By Bob Haavind, Phil LoPiccolo, SST editors
Part 1: Chip industry set for resurgence, not maturity, predict ISS speakers
A recurrent debate during the 2006 Industry Strategy Symposium (ISS) at Half Moon Bay near San Francisco, CA, was whether the semiconductor industry really is mature, or if it is still “maturing.” The consensus was that not only is the industry not yet mature, it is set for a resurgence driven by a host of new electronic devices for consumers and businesses offering wireless connectivity. As a result, compound growth rates will climb into the low double digits over the next few years, some forecasters concluded. However, they also agreed that growth will be more gradual with more moderate cycles than in the past, due to better forecasting and caution among chipmakers.
The debate over “maturity” was sparked by Duncan Meldrum, chief economist for Air Products & Chemicals, who compared the semiconductor industry life cycle to other industries such as automobiles and tires. Although he found some signs of maturing, he concluded that there is likely to be significant growth left before the industry reaches fully “mature” status.
One bright spot is that there are likely to be huge growth markets for trailing technology in developing countries that can buy more affordable computers, cell phones, and other information technology products. Meldrum also sees very small penetration in Asia for even higher-end products with increasing semiconductor content. While the rise of foundries might be part of a maturation shakeout, he suggested, they also may prolong the growth phase by allowing faster technology changes, which were very slow in some other industries, such as steel.
The industry may need to curb its enthusiasm in the near term, however, suggested Richard Iley, senior economist, BNP Paribas, because the US economy is perilously imbalanced. Growth has become overly reliant on consumer spending and a housing boom, together constituting 77% of the economy, versus 65%-70% in European countries, which are more reliant on exports. Construction spending, the most volatile and interest-sensitive sector of the economy, hit 6.1% of GDP in 3Q05, compared with an average of 4.75% since the end of WWII, according to Iley. “Real” prices of houses are up 60% since 1997, while they rose only 14% between 1975 and 1997. Even the oil price surge has not slowed down US consumers, who continue to spend as if the price spike is only temporary.
Meanwhile, global dollar liquidity, which was buoyant in 2003-2004, is now fading rapidly, and this is usually a leading indicator of a slowdown, Iley said. He still does not expect a recession, but sees slower growth ahead. On the bright side, he cited very strong productivity growth (a seven-year average of +3%) and tremendous strength in the corporate sector, with great gains in free cash flow and record cash on balance sheets. This provides a crucial buffer against a potential recession, he added; still, caution is necessary, because with a delicate economic imbalance, risks are high for policy mistakes.
A rousing, video-animated keynote by Brian Halla, chairman and CEO of National Semiconductor, led the upbeat outlook. He showed a host of cool gadgets introduced at the Consumer Electronics Show (CES) a week before, such as a family of iPods from Apple, including one with a 60GB microdrive that can play downloaded videos. His presentation also included live video views of traffic on major nearby highways that will soon be easily accessible from pocket devices or systems built into cars, illustrating the kinds of new services that will drive millions to buy new pocket devices.
“Personal technology has replaced the PC as the driver,” Halla enthused. He felt that the 2006 CES marked a turning point toward truly ubiquitous wireless access. He cited plans for Mountain View, CA, for example, to provide Wi-Fi everywhere in the town, allowing a pocket device to be able to get a download anywhere. “You won’t have to go to McDonald’s or Starbucks,” Halla explained. The industry is working on a mobile wireless mesh network architecture to spread ubiquitous access, he added, making it “the age of the client.” He cited Bob Metcalf’s Law that “the network goes up as the exponent of the number of nodes.” (Metcalf was the inventor of Ethernet, and a cofounder of 3Com.)
By 2009, users will have broadband wireless, Internet 2.0, a terabyte iPod, and machine-to-machine links that could drive infinite demand, according to Halla. Even the near-term outlook is very bright, he added. At the top of the dot-com bubble, the semiconductor industry hit $204 billion, but with a lot of that going into excess inventory, perhaps only $170 billion was real. “Now we’re above that with minimal inventories,” Halla explained. China already has 400 million high-end consumers, he said, and the market for semiconductor content grew 32% there last year, making it the highest consuming market in the world. The fastest growing market for cell phones, he said, is in Nigeria, with only 140 million people. Growth is not just in wireless, either, with much of the once-dark optical fiber now starting to fill up as Internet applications expand, he added.
The US may find itself trailing in taking advantage of emerging market opportunities, however, because of what Halla called “the perfect storm of technology neglect.” He cited several signs of US decay, such as the demise of Bell Labs, a shortening R&D funding horizon at DARPA, declining foreign student enrollment in US universities, and fewer PhDs from overseas being allowed to stay in the US. While emerging nations push technology R&D, tax incentives, and stock options, the US is falling behind in employing new technology (it is no. 13 in deploying broadband, for example).
A turnaround won’t happen, he suggested, until we get our fourth “technology President.” The first, in his view, was Thomas Jefferson, under whom the patent system started. The second was Abraham Lincoln, who received a patent for a method of dynamically pumping air into ships so they could pass over shallow spots. The third was Dwight Eisenhower, who created NASA and funded the National Defense Education Act, providing scholarships for science and technology students.
The first Russian Sputnik, he explained, beat the US into space, and spurred Eisenhower’s efforts to back technology. Perhaps another jolt like that will be needed before the US again gets a president who understands how critical it is to foster innovation and new technology.
