INDUSTRY NEWS

Special Report

Opportunities abound, but funding is tight for start-ups

THE BIG PICTURE: Veteran venture capitalists say semiconductor tool and other supplier start-ups have to work hard to secure financing. ILLUSTRATION BY ROB COLVIN

For potential semiconductor equipment, materials, and software start-ups, linewidths aren't the only things that are shrinking. The transition to nanotechnology and the introduction of new materials present fat opportunities for fledgling companies with the right solutions for yield-conscious customers. But entrepreneurs looking for funds to launch their companies will find that the pool of interested investors is shrinking as well.

Richard Whiting is the managing partner of DynaFund Ventures, a high-tech venture capital firm based in Torrance, CA. A partner in the firm with Lam Research founder David Lam, Whiting says start-ups have to search more for seed money than in previous years. "It is true that the universe of venture capital firms that are interested in funding semiconductor companies—especially semiconductor equipment companies—is a smaller universe than the universe of firms that are willing to fund some other areas; software, for example. Because of that, it's generally true that semiconductor equipment companies have to work harder to find financing, particularly in the first year."

Lam and other veterans with a wealth of experience in the equipment industry point to a number of reasons for the funding scramble. Investors see a mature industry with high barriers to entry, huge start-up costs, and a 10- to 15-year wait for a potential return. Globalization also plays a role, as start-ups in the United States face financial competition from potential new firms in growing regions such as Asia. In addition, the decline of government research funding and contractions at industrial R&D organizations such as Bell Labs and RCA Laboratories have left more research to universities, which have less expertise in commercializing generally usable ideas.

Rick Hazard, president and CEO of four-year-old Lightwind, headquartered in San Francisco, says global competition has upped the ante for companies just starting out. "You take a look at the globalization of IC manufacturing and, in order to be effective, you have to be able to deal with global issues. I don't think that was always the case. Back in the early 1980s, I think you could be successful regionally and be very successful as a company. The willingness of a fab to adopt a new technology and take risks is dramatically different now than it was in that period. Therefore, if you're introducing a new technology, some of the barriers to entry are quite significant compared with what they were back in the early 1980s."

A mature industry becomes more hesitant about accepting new technology, Hazard believes. He emphasizes that beta testing is critical for a product entering the marketplace, "because that's where you quickly separate fact from fiction." Lightwind recently completed beta tests at IDM sites of its L3 tool, an in-line process sensor that sits on the exhaust port of a plasma strip, etch, or deposition system to monitor the composition of wafer process chemical by-products. The company is preparing for a major ramp-up in 2005–2006.

Beta testing can dispel doubts, Hazard notes. "One of the things that is very difficult as the industry becomes more mature and there's more and more consolidation is that very often technologies are viewed as disruptive, or a new approach very often has a difficult time being nurtured within an existing semiconductor equipment manufacturing organization because you have to look at the impact on the existing product line and revenue stream."

Nevertheless, launches of companies such as Lightwind, Quantiscript Nanotechnologies, Metara, and Ponté Systems over the past few years prove that equipment and software start-ups can flourish in today's more barren terrain. They just need to know how best to tap sources that will help them grow. Furthermore, entrepreneurs must realize that the IPO's heyday has passed and that their future success will depend on partnerships and mergers with established, deep-pocketed suppliers.

The keys to drawing money from that shallow pool? Technology with strong IP protection certainly helps investors sleep better at night. But there are three overarching factors for convincing skeptical investors, according to Claude Brisson, president and CEO of Quantiscript: having a breakthrough technology, a strong management team, and a good business model.

Born from research conducted at the University of Sherbrooke in Quebec, Canada, the company specializes in non-spincoated resists and associated process technologies. Quantiscript sells two products: QSR-5, an evaporated E-beam negative resist; and 1st Impression, a nanoimprint template-based tool for process characterization.

Brisson says potential investors look to see whether all three legs of his three-legged-stool launch model are in place. The current environment in the semiconductor industry, particularly following the worst years of the downturn in 2002 and 2003, "has had a tremendous impact on the ability of start-ups to get financing. People scrutinize a lot more and are a lot more risk averse and prudent."

These concerns make it all the more important to investors that a newly hatched supplier has the technology, the model, and the team poised and ready. Depending on their appetite for risk, investors may prioritize these components differently, Brisson notes. "It's really difficult to say that one is more important than the other. It's important that you know how you're going to get all three in place very quickly."

Naturally, accomplishing this structural trifecta before "going through your first round of financing. . .makes it very difficult for a start-up to begin." The industry is capital intensive and "extremely detailed. The industry spends a lot of time understanding the technology, characterizing the behavior of the process and the technology, because it's under such pressure to deliver yield and throughput. It studies all aspects of the infrastructure of the fabrication process to an incredible degree, which is understandable."

Brisson insists that "the projects that have staying power are the ones that are being pursued because the investment dollars are out there. What it does mean is that they're just not as easy to find as they used to be." The pendulum, he insists, is swinging back from the position of  "avoiding risk at all costs. . .to a more balanced position of risk versus reward."

Not everyone agrees that investors have turned their backs on the semiconductor space. An industry veteran, Jerry Cutini has seen his share of start-up successes and failures. He cofounded the post-CMP cleaning tool firm OnTrak Systems and has been an executive with Novellus. In 2002, Eshinui, a start-up selling software for tracking spare- parts inventories, hired Cutini as president and CEO, but the fledgling company eventually folded during the industry's downturn.

