'Texas really started the whole revolution' | State continues to lead the US semiconductor chip industry
Semiconductor chips make modern technology function. Almost all of them are as small as a speck of dust.
Semiconductors have dominated headlines since the pandemic started in March 2020, most commonly in reference to the global chip shortage. The microscopic computer chips control the creature comforts of modern society as well as lead national security, from the touchscreens on our phones to the navigation systems on war missiles.
"I've learned that basically everything with an on-off switch has chips in it," Sen. John Cornyn (R-Texas) said.
Cornyn's not wrong. For example, the car you drive has thousands of semiconductor chips controlling pretty much every single major function, from the brightness of the radio display to updating the digital odometer. They are the fourth-most traded good in the world behind crude oil, processed oil and cars.
Texas has a deep history in the semiconductor industry. The creator of the modern semiconductor chip made the first integrated circuit in the Dallas-Fort Worth area. And a group of semiconductor industry leaders under the name SEMATECH called Austin home.
"In the '90s, there was an organization called SEMATECH that operated out of Central Texas, and it really was a consortium of new technologies to support the industry all over the world," said Robb Misso, the VP of CelLink Corporation. "It was something that we were very proud of."
Despite the near-constant semiconductor headlines, leaders in the industry say many don't comprehend how wide-reaching and important the sector has become.
"The semiconductor ecosystem [includes] many important parts of our economy. For example: cloud represented by companies like Microsoft and Google, consumer electronics are represented by Apple and companies like that. And then even electric cars [are] represented by companies like Tesla, which have a lot of compute. So a car is no longer just a vehicle, it is also a computing device in many ways because it has a lot of sensors," said Dr. SV Sreenivasan, the head of the Texas Institute of Electronics at the University of Texas at Austin.
"All of these important industries – automotive, cloud, [artificial intelligence], consumer electronics and even areas like industrial electronics and health care applications like wearables and then aerospace – these are all areas where semiconductor fabrication is a very important driver for all of these things," Sreenivasan added. "This is not just a supply chain issue. It's also staying at the leading edge because, you know, if you fall behind in the way these chips are made, then the entire ecosystem suffers."
What is a semiconductor chip?:
Experts range in complexity when defining a semiconductor chip, depending on who their audience is. Semiconductors make all modern technology operate and can be as small as 4 nanometers in size. For comparison, your fingernail grows about 10 nanometers in 10 seconds, according to people who have been in the industry for the past few decades.
"If you look at a state-of-the-art semiconductor device made by, for example, Samsung in Austin or Intel, you can fit roughly 1,000 of these transistors side by side across the thickness of a human hair. That's how tiny they are," said Dr. Sanjay Banerjee, the director of the Microelectronics Research Center at the University of Texas. "They are much smaller than, for example, bacteria or human blood cells."
"What you do with that size, which generation after generation you make it smaller, and then you pack more circuits or switches and capacitors and resistors onto this device here that then performs the function of a little computer, or it's sensing biometric data and it's working together to then change your temperature actuate," Ron Martino, the head of Global Sales at NXP Semiconductor, added.
"There's conductors like metals that are used – like wires – and there's insulators that doesn't conduct electrons," said Steve James, VP of fab operations at Infineon's Austin facility. "Semiconductors, we can control whether they conduct or they don't conduct. That's how we can make a computer work."
PHOTOS: Infineon semiconductor chip-making factory in Austin
"You've got people who are building houses and building skyscrapers that you can actually see. Picture kind of doing that type of thing on a silicon substrate that you can't even actually see what you're doing without a microscope or a tunneling electron microscope or [Structured Illumination Microscopy microscope]. It's very difficult," said Jon Taylor, the corporate vice president of fab engineering and public affairs at Samsung Austin Semiconductor.
At the company's Austin facility, Samsung manufactures semiconductors as small as 14 nanometers. At the Taylor facility, the South Korea-based company will be manufacturing so-called advanced semiconductors that can range anywhere from 4 to 7 nanometers. Taylor would not confirm a more specific range for proprietary reasons.
