Inside the Semiconductor Boom: What Manufacturers Need to Know Before They Try to Serve It
Manufacturing Happy Hour
At a live recording of Manufacturing Happy Hour at SEMICON West in Phoenix, Arizona, host Chris asked two industry veterans how manufacturers, especially in the U.S., can take part in the current surge of semiconductor investment. Danielle, global segment manager for semiconductor at Harding, a connectivity company, and Jeff, principal for digital at Rockwell Automation covering advanced electronics and semiconductor, both describe an industry that is technically ambitious and growing fast. They also describe it as risk-averse, slow to adopt new suppliers, and run on data. Their advice to outsiders: learn the language, partner for the long term, and connect to the fab's operational data.
What a Fab Is, Explained Over a Beer
The host opened by asking how each guest would explain a "fab" to someone outside the industry, such as a parent or a manufacturing leader from another sector.
Jeff calls it the fabrication plant: the place where the semiconductor itself is made, though not necessarily packaged. For people unfamiliar with the industry, he describes it as the transformation of raw material into a complex device. In that sense, he says, it resembles other complex manufacturing plants.
Danielle compares a fab to CERN. It has similar complexity but runs at production scale. CERN's ring is kilometers around; a fab covers a few hundred thousand to half a million square feet. She calls it "a really fascinating way of manufacturing," very different from automotive or food and packaging. She returns to the CERN comparison later, when explaining why outsiders need to invest in learning the industry.
Why the Boom Is Happening Now: A Circular Relationship with AI
The host framed the current resurgence as a chicken-and-egg problem. More chips are needed for AI, and the new fabs building those chips need more intelligence themselves.
Jeff points to two trends converging. First, physics knowledge has grown, so manufacturing technologies differ greatly from those of 20 years ago. Second, demand for increasingly sophisticated semiconductors has grown on the design side. He then plays up the circularity: do we need more AI chips and GPUs because modern fabs need more AI for testing and for running complicated processes? Is the demand coming from the fabs or from the data centers? He jokes that it's "a great business model." More seriously, he calls it a virtuous circle, in which advancing manufacturing technology and rising demand for sophisticated chips reinforce each other.
Danielle acknowledges that opening every conversation with the explosion of AI has become almost cliché, but she says it really is what's driving the industry. AI creates demand for more chips and more advanced chips. That creates demand for more manufacturing facilities, more automation, more connectivity, and more data collection for predictive maintenance and fab optimization. In her account, all of it traces back to the need for more processing power and higher speed.
The Speakers' Backgrounds and What They've Watched Change
Danielle says her first SEMICON was around 1999 or 2000. She lived through the transition from 200 mm to 300 mm wafers, which she says drove the need for more automation. It also drove many of the industry standards for fabs, such as how tools interface, which are common across fabs whether they belong to TSMC, Intel, or Micron. In recent years, she says, AI's growth has brought real changes in how chips are manufactured, especially more complexity on the back end. "We never would have thought we'd be doing lithography on the back end 25 years ago," she says, calling it a great time to be in the industry.
Jeff has worked in the industry on and off for about 15 years. He focuses on data and its role. He describes himself as "a manufacturing operations person at heart." He says he is less literate about what happens inside a process chamber, but he concentrates on concerns shared by every manufacturer: material handling, supply chain, logistics, the economics of running a profitable fab, capacity management, and yield. His broader digital role lets him compare semiconductors with other industries, and he notes how often semiconductors lead the way in new capabilities.
Before and After 2020: Local Manufacturing, Global R&D
The host recalled his own time working with semiconductor equipment manufacturers in Northern California. Back then, the Bay Area was mainly where design and conversations happened, and much of the building took place overseas. Now fabs are appearing around Phoenix, which he gives as one reason SEMICON West was held there.
