Applied Materials, Inc. (AMAT) Earnings Call Transcript & Summary

Thanks, Gary. As many in the audience may know, there are more than 1,000 semiconductor fabs operating globally today. Together, these factories produce $450 billion worth of chips a year and about 50 million metric tons of CO2. In semi manufacturing, our carbon footprint is made up of Scope 1, which is fab direct emissions to air; Scope 2, electrical energy use; and Scope 3, energy-related to the consumption of process chemicals and consumables produced by suppliers upstream. Considering all these aspects, with direct emissions properly abated, by far, the largest component of a fab's carbon footprint is Scope 2, electrical power. Let's consider an advanced fab with a capacity of 50,000 wafer starts per month. That fab is consuming roughly 1 terawatt hour per year, close to the electricity use of a city of 100,000 residents, like Santa Clara, California for example. When we break that power usage down, about half is directly consumed by the tools themselves, and the rest is used to run the tool support systems, facilities and clean room. And that is the industry's sustainability challenge in a nutshell, Gary. I hope that is helpful.

Thanks, Sarah. That's a great setup. As we work with our customers, there are 3 main initiatives we are driving. First, we're improving the eco performance of our products. An easy way to think about this is that a typical advanced process tool in high-volume production consumes about 1 million kilowatt-hours of equivalent power a year. To improve eco performance, we've developed hardware and software upgrades to reduce energy and process chemical use. Many are available as upgrades for the installed base as well as in our new tools. Second, we've developed advanced service products that help customers optimize their installed base eco performance as well as shorten ramp times, improve device performance, yield and cost. Our service group works with customers to tune their systems to peak performance. Third, our next-generation of tool architectures significantly improve throughput density. That's the number of wafers processed for a square foot of clean room space. We've set an ambitious goal that we call 3 by 30. On a per wafer basis, we're going to drive a 30% reduction in equivalent energy consumption, a 30% reduction in chemical consumption and a 30% increase in throughput density per square foot of clean room space, all by 2030 averaged across our semi product portfolio. Making improvements of this magnitude and, at the same time, driving the technology road map forward is not easy and requires deep partnerships with customers. So let's hear from one of them that is leading the way in environmental responsibility.

At Micron, our vision is to transform how the world uses information to enrich life for all. We embrace our responsibility to do this through the products that we develop and by ensuring the sustainability of our operations. We have identified a multipronged approach to combating climate change. We are developing more power-efficient products that can enable new high-performance computing architectures with dramatically lower power consumption. The scaling of our technology nodes delivers greater memory capacity from each kilowatt-hour of energy consumed in our fabs. Every year, we increased bits per wafer by a higher percentage than the resources required to produce that wafer. Finally, we are investing 2% of our annual CapEx amounting to approximately $1 billion over the next 5 to 7 years to enhance energy efficiency, greenhouse gas abatement as well as waste and water recycle infrastructure at our facilities globally. We are also integrating sustainable design into our future node development. By starting early with our partners, including Applied Materials, we can enable the processes and factories of the future that will help us create and realize a more sustainable world.

