中文 English

The Big Data Shift Has Started

Brewer Science’s CEO looks at what’s changing across the industry, how long those shifts will last, and who will benefit from them.

popularity

Terry Brewer, president and CEO of Brewer Science, sat down with Semiconductor Engineering to talk about different priorities for private and public companies, why AI completely changes the game for technology companies, and what impact materials will have on innovation and design in the future.

SE: What are the next big opportunities for Brewer Science?

Brewer: There are broad opportunities on multiple levels. For employees, it’s a matter of changing the work environment through education. We’ve been around for 37 years. We now have three different generations of people inside our organization. We have to have innovation to fulfill their success. That requires a lot of creativity and new ideas to manage the culture. For the community, particularly our small rural community, there are infrastructure challenges, living challenges. We have to face many of the same problems with health, security, well-being, food supply, but we have less infrastructure to do it in a state like Missouri than in places like California or New York. As private company we play a larger role in supporting that infrastructure. Another element is the customer, which is undergoing a transition from focusing on making things better to getting bigger through consolidation. In the long run, that doesn’t work. These cycles come and go, but do they last a long time—sometimes 20 years or more.

SE: There also are a lot of new markets that never existed before for electronics.

Brewer: Yes. AI is one of the big drivers. IoT is another driver. Those are pieces that will drive business and business opportunities forward at a pretty fast pace. Right now, that’s on the hype cycle. How much AI and IoT do we really have? We’re going to have self-driving and self-correcting cars. Those are coming, but we’re not there yet. We’re still using old technology. The concept is heading in the right direction. It’s like computers. When we first had desktop solutions we couldn’t do graphics. The type face was small and there was a single font. They crashed all the time, and every time the software changed, you had to start over again.

SE: So where do you see this heading?

Brewer: One of our products is sensors. Those sensors are two-dimensional, but what they sense is the third dimension.

SE: You’re talking about the Z axis?

Brewer: Yes. Those sensors also are small, flexible, and they can be integrated onto or into anything. They’re also really fast, because you can generate 30,000 pieces of data a minute. What are you going to do with 30,000 temperature readings? Nothing in today’s world. But as soon as you add in AI, it changes the game, because now you can turn it into predictable and profitable information. The world does not exist in single-digit measurements. So what happens if you can take your blood pressure or heart rate or moisture content on your skin 30,000 times a minute?

SE: What’s the shift you see here?

Brewer: Making it predictive. If you look at something like lithium-ion batteries, we know one of the problems is those batteries sometimes get hot and occasionally they explode. When it gets too warm you can predict a problem. But with a sensor on the battery and AI, you can see fluctuations in temperature changes and predict, based on a sequence of events, that the battery is going to get hot. And you can do that even before it starts to get warm. That changes the game. Now you have preventability, and you can do that under all kinds of conditions. So now you start looking at the environment as a conditional environment, not as a fixed environment.

SE: At this point programming that into algorithms is still very difficult.

Brewer: Yes, but this is where things are going and how innovation is changing the world. It will take more software and AI and new kinds of sensor, but all of that is coming. On a higher level, you’re not going to have to ask for a temperature or stability or moisture content reading. People will start to think in terms of distributions rather than points. So what is the distribution of your power usage or your energy profile or security. It’s a scientific leap because this is quantum mechanics. There is no absolute. Everything is determined by distributions. That’s can predict things more accurately than fixed numbers. Once that happens, we will understand our world a lot better.

SE: How do you see this being applied in the real world?

Brewer: I was raised on a farm, and sometimes a cow produces more milk than at other times. Sometimes chickens lay more eggs than other times. You naturally plug into that environment and accept that it’s a distribution. Nothing is fixed in nature. Probability is a view of the world that quantum mechanics uses, and it’s more predictive than the Newtonian fixed-state world. If you have 30,000 data points, that distribution will be more accurate than a single measurement, no matter how good that single measurement is.

SE: Let’s bring this back to the chip world. What do you see as the big drivers for future business?

Brewer: The lithography market has been driven forever by the nature of the equipment available, starting with UV and DUV. But if you dig deeper, that lithography has been driven largely by materials. There has been no real change in the equipment. There has been new equipment on the horizon, if you can afford it, but the last 10 years in semiconductor lithography has been executed by changing the material. I see that as a hint of things to come. That doesn’t mean equipment is not going to progress, but lithography has only been executed by the multi-dimensional and reflective layers that we create. The sensors we discussed are a material-driven technology.

SE: Do you get a better view of where things are going looking at it from the system-level down, or from the standpoint of what the materials can do?