Danielle says the biggest change since 2020, and since COVID, concerns the industry's global nature. The supply chain has always been global. She recommends the book Chip War for explaining how the industry began in the U.S. and then globalized, including why lithography ended up where it did and why much manufacturing moved to Taiwan and elsewhere in Southeast Asia. In her view, that globalization expanded greatly between 2000 and 2020. Now the industry is seeing more reshoring and nearshoring, with companies doing more "in the region for the region." She names supply chain complexity as one of the biggest changes she has seen over her career.
Asked what excites her about the current situation, she describes an inversion. Manufacturing is being localized while R&D is being globalized. R&D was once concentrated in Northern California and the Boston area. Now she engages with R&D and design engineers in India, across Asia, and in Japan. The host singled this out as a key takeaway: "Manufacturing has been localized, R&D has been globalized."
Jeff focuses on the fabs themselves. He says fabs built 20 years ago were purpose-built for particular categories of semiconductors and look inflexible by today's standards. Their subfabs could not easily adapt water, gas, and electrical energy management. Tools were hard to move within the clean room, and the physical spaces were limiting. New fabs, he says, are built for flexibility and adaptability: to handle changing market conditions, different materials standards, and different ways of handling wafers. He adds that older 200 mm fabs are still running, still profitable, and remain "the backbones of many businesses." He hesitates to call them cash cows but says "they pay the bills," even though they are built very differently from today's fabs.
200 mm vs. 300 mm and the "Two-Nanometer Node"
For listeners outside the industry, the host asked Danielle what the wafer sizes mean. She explains that they refer to the diameter of the silicon wafer the chips are produced on. Most industries think in terms of shrinking, and she notes the irony: the wafers have gotten larger over the last 20 years, going from 200 mm to 300 mm in diameter, while the circuits and features on them have become dramatically smaller, now in the single-digit nanometer range. She notes some disagreement about how the industry characterizes feature sizes. When people say the industry is at the two-nanometer node, that "doesn't necessarily mean the features are two nanometers, but it means they're really darn small."
What Equipment Builders Are Excited About: A Data Snowball
Asked what equipment makers are focused on, Jeff describes another self-reinforcing cycle. OEMs are developing their own ideas about how to "sensor the machine," a verb he uses deliberately. At the same time, chip designers are sending requirements: can you make the chip this way? Once the OEM says yes, it becomes concerned with the maintainability, performance, and health of the tool. The next question is whether it is gathering the right data from the tool, and he says the answer is usually no.
More data to gather means more to connect, more data to store, and more sophisticated analytics for machine health and semiconductor quality. "You see where the snowball goes," he says. Capable machine builders look for ways to apply their capabilities, chip makers want those capabilities, and OEMs are happy to provide them. That in turn creates demand for companies like Harding and Rockwell that help keep assets running. He calls this growth distinctive to semiconductors, because few other products combine changing physics with changing functional demands.
Danielle describes what she hears from equipment manufacturers as "do more with less." Tools are not getting bigger and fabs are not getting cheaper, so floor space is extremely valuable. Whether the component is connectivity, a wafer-handling robot, a lift assembly, or another subsystem, it generally needs to be smaller while being more complex. It needs higher precision, higher speed, and less vibration.
Parallels with Other Industries
Returning to Jeff's cross-industry perspective, the host asked for concrete parallels. Jeff starts with predictive maintenance. Across industries, machine tools are becoming more sophisticated and expensive, which raises the penalty for unplanned downtime. He points to precision manufacturers, such as medical device makers and certain aerospace equipment manufacturers, where a tool going down or a process excursion has a very high cost of quality. For them, moving from preventive to predictive maintenance becomes a compelling case, and AI is being applied to what he calls "the walls of data."
He also names new forms of non-destructive evaluation, beyond vision systems, as new ways of testing whether a product is adequate. Here he says semiconductors lead because of the physics and the tiny scale involved. In his view, inspection and quality management are areas where semiconductor manufacturing sets the pace and other precision industries, such as medical devices, are looking to adapt.
Why Phoenix, and What About Water?
The host raised what he called a possible misperception: Phoenix is not near a great lake or an ocean, so why are water-hungry fabs being built there?