"The best way to describe it, in layman's terms, it's a chip," said Mark Pollard, COO at Astute Electronics. "What does a chip do? You know, normally it's a cellphone example. Do you have any idea what makes your cellphone work, inside this really pretty glass packaging and the, you know, the shiny plastic? Does anyone stop and think what – how in the world it works?"
Chips are typically made from silicon wafers. The wafers stretch about 12 inches across and can be cut into thousands of semiconductors. A process called lithography can cut the design into the chips using a combination of chemicals and machines.
"Lithography is just a very expensive, fancy camera, to make it very simple. And let's say this camera might cost $250 million or more per camera," Misso said. "[Innovation] starts with the ability to print the lines of the circuit, integrated circuit of a semiconductor chip, smaller and smaller and smaller to add more and more devices on an a little small piece of real estate of silicon."
"The lithography equipment is where we coat and develop [a wafer] with a photosensitive material and then it would move on to an etching tool or a deposition tool or a metallization tool," said Larry Smith, the chairman of Tokyo Electron U.S. Holdings in Austin.
Tokyo Electron develops different types of tools for semiconductor manufacturers, most notably lithography equipment.
"We have lithography equipment developers, we have furnaces or thin film capabilities, actors, probers," Smith said. "There are six distinct different pieces of equipment or business units that are part of this ecosystem, that really provides technology that enables life … Our customers cannot make microchips without Tokyo Electron."
"The challenge when you produce [something that small] is getting everything right and eliminating any type of breakage or any type of contamination that could affect the performance of that," Martino advised. "To put that in perspective, a skin flake is devastating at that size. So as you go into an advanced semiconductor facility, it is designed so that the amount of particles in the air are so small and so few that they can't affect this. Which is why semiconductor manufacturing factory is so complex and so costly is because you're producing very small images and you need to keep the environment in the purest state in order for the end product to function."
Advanced vs. mainstream chips:
Most technology does not change as quickly as chips do, leading to a necessary supply of what leaders call "mainstream" semiconductors, or chips that are not the smaller size of "advanced" chips.
"Most of the chips that go into cars are 40 nanometer and above," said Tyson Tuttle, the retired CEO of Silicon Labs. "These are very old, old technologies that are still very useful. And in many, many cases, at Silicon Labs, for instance, you know, we wouldn't want to put in the most advanced technology because it would be more expensive and it wouldn't actually be better."
At Infineon's Fab 25 facility in Austin, the company makes these types of chips for businesses all over the country and the world.
"The newest, most advanced chips are usually going in the brain applications because the CPU, or the microcontroller, those are the ones that really benefit from smaller and smaller transistors, smaller and smaller memory," James said. "Around those chips, or in other applications, you've got power semiconductors to drive those motors, you've got sensors, you have other types of memory that can't go with small chips. For instance, there's lots of other applications that surround that chip. That chip can't talk to the outside world. It can't power itself. It can't send signals, you know, can't connect to the internet just by itself.
"So, you need all these other chips that are built in factories as, you know, maybe they're not the brand-new, state-of-the-art factories because you can't really shrink those chips down to the the sub-10 nanometers, you know, geometries like in the new factory," James continued. "So it doesn't make sense to build them in those factories."
Experts agree any device with advanced chips cannot function without mainstream semiconductors.
"You'll need a mix, always," Banerjee said. "The reason is, with these tiny devices, they cannot handle very high voltages because they will break down electrically. So, there are applications in automotive, for example, where you're not operating the devices at the voltages that you have in your computer chips, which can be as low as a few volts, but instead have to operate at maybe hundreds of watts. So, for those devices, you really need larger dimensions so that it can withstand the higher operating voltages."
As technology advances and society continues to rely on both advanced and mainstream chips, the growth in demand means prices continue to rise.
"If you look at [the] year 2000, about 18% of the cost of a car was due to the semiconductors in it, and now, it's 40%. And that number continues to increase," said Taylor, whose facility creates the semiconductor chips for Tesla.