Part 2: Photovoltaics market shines at ISS
Sharing the spotlight at the recent Industry Strategy Symposium (ISS) was a discussion of how semiconductor manufacturers can apply their expertise to reduce costs and improve efficiencies in the photovoltaics (PV) industry. More energy will be consumed over the next 60 years than in all of recorded history, and traditional fossil fuels are not going to meet demand, according to featured presenter Alexander Wong, head of venture capital at D.E. Shaw & Co. As a result, by 2040 the solar industry is projected to be bigger than the oil industry, he said. This realization has led to intense VC investment of about $1 billion/year in so-called “CleanTech” – alternative-energy products and services, including PV, that are regarded as friendly, and more important, make money.
“People have laughed about solar technology for many years, saying it’s the subsidies that sustain the market. But now we’re starting to see solar becoming a real market by itself,” Wong said. Over the next five years, revenues will grow from $10 billion to more than $50 billion, he predicted, with new installations growing from 1% to 5% of total sales.
To give a sense of where the opportunities are for semiconductor equipment and materials suppliers, Wong explained that solar cells are basically very large semiconductor devices with much less stringent tolerances and processing demands. “Someone at BP recently told me to tell the equipment guys and the materials guys to just go back 15 years, dust off those old plans, and that’s what is needed for the solar industry.” Indeed, while ICs need to be as small as possible and defect free, and need to sell for $1000/wafer, for PVs, larger is better, they can be defect tolerant, and they can sell for $10/wafer, he explained. “Optimizing equipment for these new parameters is an opportunity,” Wong noted, “and as we start to see new thin films for flexible technologies, there will be more opportunities for deposition and other tools.”
The other near-term opportunity appears to be in materials, particularly for polysilicon, which is currently in short supply. According to Semi, polysilicon production totaled 26,000 metric tons in 2005, with about one-third consumed by the solar industry, whose use is growing 30%/year. Prices have been driven up from $24/kg in 2003 to $36/kg in 2005. Moreover, unlike with ICs, for which the material cost is relatively a small percentage of the total bill of materials, silicon represents roughly 30% of the total solar cell cost. Given these trends, noted Wong, any change in pricing can make a big difference. “There is room for improvement in equipment and materials in this area, but the key will be to focus on the synergy between both to find ways to make polysilicon much less expensive and hit a market that’s growing.”
The key driver for PV is the pricing of solar energy vs. traditional fossil fuels. “Solar is currently about two to three times more expensive than traditional electricity, without subsidies,” said Wong. “If you look at other regions such as Europe with high subsidies, the cost is almost the same.” But solar pricing is likely to follow the semiconductor cost curve, because of its huge wafer component, and because over the next decade it will reach parity in cost with traditional on-grid energy-generation options, he said. “That will really cause a shift in thinking.”
Echoing Wong’s message at ISS was Dick Swanson, president and CTO of photovoltaics manufacturer SunPower Corp., which merged recently with parent company Cypress Semiconductor. “Things are starting to become very exciting in the photovoltaics field,” he said. “The market is growing fast, costs are coming down, and there are synergies with the semiconductor industry in that we process lots of wafers.”
According to Swanson, the market has sped along at a 40% compound annual growth rate for the last five years, achieving a number of milestones. For example, in 2000 the solar industry produced a cumulative 1GW of power, approximately equivalent to one large power plant. In 2004 it began generating more than 1GW of power/yr. Also, in 2003 the PV industry processed more sq. in. of wafers than the microelectronics industry, and by next year – even though the wafers tend to be much thinner than those used for ICs – it likely will consume more silicon as well.
After years of underfunding by large oil firms, the PV industry fell decades behind the IC industry’s manufacturing expertise, but several companies have demonstrated the benefits of synergism with semiconductor manufacturing. Sharp, the number five or six player a decade ago, now is by far the largest manufacturer today, and has a PV manufacturing plant contiguous with an IC fab. “They have had that synergy between the dynamic semiconductor industry and the PV industry, and have incorporated their manufacturing excellence into their PV operations, which has been hard for the rest of the industry to respond to,” noted Swanson.
The economics of solar technology also has undergone a steady improvement. Since about 1975, the cost of producing PV modules over time, plotted by cumulative production, has been approximately 80% – meaning that for each 100% increase in cumulative output (measured in megawatts), costs have dropped by 20%.
This caught the industry by surprise, said Swanson, because PV modules were seen as a costly energy source, given the laborious and expensive process of making wafers to produce a relatively small amount of power. But through the continual experience gained, he said, the costs of polysilicon and of producing PV arrays have come down tenfold. “The PV industry owes a tremendous debt to the microelectronics industry in so many ways,” he contends, “but this is one of the big ones.”
At the same time, the efficiencies of the PV arrays are suddenly going up. For years, with everyone focused on reducing costs, “there was a kind of generic solar cell with an efficiency of about 14% that was produced by a generic process,” Swanson said. “Now the understanding of the value of improving efficiency is changing dramatically. The projection is that the efficiency will be greater than 20% on average by 2012.”
All of these factors are driving the economics toward a tipping point. Today, the cost of making a PV module is about $2/watt, but by 2012, that is anticipated to be $1/watt, a level at which no subsidies will be needed for any application in the distributed power grid. “We think that will happen, and we have a roadmap to get there,” Swanson said.
SunPower also projects aggressive capital cost goals for the industry. Currently, a typical PV processing line costs about $1/watt, said Swanson. The target is about half that, meaning that a 200mW line/year, producing ~8000 wafers/hour, will need to be built for about $100 million.
In terms of the tools and materials markets, Swanson placed each in the $1 billion/year range within five years, but predicted that they would reach $10 billion by 2020. “Because PV will be cost-effective in most places in the world, it will start to be a significant part of new capacity installations and ratchet up quickly after that,” he said. “Now is the time to get your foot in and begin playing in the industry.” – Part One: B.H., Part Two: P.L.