Danielle lists several factors in fab siting. The first is available skilled labor. The second is geography: Phoenix is dry, and "humidity is the enemy of the fab" and of many processes. The third is that there is essentially no seismic activity to contend with. Beyond these, she mentions governmental factors such as incentives, and the preference for politically stable, environmentally suitable locations. She calls Phoenix "the perfect confluence" of these. She acknowledges real challenges around water and chemical usage in a water-constrained place. Fabs are responding by reclaiming and reusing water and by capturing chemical waste and repurposing it, which she notes is expensive. In her framing, the positives are weighed against these challenges, and Phoenix still makes sense.
Jeff says sustainability is usually discussed in terms of water, air, gas, electricity, and steam. What he has learned about semiconductors is that the interrelationships among these resources are much more sophisticated than in other industries. In Arizona, for example, water is a constraint while electricity is less so. He says companies moving into the area are telling the state they have technologies to reduce water consumption: dosing water differently and using different chemistries. He is careful not to claim outright that fabs will trade more electricity for less water. He says only that "those kinds of calculations may be out there," which a heavy-machinery manufacturer probably would not be making. For him, this illustrates how finely the industry can manage scarce resources.
Advice for Outsiders: Learn the Language, Partner, and Be Patient
The final question was aimed at listeners outside the industry, from small family-run machine shops and local integrators to enterprise engineers and early-career professionals. What should they do to be better equipped to serve semiconductors?
Danielle's first recommendation is to build a team willing to dig in and learn the industry, "even if it's a team of one." Returning to the CERN analogy, she says the complexity means someone must learn to speak the language and understand the processes. You can't understand the customer's pain points without understanding what they do. You don't need to be a physics expert, but you do need to know what type of process you are dealing with and what its key needs are.
Her second point, which she says she can't stress strongly enough, is to partner with the customer. This is not an industry where a supplier arrives with the newest thing and the customer adopts it because it is new. The industry is high-tech and forward-thinking, she says, "but it's also very risk-averse because every hour of downtime in a fab is worth tens of thousands of dollars." No one wants to be responsible for a fab, or even one process, going down. Suppliers therefore need to co-develop products with customers and be ready for long timelines. She says tool development timeframes run three to five years: start a project now and you may see production in five years. The reward is that once you are in production, you have a valued partner for as long as that tool is produced. She also warns about the industry's ups and downs and sums up the requirement as "patience."
Copy Exact
The host prompted Danielle to explain "copy exact," which she had deliberately avoided because it is something of a buzzword. The industry operates on the principle that everything must be the same every single time. A supplier cannot change anything about its product without approval from the equipment manufacturer, sometimes all the way up to the fab. She says this includes something as minor as the color of a wire. That makes a design win very rewarding, but it again demands a long-term commitment and a partnership mindset. The host added that most manufacturers would be thrilled to get locked in the way this industry allows. He said he found this unusual when he first sold into semiconductors as a Rockwell salesperson.
Jeff's Closing Advice: Data Is the Currency
Jeff's final advice centers on data. "Semiconductor is defined by the complexity of the data," he says. Every industry runs on data, but this one does so to a much larger degree, and in his words, data is "the currency of conversation." He gives two examples. A maintenance services provider should integrate with the data that shows how the fab is behaving. A materials supplier should integrate its schedules not only with on-hand material balances in the fab's stores but with what is actually happening in the fab.
He concludes that the successful partners will be the companies that speak the language of operational data, "because that is money." He ties this to Danielle's point about partnership: companies that speak the language of industrial data will be welcome partners in semiconductors.
The industry is very high-tech. It's very forward thinking, but it's also very risk-averse because every hour of downtime in a fab is worth tens of thousands of dollars. So, no one wants to be responsible for a fab going down or even a part of a fab, even a process going down.
Today's episode is all about preparing manufacturers for the semiconductor boom. If you're a manufacturer, especially in the US, you are well aware that a number of recent trends are driving unprecedented growth in the semiconductor space. So, how can you get a piece of that action? We're going to talk about that today in a special live episode of Manufacturing Happy Hour recorded at SEMICON West. Let's head to Phoenix, Arizona, and talk about what it takes to get involved in the semiconductor industry over a beer.