"It's essential that we have a continuous pipeline of new technologies," Mark Papermaster, chief technology officer at Advanced Micro Devices (AMD), said. "Why do you need new technologies? Because the demand for computing just keeps going up and up and up. So, you know, if we didn't get new chip technology, we'd be flatline. We'd have only the kind of capabilities that we have today and we couldn't grow to meet the demand."
AMD is currently a "fab-less" semiconductor company. It designs chips used in computers, laptops and gaming consoles like the Xbox Series X and PlayStation 5. Chip design, according to Papermaster, is what helps drive innovation and combination of advanced chips and mainstream chips.
"You care about the chip technology because, let's say you're a gamer, you want a vivid graphic experience, you want to be fast, you want to be responsive. You're working on your PC? You want it to be able to solve those problems. You want it to be able to edit your videos and your photo collection very, very efficiently. In a laptop, you want a long battery life, and if you're running your business, you're running on the cloud or a data center, you want it to efficiently run your business operations. That is chip design," Papermaster said.
According to different leaders and politicians, the U.S. makes zero "advanced" semiconductors. Taiwan manufactures 90% of the world's advanced chips.
"That's very unique technology that really Taiwan Semiconductor Manufacturing Company (TSMC) has virtually a monopoly on," Pollard said. "China is just starting to get breakthroughs with that technology. It's very expensive to get the transistors that small. China has exploited that by filling the void for the larger wafer size, and they focused on that. They didn't have the expertise. So, they're now filling the void of the larger wafers for the older technology, which is something I believe the U.S. is going to focus on quite a bit and the innovation to get smaller and smaller and smaller."
"What if tomorrow President Xi decided to invade Taiwan? It would put him in control of 90% of the global supply of these important advanced semiconductor chips," Congressman Michael McCaul (R-Austin) said.
For mainstream chips, Taiwan makes about 60%, China makes about 25% and the U.S. makes 12%, with the final 3% being spread across other countries throughout the world, according to McCaul and industry leaders.
The geopolitics of chips:
In the 1990s, U.S. produced between 30% to 40% of the world's chips. Since then, American production has fallen behind, while innovation has continued to surge.
"If we don't ensure that we have our own independence to create these chips, control the products that they go into, we're really at risk," Pollard said. "We're at risk of falling behind as a technological leader, but there's a true risk as far as the threats of our foes and what they could do and and how they could potentially cause harm."
"Over the years, we had this idea that if it could be made cheaper someplace else, then that was a good thing," Cornyn said. "It is good for consumers, by and large, but when it's something that's absolutely essential to our economy and the technology advances that we continue to see in our national security, that's where it became a game changer."
To try and claw back to semiconductor independence, Cornyn and McCaul led the fight to pass the CHIPS Act in 2020. After going through a few name changes, more than 18 months of sitting in Congress and some renegotiation, lawmakers approved the funding portion of the bill in July 2022 under the name "CHIPS and Science Act."
"This is something we're not accustomed to doing, which is making these huge federal investments and incentives to have businesses come back onshore," Cornyn said.
"If we didn't pass the bill, the CEOs of these companies told me they'd have to go elsewhere. They would go to Europe, they go to Asia, to those markets," McCaul added.
"There was a growing realization that from both an economic security standpoint as well as a national security standpoint, it's important to have semiconductor manufacturing done within the U.S., to a certain extent, and to kind of reverse that trend," said Tuttle, who worked with other semiconductor companies to lobby the federal government to pass the funding.
"Really, the motivation for the CHIPS Act was to provide subsidies to companies to build manufacturing facilities here and also to do advanced research and development around semiconductor manufacturing and to bring that portion of that back to the U.S. And if you look globally, a lot of the countries out there – whether it's China, Korea, Japan, Taiwan – provide significant incentives to build factories in those locations," Tuttle said. "So, it's really a bit like you're leveling the playing field in terms of making it cost-competitive to build those facilities in the U.S., whereas before we would be at maybe a 30%, 40% disadvantage in terms of cost if you built the factory here, which is why the factories were being built in other locations."
The legislation provides more than $52 billion in incentives: $39 billion for production, $13.2 billion for research and development and $500 million for communications and supply chain improvements.
