Curt Wozniak, Electroglas Chairman and CEO

Curt Wozniak joined Electroglas (San Jose) as CEO in 1996 and became chairman in 1997. Under him, the company acquired Knights Technology, Techné Systems and Statware, expanding its capabilities to include data analysis and post-fab inspection. Before joining Electroglas, Wozniak was president and COO of Xilinx Inc. Prior to that, he spent more than 10 years at Sun Microsystems, where he served as vice president of worldwide marketing for Sun's subsidiary, Sun Microsystems Computer Corp. (SMCC). Wozniak also spent several years as a production engineering manager at Hewlett-Packard. He has a B.S. in mechanical engineering from GMI Engineering and Manufacturing Institute (now Kettering University) and an M.B.A. from Stanford University. Electroglas manufactures an integrated toolset: wafer probers, inspection systems and Web-based process management software.

SI: Like everyone else in the industry, you announced lower earnings. Revenue was $11.7M for the third quarter, up 24% from the third quarter of 2001, but down 43% from the second quarter. However, you've said you expect 2003 to be a growth year. What do you base this on?

Wozniak: Obviously, the current environment is poor. If we look at what's gone on since our last quarter of last year, we've consistently had small increases in booking levels from that low point. Most of our customers are running at fairly good levels of capacity utilization — around 70-80%. That's not great, but at least they aren't putting systems into mothballs as they were a year ago. So if economic growth continues through this year and the next, and the U.S. and world economies continue to grow, eventually that capacity will be absorbed, and they'll get to the point at which they must add new capacity. Most analysts, myself included, are guessing that this will take place at the second half of 2003. But we're still in for another nine to 12 months of rough sledding.

SI: So you don't see those bookings vanishing like they did before when, instead of reaching the upturn, everything just stopped and people began canceling orders?

Curt Wozniak (Source: Electroglas)

SI: What else besides Electroglas' pocketbook has been affected by the downturn?

Wozniak: The downturn is certainly affecting our earnings and eating into our cash reserves, as we've lost money over the last few quarters. However, our strategy is to continue to invest in R&D and new products, because upturns are unpredictable. We've refined our strategies to try and crimp down any new expansion plans that we had, and are very focused on making sure we're doing what we have to do and spending money in the right places.

SI: Last August, you announced a layoff of 130 workers in the United States and added 45 to your Singapore operation. More layoffs are planned as Singapore comes on-line and you're able to shift your manufacturing there. How is this working?

Wozniak: As a company, we've been in Silicon Valley — the San Jose, Santa Clara area — for some 40 years. Unfortunately, this isn't the most inexpensive area in which to do business. This moved us to make a strategic decision and, at the beginning of this year, we announced that we were going to move our manufacturing of wafer probers to Singapore, to drive down production costs as well as get a better presence in an area of the world where there is considerable customer growth. We considered several locations, and determined that Singapore was probably the easiest move for Electroglas to make, given our size. We started this program in January of this year, and we were up producing machines by July.

So far, the team there and the results have been excellent. Because we chose to do this during a downturn, we were able to get very good people and build an excellent team. As a result, we've been able to move and reduce our operations in San Jose. We are currently in the process of moving our newer products to Singapore — this should happen by year's end or the beginning of 2003. At that point, we think we'll be well prepared for an upturn, should it come. We'll have the facilities and people in place all ready and trained, and be in the position to drive costs down as we compete in our marketplace.

SI: Should Washington do more in the way of taxes and other incentives to keep more of our industry in the United States?

Wozniak: It would help to do some things. But it's not just taxes that drive those decisions — it's also the cost of the infrastructure. In some cases it's just the economics of the comparative advantages of moving things to locations where the costs are cheaper. The places you leave then have to move up in the food chain in terms of their value. That's why we'll continue doing a lot of design, engineering, R&D and marketing in Silicon Valley, where so many talented individuals live. But many of the manufacturing activities will have to move elsewhere, where costs are lower.

SI: At what point do you think you will be doing all of your manufacturing offshore?

Wozniak: We probably will be doing all of our manufacturing overseas by the beginning of 2003, except for new product development — the building of those first few machines.

SI: So when the big upturn arrives, everything will come from Singapore?

Wozniak: Exactly.

SI: Are you looking at China as well?

Wozniak: We are. First, we're being very aggressive at trying to get business in China, which is a very competitive, cost-conscious and dynamic region. I've already been there a few times this year, and probably will continue to go there to visit customers and see in what state they find themselves. We don't see manufacturing in the short term in China, except at the subassembly level. As we move to Singapore we will probably outsource some of that component assembly to places like Malaysia, Thailand or China, to continue to drive down the costs of our machines.

SI: How do you view the intellectual property question in China?

Wozniak: I think we must be very careful to protect our IP, whether we go to China or anywhere in the world. Until China really develops its legal structure regarding IP protection, you'll have to be very careful to keep control over your IP, to be very strategic about which IP you release or put over there. We've had the experience of having some of our machines cloned in China and Russia several years ago, although they were unable to duplicate our quality and sophistication level.

SI: Have you implemented additional spending controls and cost-cutting actions?

Wozniak: Yes, many spending controls and cost cuts! In addition to having to reduce people, which is the worst part of being in this industry, we've also had every employee above a certain minimum salary take a pay cut this year, staggered up to the CEO, who took the biggest one. In exchange, all employees got stock options. We've also had additional shutdowns during each quarter to reduce costs, and of course we're doing things such as clamping down on discretionary expenses.

SI: You've been very active in the merger area. You acquired Knights Technology in 1997, and a number of software vendors. Did this require company restructuring?