Welcome to Manufacturing Happy Hour live from SEMICON West.
So, we're going to jump right in. We'll get to intros in a second for both Jeff and Danielle. We'll learn a bit more about your background here, but in the spirit of Manufacturing Happy Hour, I want to ask you, how do you describe what a fab is over a beverage with someone for the uninitiated? Because for context, there are a lot of people here that know the semiconductor space near and dear. They could answer this question. Honestly, everyone probably has their own unique way of answering this question. But I'm curious from your perspective, how do you describe it as if you're hanging out with your parents? How do you describe it for a manufacturing leader that might not spend every moment of their day in this industry? Jeff, we'll start with you and then Danielle, I'd love to hear your take on this.
The fab is the fabrication plant for production to the folks that look at fabrication of other items. The fab is where we make the semiconductor but not package it necessarily. We're just making the internals. But to folks that are unfamiliar with the industry, it's the transformation of raw material into a complex device. And that makes it parallel to other compound complex manufacturing plants. That's what a fab is for semiconductor.
Danielle, what would your answer to that be?
Well, so I would liken a fab to a CERN project. So it's got the complexity of something like CERN, but on a more production-based scale. So instead of 5 kilometers around, you've got a fab that is a few hundred thousand or half a million square feet. But it has the complexity of something like CERN. And it's a really fascinating way of manufacturing, very different from what we see in, say, automotive or food and packaging.
So in this day and age, why are we seeing such a resurgence in the semiconductor space right now? Let's talk a little bit about this in the context of artificial intelligence. Right? We're making a lot of chips to, you know, we need more GPUs for what we're doing in this space right now, but it's almost like a back and forth chicken and the egg, right? We're building all of these new fabs that need more intelligence. So, Jeff, how would you describe this? I feel like you had a really good way of doing this before the conversation.
Well, I think there's been an interesting confluence of growth in physics knowledge. The technologies by which we can manufacture are much different now than they were 20 years ago. Corresponding with that, the demand for more and more sophisticated semiconductors themselves, from the design, the demand side, has grown. The interesting chicken and egg equation here, I'll be the one to do it on the podcast. Do we need more AI chips, more GPUs because the modern fabs have more need for AI testing for more complicated processes? So is it the demand and the fabs or is it the data centers? It's a circular thing here. It's a great business model. But in all seriousness, the technologies for manufacturing have advanced dramatically, commensurate with that the demand for more and more sophisticated semiconductors has grown, and that's the virtuous circle.
Danielle, anything you would add?
Yeah, I would add that the rise in AI, and it's almost cliche to say this because everyone is, you know, starting every conversation with the explosion of AI, but it really is what's happening. What's driving the industry is driving the need for more chips and more advanced chips, which is driving the need for more manufacturing facilities, more automation, more connectivity, more data collection for predictive maintenance and for fab optimization, but it really all comes back to AI. It comes back to the need for higher processing power and higher speed.
So, one thing I'm interested to do in this interview is to understand how you've seen the industry change over time and how you've seen it change at particular inflection points. But before we get there, to put this in the context of, let's say, the film Office Space, if you will, what is it you say you do here in the semiconductor space, right? We jumped right into the conversation defining fabs for the uninitiated, but we really didn't get to do a proper intro. So Jeff, you've been leading a lot, leading off a lot. Danielle, if you could introduce yourself and briefly your experience in the semiconductor space, and then Jeff, we'll go to you after that. It'd be great to have some context as to what you've seen in this space over time as we get into that conversation.