"As far as a country like China that's raced ahead of us as far as production and their ability to do it, how that controls so much of what we need to do to protect our resources, our people, our freedoms, protect our allies, protect our economy," Pollard said. "You know, we tend to swing the big stick in the world and you see that power diminished – that's a scary prospect for a country like the U.S. We've let it slip because we were the innovators of this product. I don't think anyone knew how impactful it was going to be, how the technology was going to evolve to the extent that it is today."
"In many regions of the world, the cost of building that capability can be subsidized, and many people drive incentives to attract companies," Martino added. "By passing the CHIPS Act, the U.S. is taking a step to bring that technology to the U.S. for all the reasons we've talked about because the dependency on semiconductor chips and what they do for our life and our communities is pervasive now."
"The [Texas Instruments], the Intels developed a lot of this early technology, and because it was more cost effective to offshore a lot of this, we didn't think a whole lot about it. The companies were profitable. We still had the cachet of being the developers of this, but all the while, these countries took what we had innovated, built on it, built the infrastructure to really outpace us, both from an educational research and development, to really building the infrastructure to continue at scale for the pace of the semiconductor industry," Pollard said.
Tuttle added that most major global events influence the semiconductor industry, whether they be natural disasters or man-made events like Russia's invasion of Ukraine.
"If you look at globalization and how, you know, any one product that's designed maybe getting components or materials from all across the globe, you know, whether it's, you know, mining for lithium or neon in the case of Ukraine or titanium or gold or, you know, a lot of these raw materials are where the manufacturing facilities are," Tuttle said. "If an earthquake or tsunami or something happens, certainly as a semiconductor person and then in the industry you think, 'Oh, gosh, there's some stuff that's coming from there that's going to impact and how are we going to deal with that?'"
Because of the world's reliance on semiconductors in modern society, Tuttle worries any monopoly within in the industry needs to be spread to other companies and countries around the world.
"Fifty percent of the global fab capacity is in Taiwan. If there was a real conflict that broke out and let's say something happened to one of those fabs or there was an embargo, you couldn't make those chips," Tuttle said. "Inside of a mobile phone, for instance, you might have 20, 30 different chips inside that phone. If one of them comes from a factory in Taiwan and you lose access to that, you can't make that product. If you lost truly lost the capacity that's in Taiwan today, you almost couldn't build anything."
"During the Cold War, nuclear missiles were the mutually assured destruction. Whether we want to call, you know, the economic or the competition with China a 'new Cold War' – but in that competition, the mutually assured destruction is a few semiconductor factories in Taiwan. And it would dramatically impact us that, you know, we couldn't build a lot of the products, you know, a lot of those end products," Tuttle added. "For instance, iPhones are built in China. And you would think that, you know, hostility breaks out, that's where the end products are designed. You wouldn't have access to the semiconductors. It would drive the world into a global depression if that truly came to pass."
Cornyn and McCaul believe the CHIPS and Science Act will bring semiconductor makers back to the U.S., further diversifying where chips are made while improving competition. However, business leaders say the one-time incentive won't be enough.
"Most of the countries that have a big semiconductor industry in the world have their own incentives already," James said. "So, starting with China, Taiwan, Korea, all over Europe, there are already incentives. In a way, we're just trying to level the playing field with the CHIPS Act."
"Technology companies could all move out and put China in a very bad place," Misso added. "So, it'll be interesting to see. It's not going to be driven by nations. It's not going to be driven by politicians. It's going to be driven by companies. And it's going to be driven by the semiconductor companies, I believe. So, in the geopolitical realm of how things are done and the power around the world, I think you're to look at the Samsungs and the Microns and the Intels and the TSMC. They're going to really have a lot of a lot of pull in what happens."
Innovation vs. production:
While the U.S. has fallen behind in the market share for global production, industry leaders in the country have led innovation from the outset.
"Texas really started the whole revolution," Banerjee said. "I started my life at Texas Instruments. The first integrated circuit was invented by Jack Kilby, who, in fact, won the Nobel Prize in physics for our invention of the IC. So, that happened in Dallas, Texas. Texas Instruments was also the company that made the first silicon transistor."