Brewer: It’s both, but there is a big gap there. It’s a combination of, ‘What can I do with this composite of molecules,’ to, ‘What does that have to do with fulfilling the outcome or need of larger properties or functionality?’ We’re still a long way from that bridging those two. We have elected to use the evolution of materials to fulfill the need, and we’re trying hard to make that connection. We’re trying to promote the computer modeling of materials and surfaces. That’s a big step, and an important one, to bringing those worlds together, and we’re trying to master that. We are basically utilizing the integration of materials outward. With sensors, we started with the belief that if we had a better nanotube, that would empower people to create new functionality. It didn’t happen. People didn’t know what to do with that technology, so we made sensors out of it. Then we thought, now they can create better functionality. That didn’t happen, either. So we started adding in electronics and software, and by the time we were done we had provided better functionality.

SE: One of the big changes now is we’re moving beyond the box. With 5G and some of the other technologies, computing becomes additive and distributed rather than defined by a single device.

Brewer: It’s starting to happen again. We lived through the last era and saw the evolution to the phone. Now we’re getting a hint that we’re going through another big shift. This is going to be a long period of change. In a nutshell, this is an accumulation of AI, advanced memories or new memories, and a new dimension in sensors.

SE: And probably new architectures to take advantage of this, right?

Brewer: Yes, there will be new software and new architectures. Those new architectures will include the chip, the package and the output. All of this will be distributed more than in the past. So you have a distributed architecture, but everything else is getting distributed, too. Everything will be distributed—the software, the memory. This will be a distributed digital world, and eventually it will go analog. But right now we’ll go from digital to distributed digital. That will help with AI, because the difference between 3 points and 3,000 points is distributed digital. We’ll start to see distributions that will start to approximate analog. The big drivers are AI, new memory and sensors, and distributed architectures. But all of them are in the process of distributing information, and to do that you need stronger analytics. You can’t look up a single piece of information. That won’t have any value. We need analysis of everything, and that will make us think on a broader level.

SE: And that’s really what we’re moving to, right, a distributed data set.

Brewer: Yes, but the outcome won’t be a distributed data set. It will be the analysis of that data. We keep hearing that whoever owns the data owns the economy. That’s not quite true. Whoever owns the analysis owns the economy. That’s why the consumer gadget is so powerful. It’s the analysis of the information that makes it so powerful.

SE: How will that affect the chip industry?

Brewer: You have to ask, ‘Why don’t we have pharmaceutical companies making bio chips?’ Chips followed straight lines in the past, and those lines led to three or four companies and lack of diversity and opportunity. If you look at who’s doing AI, it’s a whole bunch of companies. The big chipmakers wouldn’t have done that. All kinds of memories have been around for years, but until recently we still built them the same way. Now there are a whole bunch of different kinds of memories coming out. If you look at materials, they used to be for a single use. Now it’s gone from one layer to two layers to three layers. Materials will take on more properties that were achieved by equipment or by design. That will happen inside the materials, taking on new functionality and capabilities. That will be compatible with the spreading out of data and other technologies. Entropy and diversity will continue to increase.

SE: What are the big changes you’re seeing in your business?

Brewer: To answer that question, you have to put it in perspective. We’re a commercial business, and it’s important we’re able to run a solid financial company. But we’re also a private business, which allows us to choose to have other priorities besides financial. We don’t have to worry about the stockholder vision of what a good financial ship is.

SE: Was that a conscious effort to stay private?

Brewer: Yes, because as a public company you have to adhere to the vision of the owners, which are the stockholders. They have a very specific view of what a good financial business is. As a private company, you can have a different view. It’s not totally independent, because you still have to meet the interests of the customers and the banks and the suppliers and employees. But it’s a lot more independent than a public company.

SE: You still need to make the business work, but not necessarily on someone else’s time schedule, right?

Brewer: Yes, or with someone else’s priorities. We can have other priorities than just the financial priority. It’s based upon value. It’s looking at our people, our customers and the community, and then providing a return value of success that is not always financial.

SE: To make that work you have to sell something valuable, though.

Brewer: Yes, and what we sell is innovation. It’s the value we can add, and we add it at the point of high need, which is where there is a transition or change or challenge.

SE: Define innovation. In semiconductors, it’s typically a problem that gets solved and it’s a matter of who solves it the best.

Brewer: There are two kinds of innovation. One is based on short-term need, where you solve a problem. And then there is innovation that helps create a better path for the future. You need both. And you have to create something that doesn’t exist at the present time.