Yeah. So, I am the global segment manager for semiconductor for HARTING. HARTING is a connectivity company, and I started in the semiconductor industry, I think my first SEMICON was in 1999 or 2000. So, a lot of change. I witnessed the transition from 200 to 300 millimeter wafers, which drove the need for more automation and the standards. Really, that transition drove a lot of the standards that SEMI has put together for fabs and how tools interface and all of that kind of, all the things that are common between fabs, whether it's TSMC or Intel or Micron. So it's been a lot of change, and especially in the last few years, again, as AI has grown, has exploded, we're seeing real changes in the way chips are manufactured in terms of the complexity, especially on the back end, and it's really a great time to be in this industry because it is changing in ways that no one would have thought of in 2000. We never would have thought we'd be doing lithography on the back end 25 years ago.
Yeah. Yeah. And I want to ask you what it was like 25 years ago in a bit, but Jeff, jump in with your little quick background here as well.
So, I'm the principal for digital at Rockwell Automation for Advanced Electronics, Semiconductor, and a couple of other industries. I tend to look at this industry that I've been in off and on for about 15 years around data and the role of data. I'm a manufacturing operations person at heart, and so while I'm not as literate about the goings-on inside of a chamber and the actual fabrication of the wafer, certainly the material handling, the supply chain, the logistics, the economics of successfully running a profitable fab, capacity management, yield, these are all things that are very core to every manufacturer and are no less so in semiconductors. So it's that area that I've looked at and sought to improve through technology for the last 15 or so years, and what we find, and I know we'll talk about this a little later, Chris, is the parallels to other industries and how much semiconductor leads the way in some of these new capabilities.
Yes, you're right. I am going to ask you about more of those parallels as we get further in this conversation, because you are unique in your digital role, broader digital role, that you get to see more than just the semiconductor industry on a regular basis. But let's go back in time first. You know, what was this industry like pre—we're going to draw a line in the sand here—what was this industry like pre-2020s? Because I reflect on my time working with the semiconductor space out in California, Northern California, working with a lot of the equipment manufacturers back then. It was very much, you know, yes, we were designing some complex equipment, building complex equipment, some of it here in the US, but a lot of it was taking place overseas, and the Bay Area was just where a lot of the design and the conversations took place. But now we're seeing fabs cropping up here in the Phoenix area. That's one of the reasons that we're doing SEMICON West here this year. So Danielle, maybe if you can start us off here also. What was it like pre-2020s? Then you can go back as far in time as you want to answer this question. Doesn't need to be just the years leading up to 2020, right? Got it.
Yeah. So, one thing that has really changed, especially since 2020, since COVID, has been the global nature. I mean, the semiconductor supply chain has always been global. There's a great book called Chip War that explains how the industry kind of grew, started in the US and then became a truly global industry, and why some of those changes happened, especially why lithography ended up where it is. A lot of the manufacturing went to Taiwan and to other places in Southeast Asia. It's a great book. Please, if you're interested in the industry, I highly recommend it. But that global nature has really exploded, or really exploded between 2000 and 2020, and what we're seeing now is more reshoring or nearshoring and companies doing more in the region for the region. But the supply chain complexity, I think, is one of the biggest changes that I have seen from 2000 to now.
Follow-up question, what excites you about the current supply chain? Like, what are you doing differently now than you were before in 2020?
Yeah, I think the fact that we're engaging with R&D and design engineers across the globe. So manufacturing is being localized while R&D is being globalized, if that makes sense. So manufacturing is local for local more and more, but R&D has moved from really, really concentrated in Northern California and the Boston area to globally: India, Asia, Japan. Yeah, I think that's one of the biggest things.
No, that's a great point and a great quote as well. Manufacturing has been localized, R&D has been globalized. That definitely is going to be one of the main takeaways from this conversation. Jeff, what would you add?