Kilby introduced the first integrated circuit in 1958. Since then, the semiconductor industry has boomed in Texas – especially in the Austin area, which is always growing and innovating.
"Austin has a incredible tech legacy. Semiconductors are the foundation of that," said Ed Latson, founder of the Austin Regional Manufacturers Association. "You know, you look back to Motorola and the work that they were doing, IBM – it's really evolved and accelerated over the last 10 years, especially. You know, now you have world-class companies like Samsung Semiconductor, you know, Infineon, you have this incredible commitment from Applied Materials to the region, Tokyo Electron. They're all incredibly busy right now and trying to keep up with this global demand for chips."
As big-picture demand for chips grows, the short-term demand may ebb and flow.
"There's going to be ups and downs," Pollard said. "There's cycles, [are] very much like the energy markets. I believe that we're going to see a little bit of softening in terms of the constricted supply. It's going to ebb and flow a little bit, but it's largely going to continue for the next 12 to 18 months. We can really only see and predict three-quarters ahead of time, roughly, but it's going to continue. And then, much like an energy correction, it's going to correct. We're going to be oversupplied, whether it's a global recession that just, you know, related to inflation, you know, the interest rates that we're reading about just to control the growth and control inflation."
"Historically, what happens in this market – and this is kind of in cycles – we overshoot the market and all of a sudden then we have a period of time where we have too much production, too much equipment, there's been too many people hired. And we go into a downturn or a downswing in the market," Misso said. "My biggest concern is that's exactly what will happen. Right now, there's a lot of overreaction for us not having the capacity today to support the need from three months ago or six months ago. But I do believe it'll catch up very quickly. And when it does catch up, I think we'll overshoot."
Even if the industry oversupplies in the coming months or years, researchers believe production and innovation go hand-in-hand to address growing demand.
"Innovation feeds manufacturing in five, 10 years down the road," Banerjee said. "You have to innovate, and that's where the universities come in. Because unless we do the leading-edge research and produce the Ph.D.s who work at these companies, you know, you kill the seed corn. You kill the innovation pipeline, if you will. And then manufacturing will stop – not immediately, maybe, but, you know, maybe five, 10 years down the line."
"In semiconductor manufacturing, if you go into the fabs, you're building equipment and developing processes and measuring things at the nanoscale. We're down to sort of a few nanometers … To give you a sense, for the smallest atom, hydrogen is a 10th of a nanometer. So if you lay 100 atoms of hydrogen across, that's about where we are today in what we do in a fab. That's by no means, you know, achieved without innovation," Sreenivasan said.
To ensure innovation continues to happen in the industry, Papermaster sees university partnerships as one possible solution to expand the research and development pipeline.
"You've got such a strong university system, and it's University of Texas in Austin, but it's also, of course, Texas A&M and Texas State and across the region, a great feeder pool of talent," Papermaster added. "That's sort of the foundation. If you can't get talent, you can't be a leader in an industry."
"Innovation pervades all these fields, whether you're doing design or whether you're doing manufacturing. And so, when you say 'America leads in innovation,' not in all fields, not in manufacturing. So, TSMC is today the leader in semiconductor manufacturing and in innovating. So they have gotten ahead," Sreenivasan continued. "And so, you know, it is somewhat naïve to think that we are No. 1 in innovation in every field. You know, there are just a lot of smart people around the world. And in some fields, we are lagging behind. But, by and large, I agree that we innovate in many, many fields. But in semiconductor fabrication, we are not No. 1 today."
Banerjee added that advanced semiconductors are coming down to the size of viruses. Chips have a physical limit to how small they can be just by the nature of the size of atoms.
Tuttle agreed, citing not just the physical constraints, but the economic constraints of making chips smaller and smaller.
"As time has gone on and things get smaller and smaller, the costs go up," Tuttle said. "It actually today is more expensive to design a logic gate or a transistor on a chip. The per-transistor cost is starting to go up because of the difficulty to continue to follow Moore's Law. I like to say that the end of Moore's Law – because we are rapidly approaching the point where we will not be able to shrink further, or if we shrink further, it will cost so much that it won't make sense, it won't make economic sense."