Cutini is now president and CEO of Aviza Technology, which acquired ASML's thermal division nearly two years ago, and in March 2005 announced a merger agreement with Trikon Technologies, a Wales-based supplier of plasma etch and deposition tools. The executive believes two issues dominate the matter of new launches.

"First, and this is a general statement, getting funding isn't too difficult, as there are both individuals and VC firms that are always looking at this space. I can think of a number of new companies that have started in the past few years. It's certainly down from the early 1990s, but it shows there is funding available. Second, the real issue has changed from securing funding to getting customer interest."

The diverging fates of OnTrak and Eshinui are instructive, Cutini says. In addition to a can't-miss CMP-related technology, the former company had real customer pull. The latter, while drawing interest from "various parts of the customer base," could not connect with "a driving need."

People in different fabs expressed keen interest in Eshinui's software, "but getting them to 'pull' or buy the product was not an easy proposition. The point being, if your new product doesn't really solve a problem, you're going to have trouble. Saying you have a better technology isn't enough. You have to be in the unique position to address a problem in the customer base and provide a solution that no one else can."

"The commercialization is very easy [when] the market is pulling you instead of you pushing the market," agrees Nitin Deo, senior vp of marketing for Ponté Solutions (formerly E-Z-CAD). As a new company, the Mountain View, CA–based firm had a very smooth transition to the marketplace precisely because it was "addressing the right problem at the right time."

A fundamental lesson is that the successful start-up exists to serve the customer's problems, he insists. "There is a problem first, and we have a very innovative and robust software solution. If you have a solution looking for a problem, it's really a hard sell."

With its technological roots in embedded-memory design, Ponté understood that engineers needed an earlier picture of potential yield-impacting design problems than they were seeing. As chipmaking processes dip below the 90-nm level, "designers are finding that they need to know what's in their silicon, what's in their design, what is the design content of their chips. And technology like ours, which gives them a very clear view of what will happen when you actually manufacture the chip, is very, very timely." Validation for Ponté's approach came from NEC Electronics, which announced in early June that it will use the supplier's model-based yield-analysis software with the chipmaker's own methodologies for analyzing random defects in its fabs.

Deo acknowledges that because it's a software firm, the company enticed investors more readily than a hardware supplier may have. The amount of resources needed to design software "is fairly low compared with any hardware company. That's the traditional difference between software and hardware. Software can be changed and customized fairly easily."

Alain Harrus, technical advisor and venture capital expert with Metara, says it's not enough that a start-up have the requisite breakthrough technology. "The breakthrough has to be associated with a pain point for the customer, especially in the semiconductor industry. . .You have to be able to address and pinpoint where the customer is then willing to work. The most difficult thing is to establish credibility."

Slightly more than five years old, Metara began as "a bunch of parts in Ziploc bags," Harrus jokes. The founders of the in-line chemical metrology firm determined that "the chemistry and the chemical baths in the process are kind of the last area that is not being utilized to improve yields and improve cost saving through scrap [wafers]," says Randy Clegg, the company's vp of sales and marketing. Metara launched its new Sentry platform in mid-June. The metrology tool is designed for real-time, noninvasive characterization of process chemistries.

"Everybody was worried about defects on the wafer, but they would send their chemistries out to a lab with a turnaround time from four hours to as much as two weeks. A lot of wafers can be jeopardized during that time," Clegg points out. Clients processing huge wafer lots would discover that by the time they received the lab results on the health of their chemistries, "they were on to a new batch of chemistries."

Asked to define success for a new company, Harrus replies rapidly: "Profitability, absolutely." Shepherding the technology from the drawing board to the boardroom means one thing. "I don't like sports metaphors, but to me it's the salespeople who carry the ball the last five yards.  It's your touchdown teamÉ. Very seldom do the best technology and people beat a path to your door. It does not happen."

Sooner rather than later, you're going to need a good sales team, insists Brisson of Quantiscript. "Start-ups are initially driven by the technology, but at a point in time you've really got to reverse that completely, and the company has to be driven from a sales and marketing and product perspective. The sooner you do that, the better."

Cutini says that the definition of success has changed. When OnTrak began back in the early 1990s, success was typically delineated by a much-anticipated initial public offering. "Today," he says, "I'd argue that there is little to no chance of a start-up front-end equipment company to get to the IPO stage on its own." The goal now is "to get purchased by one of the larger equipment companies."

David Lam, who founded the plasma etch equipment company bearing his surname in 1980, says an IPO for a fledging equipment supplier "is like graduating from college or getting married. Unfortunately, when this is not available, we all turn to mergers and acquisitions as a way to continue growing, albeit under someone else's name."

The "mentor capitalist" heads the David Lam Group, a high-tech consulting firm. He was recently named chairman of the board of Qcept Technologies. The new metrology tool supplier is marketing Chemetriq, an in-line inspection system that measures chemical variation and uniformity on wafers to atomic-level sensitivity. The tool is based on technology transferred from the Georgia Institute of Technology in Atlanta, and Lam is quite excited about the company's future.

Lam says any economic impact of the industry's shift into nanotechnology "will be years away." Cutini says the opportunities in the nanosphere "are getting harder to find. There are obvious materials changes that are nontrivial. However, it's rare to find one solutions provider that can address all the varying needs of customers' advanced manufacturing challenges."

Ultimately, Cutini doesn't see any major process modifications "until you get below 32 nm, then things will change pretty significantly again. So we might be in a period of hibernation for start-ups. Anyone that has solutions for the sub-32-nm realm won't have many customers that will buy products [now], so there's a few years to wait."—JC