I think the fabs of 20 years ago were in their time purpose-built for certain categories of semiconductors, and by today's standards you would look at them and say they were kind of inflexible, whether it was the subfab and the ability of the subfabs to adapt to different, you know, water, gas management, electrical energy. The same thing, the ability to move tools within the clean room space, the physical size of the spaces. They were inflexible, again, by today's standards. Contrast that with the fabs of today and the ones that are being built. Never mind ones that are just for very complex chips, AI chips, but there's a level of flexibility, adaptability, ability to accept changing market conditions, different materials standards, different ways of handling different wafers. And so we've learned so much, I think, in general manufacturing, but certainly applicable to semiconductor, that causes us to build fabs that are far more capable and flexible and adaptable, I think, than they were, say, 20 years ago. Which isn't to say those fabs aren't still running and aren't profitable. I don't know whether you call them cash cows, but they're producing products today, 200 mm wafer fabs generally, that are still the backbones of many businesses. They pay the bills, but they're remarkably different from the fabs of today in terms of how they're built for flexibility for the future.
One of the things we do here on Manufacturing Happy Hour is go beyond the buzzword. So another question for those that are outside of the purview of the semiconductor space on a regular basis. What is the difference between 200 mm and 300 mm? Danielle, this is a question for you to answer.
It just refers to the size of the silicon wafer on which the chips are produced. So normally in most industries you think shrink, shrink, shrink. So you think things get smaller, and the actual wafers themselves on which the chips are produced are getting larger, or have gotten larger. In the last 20 years they went from 200 mm diameter to 300 mm, but the actual circuits and the features on those wafers have gotten incredibly smaller. I mean, we're talking in single-digit nanometer ranges. Arguably, there's disagreement about the way the industry characterizes the features, but now we say we're at the two nanometer node. Doesn't necessarily mean the features are two nanometers, but it means they're really darn small.
Yeah. No, excellent for providing context there. I'm very curious what type of conversations you're having in this day and age with the equipment builders. What are the areas that are of focus for them in this new high-tech semiconductor environment? You know, it would be cliche to be like, what were their areas of concern, but what are they excited about? What are they cautiously optimistic about? Jeff, you've been nodding along with every variation I could come up with of this question. So, I think you already had an answer like 10 seconds ago.
Well, what's exciting? I don't know if this is a good one or not, but what's exciting is that the machine builders, the OEMs, are coming up with their own ideas about how to produce, how to sensor the machine. That's a verb, how to sensor the machine. But they're also getting a lot of requirements from the designers of the semiconductors. Can you make the chip this way? So, okay, the OEM, the machine builder, says yes, we can. And then they become concerned with the maintainability, the performance, the overall health of the asset. And then they ask, well, are we sensored? Are we gathering the right data from the tool that we need? And usually the answer is no. There's more to gather, which means there's more to connect to, which means there's more data to store, which means my analytics that I use to determine the health of the machine, the quality of the semiconductor, that becomes more sophisticated. You see where the snowball goes. That to me is very exciting. It's, again, another form of a virtuous cycle here. Not only are the machine builders far more capable and they're trying to find ways for that hammer to find a nail, but the chip manufacturers, the chip designers want it as well, and the OEMs are more than happy to provide it, which then creates more demand for HARTING and Rockwell and other businesses that help make sure those assets are running well. And that growth, I think, is distinctive in this industry, because there are not many other products that are like semiconductors in this regard. Physics are changing, demand of functionality is changing. You don't find that in even the most sophisticated of the products. That makes this industry unique in that regard, I think.
Yeah, there's a snowball effect taking place across the industry in the way you described with the type of data that's within these machines. You described it at the very start with the fact that with all this artificial intelligence that's on equipment that's requiring GPUs, it's like, hey, are we building this because we need more GPUs, or, you know, is it the chicken-versus-the-egg thing? Danielle, is there anything you would add to this portion of the conversation?
Yeah, I would say from the equipment side, what we're hearing from equipment manufacturers is the need to do more with less. So do more in less space. Tools are not getting bigger and fabs are not getting cheaper. So real estate is extremely important. So whether it is the connectivity or it is the actual, you know, wafer handling robot or, you know, spin lift assembly or whatever type of, you know, assembly or subsystem we're talking about in the tool, it needs to be smaller and it generally needs to be more complex, do more, higher precision, higher speed, with less vibration. So more with less is really a big theme of what we're hearing right now.