Moore's law is named after Gordon Moore, one of the founders of Intel. Moore theorized semiconductors would basically shrink in half every 18 to 24 months with more transistors on each chip. In the past 40 years, that's held true, according to Banerjee. But as shrinking semiconductors continues to drive up costs and push up against the physical limits of chips, innovators will have to find another way to advance semiconductor development.
"You've seen the evidence of that already because, you know, when I started in the field, there probably were a few dozen companies in the U.S. which worked on semiconductor devices. And over time, many of them have kind of fallen by the wayside," Banerjee said. "Scaling of these transistors will stop long before you reach the sizes of individual atoms because other limits of quantum physics start playing a role."
Growing pains for a strong industry:
According to the Texas Comptroller for Public Accounts, the semiconductor sector of the manufacturing industry is one of the fastest-growing areas of the Texas economy. In 2020, the industry contributed $15.3 billion to the Texas GDP. Texas semiconductors made up 15% of the U.S. semiconductor GDP in that same year, according to the comptroller's office.
In 2020, the semiconductor industry made up more than a quarter of all Texas exports. In terms of jobs, more than 41,000 Texans work in the sector. Fourteen-thousand of those jobs are in Travis and Williamson counties.
"That kind of rapid growth is really good. That does put some short-term pressure on the workforce," said Bryan Daniel, the chairman for the Texas Workforce Commission. "There's really just this need for us to do a continued education, reevaluation of our skills and just upskilling and reskilling in kind of a continuous basis in the workforce."
As more companies in the industry move to and expand in Texas, that means they all pull from the same pool of resources for the workers, water, energy and raw materials needed to make semiconductors and the machines that make those chips. To widen the pipeline of workers, state leaders like Daniel, as well as industry leaders, look to job-training programs, school districts and local colleges to train the next generation of semiconductor employees.
"Those jobs have now changed," said Dr. Laura Marmolejo, associate dean for manufacturing education at Austin Community College. "Everything's robotic, controlled automation. And so, all the jobs now require some degree of knowledge."
"Central Texas has brought a lot of high tech businesses into a single area like a Silicon Valley, and they also are investing in different educational initiatives that helps build a strong workforce population that can work in the factories and related jobs," Martino said. "Also, there's a university, very strong university capability in Austin and Dallas and throughout Texas that helps create innovative, innovative environment."
"Even the very entry-level ones need a little bit of knowledge because these machines are much more sophisticated than they've ever been," Marmolejo said, adding, "It's not all engineers. There's so many levels that people can get into. So, if you're the guy who just doesn't want to go to college, no problem. We can get you in that industry in eight weeks with a short program that gives you some fundamentals."
Workforce development will need to lead the growth in the industry, experts and leaders agree. But that means supporting that workforce, too.
"Some of my concerns, having spent a year in Silicon Valley, I know very well what it can end up being, and that is the massive growth," Taylor said. "You know, you've got to have the roads and you've got to have the housing and all those things. The technology in these great companies are going to bring people in, but we also have to care for the people who are going to live here. You know, when I worked in California, I had technicians that would travel three hours one way to get to work in Santa Clara and then three hours home. They're on the road for six hours a day."
Those looking at the big picture for the industry agree that there needs to be support systems in place that allow the growth to continue.
"We're going to grow into this, but we've got to be deliberate in that growth," Daniel said. "We can't hope for the best. We've really got to do some planning, and we've all got to kind of get on the same page. That ripple effect for the economy becomes part of the economy at some point."
The ripple effect in the economy leads to job growth outside the semiconductor industry too.
"The semiconductor industry sector has a jobs multiple of seven, which means we roughly have 250,000 Americans working directly in the semiconductor industry, but they also spin off 1.8 million more jobs," said Tony Bennett, president of the Texas Association of Manufacturers. "So that is 250,000 direct jobs times seven, you get the 1.8 million that is critical to all of our sectors. But, you know, this is one sector that produces an exponentially large amount of indirect jobs, which is important for the U.S. and the Texas economy."
