2006 was a turbulent year for solar photovoltaics, write Travis Bradford and Paul Maycock, but the industry showed substantial resilience. Once again, prices have resumed their historic downward slide.Solar photovoltaics (PV) continues to be a dynamic market segment within renewable energy. Not only has the global PV market continued to grow at astounding rates, reaching nearly US$20 billion of installed systems and components in 2006, but it has also been one of the most dynamic and unpredictable. In just the last year alone, the industry has grappled with polysilicon shortages, massive new capital infusions to support new Chinese and Taiwanese production capacity, the emergence of viable thin film technologies, and substantial changes in dozens of global support regimes.
PV-shaded parking in Los Angeles, California LADWP
Despite it all, the industry has shown substantial resilience. Due in part to the substantial economic support available in a number of places like Germany, Spain and California, and in part to deeper customer desires for locally generated, clean electricity, the PV market’s momentum has remained powerful. Even as average system prices rose in 2006, with little apparently available supply, the industry managed to grow. Now, new capital has raced in to provide new production capacity and alternative technologies, which should simultaneously help to grow volumes and re-establish the historic price decline of PV.GLOBAL CELL PRODUCTION SURPRISINGLY STRONG
In 2006, global cell production grew by 41% to 2520 MW, despite widespread fears of insufficient polysilicon feedstock available to producers. Our survey methodology for global PV production uses direct confirmation of cell production levels from each manufacturer, of which there were 76 globally in 2006. Our primary focus has always been on cell production because of the higher reliability of those numbers across companies and technologies, and the ability to avoid potential double counting.
The global cell production we recorded showed growth across all of the geographic regions, with Japanese production the slowest because of the effects of the polysilicon shortage, and the production in the Rest of the World (RoW) category the strongest, reflecting the rapid emergence of the dozen Chinese and Taiwanese companies attempting to serve the PV market (Table 1).
TABLE 1. Global cell production until 2006 (MW DC). Source: PV News, April 2007
Looking at the top 15 producers of PV cells in 2006 (Figure 1), a few trends emerge. First, Sharp of Japan, Q-Cells in Germany and Kyocera of Japan still represent the top three producers in the world, but of the three, only Q-cells grew at or above the overall industry average, reflecting its strong ability to acquire polysilicon from multiple sources, including its emerging strategic partner Renewable Energy Corporation of Norway.
FIGURE 1. Top producers shift to emerging players – top 15 global cell producers. Source: PV News, April 2007
Beyond the top three, some substantial changes began to occur in the major PV producers in 2006. Suntech of China climbed from eighth place to fourth place and looks set to break into the top three in 2007. Motech of Taiwan climbed to seventh place and will likely be in the top five soon. Other Chinese and Taiwanese producers are placed to enter the top 15 by 2007 as well. A couple of US-headquartered firms showed strength in 2006, with SunPower and First Solar both entering the top 15. More traditional PV producers such as Schott of Germany, BP and Shell fell in rankings due to a lack of access to polysilicon and, at least in the case of Shell, it caused the company to exit the crystalline PV business by selling its operation to SolarWorld of Germany.GLOBAL INSTALLATIONS UP 35%
While cell production is the easiest element to measure and compare, it is equally important to understand how those cells get converted into useful modules and products and then installed so they generate electricity. We survey downstream users and installers globally to understand the installations and applications of PV and the prices and economics affecting that demand.
TABLE 2. Global PV cell use by application. Source: PV News, July 2007; PV Energy Systems
What was most surprising in 2006 was the simultaneous growth in volumes of installation and the prices for PV cells and modules, evidencing a very strong underlying demand for PV. Installations grew in line with global production, as supply shortages meant that any available products were sold and installed with the normal lag time in the supply chain. Our cell use by application chart (Table 2) shows how the spread of PV across different markets remained concentrated in the grid-tied and off-grid applications, representing over 90% of PV installations.
