ISS 2008 Focuses on the Economy, Lack of Test Standards, and “Profitless Prosperity”
Key economists, technologists and financial analysts converge at the SEMI Industry Strategy Symposium, where apt discussions and keen insights take place about the economy, emerging technologies and business trends. ISS is the only forum where senior executive suppliers and device manufacturers go to get a first-hand view of what the industry will look like in the year ahead, and gain strategic insights into the shifting business environment.
The following article is a day-by-day summary of the key points of the presentations at ISS 2008.
Day One: ISS 2008
High Tech Industry Insulated from U.S. Economic Woes
IC Unit Volumes Continue Double-Digit Growth Trend
Excess inventory is the major concern facing the U.S. economy. That’s excess housing inventory, not semiconductors. If the U.S. economy does slide into recession in 2008, it will be due to the housing market slowdown and high oil prices–while the technology sector will be relatively unscathed, according to speakers at SEMI ISS 2008 on January 14 at the Ritz Carlton Hotel in Half Moon Bay, California.
“Housing is at the epicenter of the crisis,” said Nariman Behravesh, chief economist with Global Insight. “High tech is relatively insulated. There are no excesses, nothing to be worked off [in high tech],” he said. Although “recession” was mentioned frequently during the economic sessions on the first day of ISS, most speakers did not see it as a certainty for the U.S. economy. “Don’t get hung up on the R-word,” said Robert Fry, senior associate economist, DuPont. “If we do fall into recession it will be mild.”
Behravesh noted that stronger growth in exports–helped by the weaker U.S. currency–will cushion the blow of a U.S. downturn. Furthermore, he said the low inflation and low interest rate environments would be favorable to capital investment. DuPont’s Fry added: “This is the best climate in a decade for U.S. exports. I hope you are taking advantage of that.”
Shawn DuBravac, economist for the Consumer Electronics Association (CEA), pointed out that if the housing market is excluded, U.S. GDP growth in 2008 is forecasted at three percent. “A lot of credit issues are contained within the financial community. Businesses do not seem to have any problem getting credit,” he said.
ISS 2008 also presented an opportunity for semiconductor industry analysts to present their forecasts for 2008. According to Semico Research, the 2008 semiconductor device market will grow 12 percent in term of revenues and 15 percent in terms of units. Jim Feldhan, president of Semico, pointed out that component inventories have come down significantly and predicted that utilization rates will remain above 90 percent throughout the year, resulting in stabilization of average selling prices (ASPs). He added that consumers are still enamored by technology and will continue to spend disposable income on new electronics products.
Bill McLean, president of IC Insights, highlighted the continuing strong unit growth in the IC industry. “The demand for ICs, as far as unit growth is concerned, has never been better,” he said. In fact, 2008 will mark the seventh consecutive year of double-digit unit growth in chips. Growth in the past five years has averaged 14 percent per annum.
McLean is also optimistic that ASPs will stabilize, and even start to rise. A strong upward force on ASPs will be the trend for more IDMs to become “fab lite”, and the corresponding trend for foundries to reduce capex spending. “We’ll see better pricing in the IC industry because of this collision course”, he said.
VLSI Research forecasts 7.5 percent growth in the IC market for 2008, based on 6.8 percent growth in the electronics market. The semiconductor equipment industry will decline about 5 percent in 2008, after experiencing 7 percent growth last year, according to VLSI. The bright spot in equipment in 2008 will be the FPD market, according to Aida Jebens, senior economist for VLSI Research. “[Flat] panel makers underinvested in the past two years,” she said. As a consequence, display equipment sales will grow 19 percent in 2008 vs. an 18 percent decline in 2007, according to VLSI.
Gartner Dataquest forecasts 6.2 percent growth in semiconductors this year (including solar), while the capital equipment market will decline by 10 percent. “We anticipate return to growth in 2009,” said Dean Freeman, research vice president at the market research company.
The outlook for the semiconductor materials market was presented by Dan Tracy, senior director, industry research and statistics, SEMI. The worldwide wafer fab materials market is forecasted to grow 12 percent this year to reach $28.4 billion. The largest single segment in fab materials is silicon wafers, worth $14.4 billion this year, according to SEMI data. The semiconductor packaging materials market is forecasted to grow 9 percent this year to $18.4 billion. The largest single segment in packaging materials is organic substrates, a market expected to be worth just over $7 billion this year.
Day Two: ISS 2008
ISS Speakers Discuss Challenges and Opportunities Facing Industry
Lack of Test Standards Slows Technology Adoption
Speakers on the second day of SEMI ISS 2008 addressed business and technical trends impacting the semiconductor industry, including automation challenges in 300mm fabs, merger and acquisition strategy, and lack of standards in test equipment.
Steve Schwartz, chairman and CEO of Asyst Technologies, pointed out that automation spending as a percentage of wafer fab equipment has grown from less than 1 percent in 1990 to between 4 and 5 percent in 2007, which is comparable to what is now spent on ion implantation, CMP and PVD.
Schwartz argued that the current automation market structure inhibits innovation because there is no synergy between the tool interface, designed by the equipment company, and the automation handling system which is designed and controlled by the fab. What’s needed is a unified automation architecture where the process tool front end and AHMS system work together, according to Schwartz.
The problems with automation have been exacerbated by the shift from 200mm to 300mm fabs. In the mid-1990s a typical 200mm fab producing 30,000 wafer starts per month required 400 process tools with 1,000 load ports and 5,000 stocker shelves. Up to 200 operators per shift were required to manually move wafers around.