Wozniak: We've gone through big changes. When we started off as a public company in 1993, we were strictly a wafer prober company and, given the size of that market, one of the decisions that we made was that we had to diversify; we needed to add product lines. However, I've always been a believer that products must have a very strong strategic tie to your core business. So one of the things that we looked at was where wafer probing was headed and where equipment in general was going in manufacturing. In doing this, we determined that one of the core advantages semiconductor equipment companies were going to need to develop over time was a software capability. This is why we have acquired two software companies, and put out a lot of effort to develop software applications and a software infrastructure that allows you to manage data throughout the wafer probe floor but also throughout the fab.

We also looked at software as a way to do more hardware products that tie into the data infrastructure. So the first move we made was to develop an optical inspection capability for what we call the post-fab area — the back-end manufacturing area — to provide a data source. That strategy of software infrastructure and hardware that ties in with the data flow of the factory gives us a platform to do further acquisitions and development. We've gone from a hardware-centric company to being a hardware/software systems company culture.

SI: Your software operation is already distributed?

Wozniak: Yes. We have software in San Jose and Corvallis, Ore. And then we have about 40% of our software development in Bombay, India. This has proven to be a very strategic and critical resource for us, helping to get the people we need to build the products our customers need.

SI: What are your short- and long-term plans and strategies?

Wozniak: Short term, it's survival. We don't know how long this downturn is going to last or how deep it'll be. We must be sure we keep an eye on the cash flow, while ensuring we're ready for the upturn when it does come. Our longer-term strategy is to come out of the downturn stronger than we went into it from a product and capabilities standpoint, as well as with better customer penetration and market share. Over the last five years, we've tripled our R&D effort and are spending almost $30M a year. We'll also be looking to further acquisitions for value-added and new product purposes, which will largely be software-centric and strategically consistent with our vision.

SI: Is growth getting difficult?

Wozniak: Growth, particularly in our industry, is an interesting thing. The industry grows at an average of 15 to 17% per year and, for our sector, that's ±60%, so we'll have years where we grow 80% and others where we shrink 60%. Of course, growth is an average problem, much like the story of the guy who drowned in a stream that is 3 feet deep on average, because he found that one spot that was 8 feet deep.

In our industry, growth occurs in spurts. So we must take advantage of the upturn periods that come every couple of years, and new market segments that appear as a result. However, if you look at the development of new market segments and technologies, and that could be CMP, SMIF or copper, all the new technology development in our industry has long gestation periods — developments can take 10 years before moving into manufacturing. In the case of the back end, bumps are a good example. IBM has been making bumps for the last 20 years.

It's contrary to what people think of our technology — as being fast-paced. But, in reality, the transitions to manufacturing are very slow. It's true for software, packaging, for moving optical inspection to the assembly and test areas — it all takes a long time. Meanwhile, you're going through these vicious cycles, which add even more complexity to the matter. In our industry, growth is slow but also violent.

SI: What changes are your customers asking for?

Wozniak: There are changes on a number of fronts. On the technology front, the pressure of Moore's Law in the front end is driving changes in the back end. One example is that more sophisticated technologies in the front end are driving changes in test, whether higher frequencies or more diverse temperature test capabilities.

The second trend is cost. With the move to Asia, the drive into China, the goal is to expand the marketplace to the 1.2 billion people there, and cost is becoming more of a factor. It's always been so, but now it's more so. Finally, customers want to maintain overall productivity. They're recognizing that they must increase the productivity of equipment, the net processing time, and we think they'll realize that the only way to do this is through sophisticated networks and software.

SI: The OEM market isn't exactly a high-volume one. Your customers demand increasingly more sophisticated platforms, capable of doing more with less human intervention, but they also want lower costs. How do you reconcile the two?

Wozniak: This is a very significant problem for the industry because it puts a constraint on innovation. Twenty years ago, there was greater R&D partnership on the part of equipment and semiconductor companies. If you wanted to develop a new piece of equipment, you'd build a couple and put them with customers and work with them to develop it further, and then market it. The environment has now shifted, and customers require that equipment come into the factory plug-and-play.

There's nothing wrong with this, except that we must recognize that there is a learning curve involved, and that to get a plug-and-play piece of equipment you must build a number of them. The cost of that development — from prototype to plug-and-play — is significant. This puts constraints on innovation in our industry and stifles new companies due to the $50M investment it takes to bring each new product out. The stakes get bigger as we go through each of these transitions, and I am unsure of what the solution might be.

SI: Another aspect is that companies now want to concentrate on their core competencies, which means they've reduced their in-house expertise, and expect the OEM to fill the gap. How do you view this?

Wozniak: In addition to getting the machine, we're finding that customers expect the equipment company to dedicate more application development resources on site. There have been cases where we have had people on site for well over a year, working with the customer to develop application-level specific capabilities that they no longer wish to develop internally.

SI: When you devote that amount of resources for such a long time, are you working at a loss?

Wozniak: Yes, because customers don't want to pay for it. You're doing it in the hope of developing a relationship with the customer and locking yourself in with them for a long time.

SI: Are you moving toward owning your processes?

Wozniak: I believe there's an industrywide progression toward this. I see it in our area, particularly in data networks. The core process of probing is relatively straightforward, and customers know how to do it. However, when it comes to really integrating 150 wafer probers on the floor, networking them all together, extracting data from them, and building applications to automatically manage them either locally or remotely, that requires a level of sophistication that our customers lack today.

That's where we'll begin to own the process. It's really about providing them with a comprehensive data networking solution — process management and control. It's about taking the data and analyzing it, not only for the process but for the equipment as well, to get the best overall productivity.

SI: Any advice for the industry?

Wozniak: Regardless of upturns and downturns, don't give up on your R&D spending. It's the only thing that ensures your future and the industry's. Always make sure you have enough cash to survive the downturns.