Thank you, Manish. At Applied, we're passionate about this topic, and we want to work with companies who share our vision and goals. As well as working closely with customers, we're also making important changes to the way we partner with our supply chain. Today, we're pleased to announce our SuCCESS2030 program. This is our shared 10-year road map, and there are 3 key components. First, we're going to hold our supply chain to the same standards that we hold ourselves in the areas of environmental impact, labor standards, and diversity and inclusion. We're introducing a sustainability scorecard into our supply chain selection process alongside our traditional metrics for performance, cost and quality. Second, we're going to improve together handoffs between our suppliers and us. Here, the call to action is reduce, reuse, recycle. And third, we're going to partner with our suppliers to share best practices and key learnings. The response has been great, and we have 6 key partner suppliers already signed up to help us kick off this program. Thank you to all of these companies for sharing our vision and passion to make possible a better future. I also want to deeply thank all of our suppliers for stepping up during the COVID crisis. We really appreciate it. COVID-19 restrictions have stimulated many changes in the way companies in the industry are operating. While these actions are being driven by the need to achieve our near term goals, I'm very excited about the long-term benefits of working in new ways, especially remote support and remote R&D. In a traditional way of working, flying an engineer from the U.S. to Asia and back generates about 2 metric tons of CO2. Further expanding secure data sharing to enhance remote support and remote development is a major opportunity for the industry. I strongly believe that innovate anywhere should be a significant and permanent change to the way the industry operates. It's a game changer. Over the past several years, Applied has made significant investments in state-of-the-art digital infrastructure, sensors and metrology, data science, machine learning and stimulation. The combination of these technologies enables us to reduce product development cycles, speed up transfer of new technologies from lab to fab and optimize cost, output and yield for our customers in volume production. While operating our businesses the right way is incredibly important, our impact on the world goes well beyond our own labs and factories. When you think about the world's grand challenges about how we make the world a better place, semiconductor technology has a critical role to play. Today, I will focus on one specific theme, AI, because AI has the potential to change everything. It can provide us with the analysis, prediction, actionable insights and decision-making capabilities well beyond the abilities of human thinking or traditional computing. We can accelerate research to better understand climate change, diseases and public health. And smarter manufacturing, smarter health care, smarter vehicles can create a safer, more secure world, where we optimize the use of precious resources and reduce waste. The applications and possibilities are endless, but AI has an Achilles' heel that, unless addressed, will prevent it from reaching its true potential. That Achilles' heel is power consumption. Training neural networks is incredibly energy intensive when done with the technology that's available today. Last year, at this event, I shared our perspective about the magnitude of this problem. Since then, we've worked with our partners to refine our assumptions. Today, traditional data centers consume about 2% of the world's electricity. Our models show that if we build out AI data centers using current-generation technology, they could potentially consume 15% of the world's electrical power by 2025. This model is dominated by a new and fast-growing workload, training AI models with visual data, but there is good news. This is a chart that I first showed 2 years ago that describes the gaps between computing solutions that are currently available for AI and where performance needs to be. At the time, we said there was a 1,000x improvement needed in the compute performance per watt for training in the cloud. When we update this chart for 2020, you can clearly see progress is being made, but there is still a long way to go. The other good news is there is a lot of promising new technology in the pipeline and a vibrant ecosystem of innovators, including many new startups focusing on accelerating innovation at the edge and in the cloud. Customized design will be critical. New system architectures, new application-specific chip designs, new ways to connect memory and logic, new memory technologies and in-memory compute can all drive significant improvements in compute performance per watt. To unlock the potential of edge and cloud technologies, major advances in the power, performance and area cost or, PPAC, of semiconductor devices are needed. This playbook has 5 elements: new system architectures, new devices and 3D structures, the introduction of novel new materials, new advances in 2D geometric shrinks, and new ways to connect chips together through advanced packaging. Based on recent input from customers, we've updated our PPAC framework to PPACt, where t stands for time to market, acknowledging the enormous value of speed. Applied has, by far, the largest and broadest portfolio of technologies and products to accelerate the PPACt playbook, which spans creating, shaping, modifying, analyzing and connecting structures and devices. Let me give you an example using a product that we're launching today that illustrates how a new material's engineering technique selective deposition is being used in conjunction with advanced lithography and patterning to improve PPAC in leading-edge logic devices. Transistor innovations in materials and structures is driving higher and higher performance. However, as the wiring inside the chip gets narrower, electrical resistance grows, increasing power consumption and impacting performance of the chip. The modern transistor becomes like a finally engineered racecar stuck in traffic. The most critical issue today is at the contact, which is the wiring that connects the transistors to the rest of the chip. Until now, the contact wire has been formed by creating a via in the dielectric layer above the transistor, cladding the via with a titanium nitrate liner and adding a tungsten nucleation layer. These act as a mold as the via is filled with tungsten. Unfortunately, that mold doesn't scale, and it's taking up a larger percentage of the via space. New innovation was needed, and our solution was to eliminate the mold altogether. It's like going from injection molding to 3D printing at an atomic level. Our selective deposition technology creates thicker contact wires, which is like widening the roads again for our racecar transistor. This seems like an obvious solution, but in reality, it's incredibly difficult to do. We need to create an ultra-pure environment that's hundreds of times cleaner than the clean room. We need to carefully pretreat the surfaces. Then we need to selectively grow the tungsten atoms, so they perfectly fill the via from bottom to top. There are hundreds of billions of contacts across the wafer, and they all need to be perfect. So the entire process flow takes place under ultra-high vacuum to keep impurities out and to optimize material structures and interfaces. This next-generation of equipment, which we call Integrated Materials Solutions, is one of the ways Applied is bringing to bear our broad technologies and capabilities to address our customers' most complex challenges. As I wrap up here, I want to again emphasize that I strongly believe our responsibility as leaders is to leave the world in a better place. At Applied, we're taking actions that are aligned with our vision to make possible a better future. Today, we announced our goals to move to 100% renewable energy and reduce our own carbon footprint by 50% over the next decade using science-based targets and transparent reporting. We're working with our customers to increase the performance of our tools, eliminate waste and drive productivity. We launched a new 10-year road map for supply chain sustainability, our SuCCESS2030 program. And most significantly, we focused our company, our strategy and our investments to build a strong pipeline of meaningful innovations that will accelerate the PPACt road map and enable sustainable AI. To achieve our potential, across our industry and the electronics ecosystem, we need to think and act differently. We need to break down barriers from materials to systems and connect dots in new ways between system designers, integrators, chipmakers, equipment and material suppliers. The challenges the world is currently facing are an accelerator for key technology inflections, and we will see significant and permanent changes in the way companies operate. So my call to action for all of you is to innovate anywhere and collaborate everywhere and join us as we make possible a better future. Thank you.