Jeff, I want to go back to something that you had brought up not too long ago. You
...were talking about what are some parallels we can draw from other industries. You're focusing on a lot of different digital industries right now. Do you have any specific examples that come to mind?
Well, predictive maintenance is always one because most industries will say that their machine tools are becoming more sophisticated, more capable, more expensive. And with that goes the greater penalty of downtime. We were talking about that a little bit before the show and the cost of unplanned downtime. So I think one of the parallels we see in companies that manufacture precision equipment, I'm thinking of medical device manufacturers, certain manufacturers of precision equipment in the aerospace industry. They have machine tools and processes that if they go down or if there are excursions, the cost of quality is very high. So I think preventative into predictive maintenance, those become very compelling cases, and once again AI, we're looking at the walls of data.
So we see that. We also are looking now at new kinds of non-destructive evaluation, not just vision systems but new ways of testing the adequacy of the product. In this case semiconductor leads the way because of the simple physics and sizes of things. So I think inspection and quality management, semiconductor is leading the way, but other industries are looking to adapt, again in medical device and industries that are precision manufacturing oriented. So that's part of the interplay.
We're going to switch gears with one of the next questions I'm going to ask, and that's related to the environmental impacts of all the work we're doing right now. Because when I look at an area like Phoenix, Arizona, for example, and I'm thinking about all the resources and water that's required to allow these fabs to run. It's not next to a Great Lake. It's not next to the ocean, right? So it's not necessarily the most intuitive spot. But maybe help me debunk that misperception that I have. Why are we seeing such a surge here in areas like Arizona? And then what are manufacturers doing to ensure that this is, a term we use frequently, a sustainable practice? Danielle?
So there are a lot of factors that go into especially where you build a fab, and one of those is the availability of skilled labor. Another is seismic activity and actually geographic features of where you are. So Phoenix is dry. There's no humidity here. Humidity is the enemy of the fab, the enemy of a lot of processes. There is no seismic activity really to contend with here. There is a good labor pool here. And of course there are governmental issues, you know, incentives and things like that, but you want fabs in politically stable, environmentally suitable locations, and Phoenix is just the perfect confluence of those.
But there still are issues with things like water and chemical usage and reclamation, and fabs are trying to, and it's expensive to do. But one of the things that they're doing more of is reclaiming and reusing water, repurposing, capturing chemical waste and repurposing that to try to reduce the environmental impact. And there's obviously, you know, a lot of challenges with that, especially in a place where water is constrained like Phoenix, but you balance the positives with the challenges, and Phoenix makes a lot of sense, like I said, for a lot of environmental reasons and then governmental reasons as well.
Add something to that. We look at sustainability in a lot of different industries as usually around WAGES: water, air, gas, electric, steam. One of the things I've learned about this industry is that the interplay and interrelationships between those are much more sophisticated than they are in other industries. For example, while water is a constraint here, electricity is less so. But let's talk about water.
If the companies that are coming into this area are speaking to the state of Arizona, they're saying, look, we've got technologies now that will reduce the amount of water we consume. We'll dose the water differently. We'll manufacture products in new ways that use different kinds of chemistries. They're doing things actively to better manage that one parameter, that one component, water, and it's influenced by electricity. And I won't sit here and say that, oh, we'll use more electricity to use less water, but it's those kinds of calculations that may be out there, which if you were talking about a heavy machinery manufacturer, they probably wouldn't say that. So again, it just speaks to the complexities here of what this industry is able to do, especially when it comes to things like proper management, optimal management of scarce resources like water or electricity.
So I'm going to go back for our final question. I'm going to go back and reference the manufacturing folks out there that are listening to this podcast that aren't as ingrained in the semiconductor industry as we are. What would be your advice to them for being better equipped to serve this industry? And I'll add some context. There are a lot of folks at small mom-and-pop manufacturers. There are a lot of folks at enterprise manufacturers, engineers, folks that are just starting their career that are listening to this. I'm curious to hear your take as to what are things like the local machine shop, the local integrator, the things they could be doing to learn more about this industry and be better equipped to serve it, based on your decades of experience that you've been doing to work with this space. Danielle, if you could lead us off here, and then Jeff, if you can take us home on our live podcast here, that would be excellent.