Between the direct and indirect jobs, James added there are those jobs that support the semiconductor manufacturers like Samsung, Infineon and NXP in Austin, like Applied Materials and Tokyo Electron, which make the machines to make the chips.
"There's, you know, four to five times ancillary jobs that are supporting us that are out there," James said.
Even as a design company, Papermaster noted the industry in Central Texas both works together and competes against each other.
"What you'll find when you look at the semiconductor industry is it's a partner-and-compete model that all of us have in the industry," Papermaster said. "Of course, we compete. We compete on the design of our products and the markets that we go after and tailor those products to. But we partner on helping make sure that common problems can be addressed, and so that is that the universities can provide us the kind of skilled pool that we need, and that's to make sure that we have the right standards so that any one of our chips can work [together] with one of the other [company's] chips."
While semiconductor facilities hire thousands of people, they also use a huge amount of resources. Experts are hopeful.
"The workforce aspects, you know, Texas is very favorable there," Tuttle said. "The construction costs, the land costs, the availability of water and power, which each of these factories requires, you know, reliable and substantial energy and water resources to be able to continue to grow."
"The past couple of years, certainly, we had some issues with the energy, you know, for Winter Storm Uri. But I feel pretty confident those are resolved," Latson said.
"We had the winter storm, and now the war in Ukraine can have some effect on raw material supplies, where chemicals go, and what's available and that sort of thing," Bennett said. "There seems to always be something here of late that is challenging the industry."
While the industry faces these challenges head-on, a combination of new research and near-constant growth in demand may lead to another issue: what to do with older chips that are no longer made but are still needed in products like laundry machines or used cars.
"You think about all this demand, demand, demand, or people keep consuming. Where does all that go? Because this is actual hardware, even if it's getting smaller, where does it all go at the end of the day?" Pollard asked. "The e-waste, the e-recycling, the environmental impact of all this material is a huge question. I don't have the right answer for it right now … But when your phone becomes obsolete, doesn't work and it's already been changed hands three or four times and it's finally just dead, how many hundreds of millions of phones and computers and devices are like that? What does the industry agree upon and what do the governments agree on, as in how do we manage that? And that's something that's probably not a priority in the current situation, but should come up and should be discussed. And we want to be on the forefront of that. I believe in the near future there's going to be requirements on companies that are procuring chips for their assemblies and their devices, and a responsibility to what happens at that product at the end of that life. And like we do in the supply chain, we will probably help facilitate that process in some way."
With these water, power and workforce challenges, the cost of building new facilities continues to rise, too. Samsung has already dedicated $17 billion to a single new fab in Taylor. That price tag is consistent across the industry.
"Most of the factories, you know, the most expensive tools in any of these factories are these photolithography tools," Taylor said. "We're talking over $100 million for one tool. So when you build a factory, about two-thirds of the costs are actually the equipment for the process. About one-third of the cost is the building itself."
Within the building and inside the manufacturing area itself, companies have to install so-called "clean rooms." Even a single speck of dirt or static electricity can ruin an entire wafer of chips. When workers arrive for their shift, they put a "bunny suit" on over their clothing, walk over a rubber mat to ground their natural electric charge and enter through an air shower that sucks away any spare dust or particulate matter.
"Why is that important? That's important because if one tries to make these computer chips in a room like [my office], you know, to our naked eyes, it looks pretty clean, right? But if you used a high-powered microscope, you would see these dust particles floating around, and these would be like gigantic boulders that are falling on your chips when you try to manufacture them, and nothing would function properly," Banerjee said.
The semiconductor industry won't go away for decades, according to many of the people interviewed for this "CHIPS 101" project. For the time being, almost all are optimistic about its growth and the role Central Texas will play.
"We're adding the technologies of the future," Latson said. "You know, we have really interesting things happening around energy electric vehicles. You know, communication and semiconductors are empowering that. But they're not only powering it, they're being made here as well. So, you know, when I think about the future technology, I think Central Texas is really well poised for several decades of growth and being a significant contributor to a global economy in a way that most regions can't be."
PEOPLE ARE ALSO READING:
Uvalde Police investigating shooting at city's Memorial Park; two juveniles being treated in San Antonio hospitals