Nowadays, a typical 300mm fab producing 150,000 wafer starts per month requires 1,000 process tools stacked over three floors, with 2,500 load ports and 25,000 stocker shelves. “There is no way we are going to operate these [types of fabs] with people,” said Schwartz. The impact of today’s high throughput process tools has resulted in what Schwartz called tool starvation. “$150 to 200 million of spent capital is sitting idle waiting for automation systems to feed it,” he said.
Leo Berlinghieri, president and CEO of MKS Instruments, provided advice on conducting mergers and acquisitions, based in his company’s experience in doing 15 such deals. He said in most cases the senior management of the company being acquired does not stay around for the long term. “The reality is that it doesn’t work for them or it doesn’t work for you,” he said.
Acquiring multiple new locations is often the end-result of an acquisition, but the buyer needs to be careful not to shift or close these down, according to Berlinghieri. While it’s advisable to streamline volume manufacturing into a few locations, “you will lose the engineering assets if you move locations,” he warned. Different company cultures are another pitfall. “If you have cultures that conflict it is really difficult to get anything done post-acquisition,” he said.
Klaus Luther, senior director of corporate test technology for Infineon Technologies addressed the issue of rising test cost. He pointed out that the average test company spends more money on selling tools to customers than it does developing them. “I consider this insane,” he said. For example, R&D costs for an average ATE vendor are 19.6 percent of sales, while SG&A costs are 23.7 percent, according to data cited by Luther.
Luther, who is also co-vice chairman of the Semiconductor Test Consortium (STC), explained that the goal of the STC is to promote open standards for test equipment as a means to unlock cost saving potentials in the industry.
ATE vendors argue that standardization works against their ability to innovate or differentiate their product, according to Luther. “I think the opposite is true. The absence of standards adds to customer irritation and slows down technology adoption,” he said.
Greg Linden, a research associate with the Center for Work, Technology and Society at UC Berkeley, provided data on the global labor market for engineers and noted that there was not a huge flow of jobs to Asia. Rather, it was a slow shift and the offshore jobs tended to support high tech activities taking place in the U.S.
Whether or not there was a shortage of qualified engineers in the U.S. is the wrong question to ask, according to Linden. The right question is: “Are we training enough for the long term?” The so-called “Google effect” – where young people see the Internet as an opportunity to get rich quick -- was contributing to the decline in the numbers of engineering students in the U.S.
Linden provided data to show that non-U.S. citizens receive three times as many U.S.-issued PhDs in electrical engineering than U.S. citizens. Of the foreign nationals who earn PhDs in the U.S., 35 percent come from China, with most of the remainder from Taiwan, India and Korea.
Many of these newly minted PhDs end up going back to their country of origin to find work because they are unable to get a “green card” to work in the U.S. before their H1B work visas expire. “The process of transforming H1B visas into green cards is broken,” said Linden.
Day Three: ISS 2008
Profitless Prosperity Plagues Semiconductor Device Industry
Cost Reduction is Not the Solution: Lam CEO
The semiconductor industry has slipped into a state of “profitless prosperity,” according to a speaker on the third day of SEMI ISS 2008. “The trend since 1995 has been down and to the right. You can’t be prosperous if price declines faster than cost,” said Steve Newberry, president and CEO of Lam Research.
In 2004, the average operating profit among 40 semiconductor companies was 23 percent of revenue. By 2007, average operating profit had fallen to 12 percent, with 15 of those experiencing operating losses. Excluding the 17 most profitable players – which include Intel, Texas Instruments, TSMC, and the analog and fabless companies – the average operating profit is minus 1 percent. “If you can’t make money in a 50 percent unit growth market, you are in trouble,” Newberry said.
No single IDM has enough volume to effectively compete with the economies of scale of the foundries, according to the Lam CEO. In a telling sign, Texas Instruments, the most profitable of the large IDMs, is going from “fab-light” to “fab-lighter,” and is even beginning to look more like a fabless company, he explained.
In the foundry sector, TSMC takes all the profit while the other players lose money. Yet TSMC is not ahead due to cost leadership because its costs are similar to its competitors. Rather, it is ahead because its wafer prices are higher. “Somehow TSMC is getting a price premium,” said Newberry. “Although manufacturing cost is important, it is not the root cause impacting the… profitless prosperity problem.”
Instead, other factors are contributing to the destruction of profits, including imbalance in supply/demand, a continuous push for higher market share in commodity markets, too many players in each market, and business models that are too costly. In general, cost reduction works but it can’t overcome a supply/demand imbalance or the use of more effective business models, Newberry explained.
He sees strong parallels with the aluminum industry, which used to face similar problems to the semiconductor industry until it took steps to improve profitability. Newberry urged the chip makers to look at other industry sectors and learn from their mistakes. “If we continue to behave as other industries before us have behaved, we will suffer the same fate,” he said.
What does this profitless prosperity mean for the equipment suppliers? Although device makers constantly put downward pressure on equipment prices, the problem will not be solved through cost reduction. Even if equipment suppliers gave away all their profits to customers, the profitless semiconductor players would still lose money, according to Newberry. “We are not going to help our customers by giving away our profits,” he said.
Another ISS presenter, Jay Deahna, an analyst with JP Morgan, said the desire to perpetuate profitless prosperity seems to be in the DNA of the semiconductor industry. If device makers continue their price wars, he argued, then the best place to invest would be the “arms dealers” (i.e., the tool suppliers). “I don’t think the equipment industry is the disastrous place to be,” said Deahna.