Sounds good.
Yeah, I think for companies who want to serve the semiconductor industry but aren't familiar with the industry, I mean the first thing is to build a team, no matter how small it is, even if it's a team of one, who is willing to dig in and learn about the industry, because it is so complex. I used the analogy earlier of a CERN project, and it really is not a stretch to say that it is like a CERN project but on a production scale. So you need someone who's willing to dig in and learn about the industry, speak the language, understand the processes, because you can't really understand the customer's pain points unless you understand what they're doing. You don't have to be a physics expert, but you have to understand, you know, what type of process you're working with and what their key needs are.
The other thing is to be ready to partner, and I can't say this strongly enough, to partner with your customer, because this is not the type of industry where you come in with the latest, greatest, newest thing and the customer is going to adopt it because it is the latest, greatest, newest thing. The industry is very high-tech. It's very forward-thinking, but it's also very risk-averse, because every hour of downtime in a fab is worth tens of thousands of dollars. So no one wants to be responsible for a fab going down, or even a part of a fab, even a process going down. So you have to really partner with the customer for new product development and be ready to be in it for the long haul, because tool manufacturing time frames can go from 3 to 5 years. So you start a project now and you'll see production in five years. The good thing is once it goes into production, you know, you have a valued partner there that's going to be with you for as long as that tool is in production, but it's a long journey and you have to be ready for that. You have to be ready for the ups and downs of the industry as well. So, patience.
Yeah, copy exact is a term that hadn't made its way into the conversation yet, but one that's definitely worth mentioning for the folks that aren't living and breathing this space every day.
All right, copy exact. I didn't say it because it is kind of a buzzword, but you know, the industry runs or operates with this mentality of copy exact, which means everything has to be the same every single time. And you're not allowed to change anything on your product, even the color of a wire, without getting approvals from the equipment manufacturer and sometimes all the way up to the fab. So, you know, once you have a design in, it can be very rewarding from that standpoint, but you have to be in it for the long haul and you have to be a partner.
Yeah, I think most manufacturers would be thrilled with the opportunity to get locked in the way you can in this industry. It's certainly something that's unique to this space that I wasn't used to when I first started serving this industry as a salesperson at Rockwell. Speaking of which, Jeff, if you can add your take on this one, I'd love to hear your final piece of advice to the folks that aren't working with this industry on a regular basis.
Semiconductor is defined by the complexity of the data. This industry runs on data. Every industry does, but this one to a much larger degree. In terms of the economics, data is the currency of conversation. If you want to work with this industry, let's say you're a maintenance services provider, integrate to the data that's telling you how the fab is behaving. If you're a material supplier, integrate your schedules, not just with on-hand balances in the fab of materials in the store, but with what's going on in the fab. The successful partners in this industry, and it's not unique to this, but this is leading the way, will be the companies that speak the language of operational data, because that is money. And that's what Danielle said, you said it. I think that's the key. Companies that want to work in semiconductor will partner with them. Just speak the language of the industrial data and they will be a welcome partner.
Well, Danielle, Jeff, I appreciate the advice, the insights that you've provided on this show, both for folks that are in this space and outside of it. I'll tell you what, behind the scenes, for the folks that are just listening to this on audio, we're in very tall chairs, and I've been holding a microphone and looking at my notes on my phone and trying to grab my beer. So I've been getting an aerobic exercise, some ab work, just trying to reach down for my beer this evening. So hey, thank you both for being here on Manufacturing Happy Hour. And cheers to all of you. Thank you for being here at our Manufacturing Happy Hour party tonight. Thank you to Harding Control Soft, Rockwell Automation, and Mouser Electronics for making this all possible. Stay innovative, stay thirsty, cheers, and continue to enjoy the party. We'll see you soon.
Article published