of the risk on scrap. And so I don't know that any one company is going to solve it. I think there is still a problem of volume as it applies to the big OSATs. Are they really interested in the volume around silicon photonics specifically? If we were to be able to form somewhat of a consortium of a tight-knit group of companies who would then address those critical few processes that you could aggregate, then I think that's the way to address it.

I think there's a starting point ASP. I think that competition in the space will always have a downward force on it to try and take cost out. I do think that there is some value in the fact that customers will always want or need to have diversity across geographies, and so becoming dependent on Asia-specific suppliers and allowing consolidation with the China supply base probably not going to happen. I think you're going to have to chase that market a little bit because if the price point discrepancy in the ASPs become too different, it's going to really hurt the fact that the geographical diversity has value. But I think you can do that and still remain relatively profitable, especially if you're focused on the return on invested capital.

I don't know. I think for silicon photonics, there has to be some consolidation, and I think it's going to be between the China supply base and the non-China supply base for the fabs. I think there's going to have to be an inflection point at which there is enough commonality around the technology that you don't have five different fabs trying to deploy and design five different specific technologies. And so the more commonality that we can find so that the price points become in line with customer expectations in the hyperscale space, the more volume you'll see. And I think it's just going to have to take some level of consolidation or aggregation to get there.