TABLE 3. Global PV module installations (systems only) from 2000 to 2006 and forecast for 2007. Source: PV News, July 2007
The geographic spread of the modules used in those system-based applications (Table 3) showed that Germany remained the dominant market for PV, with over 56% of the system-based modules installed, and a growth of 50% over 2005. The rest of Europe showed extremely high demand growth as Spain ramped up its generous PV programme, and a number of other countries grew by a little. The US market grew by 31%, with a strong showing in California and New Jersey. However, the Japanese market grew by only 10% as that market continued to adjust to the elimination of the government support programme in 2005.MODULE PRICES RISE BUT TOP OUT IN Q4
The result of strong demand – particularly in Germany, Spain and California – in an environment of limited polysilicon supply (and resulting cells and modules) was that prices continued to rise in 2006. Module prices rose from an average of about $3.50/W in 2005 to $3.75/W or more in 2006, and were about $4.00/W in the fourth quarter of the year. The price rise was consistent throughout the year as more and more polysilicon and modules were contracted and available spot market quantities shrunk. The rise in module prices topped out in the fourth quarter, and reported average selling prices for modules had dropped by at least 10% by the second quarter of 2007, with expectations of further price reductions in late 2007 and 2008.
Building-integrated PV in Shenzhen, China BEIJING JIKE CO
The rising prices for modules led to rising process for polysilicon as well, with long-term contract prices going from $55/kg to over $65/kg, and spot prices being reported at over $300/kg. These prices allowed for an unprecedented expansion of profitability at the polysilicon production level. It has also created a market opportunity for thin film competitors such as First Solar in the US to emerge.US PV PROSPECTS BRIGHTEN
The US PV market remains murky and uncertain, but a few powerful elements emerged in 2006. First, US cell production has continued to shift away from traditional wafer-based technologies in favour of thin film based PV. Second, US demand was dominated by California and its new $3 billion PV rebate programme, but nearly a dozen other states have shown significant interest in stimulating PV use, setting the stage for strong and more broad-based PV growth in the US market in coming years.First Solar alone drives US production
PV production in the US managed to grow a robust 31% in 2006, but a closer look at the numbers shows that almost all of that net growth came from a single company, First Solar. This company exploded on the market with 60 MW of its cadmium telluride product onto glass that was substantially cheaper on a per watt basis than its silicon-based counterpart. First Solar’s growth will continue in 2007 as the US plant reaches 100 MW of capacity, and new plants in Germany and Malaysia come on-line as well.
TABLE 4. US PV cell production from 2000 to 2006 (MW DC). Source: PV News, April 2007
Beyond First Solar, Unisolar’s amorphous silicon-based thin film showed a strong 27% growth and plans to scale up through 2010 to 300 MW of production. The polysilicon-based PV production in the US was limited by available polysilicon feedstocks, and collectively Shell (now SolarWorld), BP, GE, Evergreen and Schott barely held their ground from 2005 (Table 4).Grid-tied demand strong in multiple states
While US production growth was concentrated in thin film technologies, US demand growth was concentrated in California and New Jersey (Table 5). Together, these two states represented nearly 90% of all US grid-tied applications in 2006, not including the 40 MW of off-grid PV used in the US. The funding of support programmes in both states was the primary driver, but a number of other states also began to show substantial support, which should translate into more broad-based PV market growth in coming years. States like New York, Nevada, Maryland, Connecticut and Oregon all increased the rebate support for PV and should see surges in growth, while Washington State installed the first feed-in tariff in the US, trying to emulate the success of similar programmes in Europe.
TABLE 5. US PV grid-connected module use in 2006 by state (kW DC). Source: IREC and Larry Sherwood
The market in the US was driven primarily by the commercial sector (Figure 2), with grid-tied commercial systems making up 60% of all US grid-tied installations. The reason why commercial users adopted PV at such high rates was the emergence of very innovative financing models for deploying commercial PV using power purchase agreements. This approach, pioneered by SunEdison of Maryland, allows customers to use their roofs for PV without having to fund the installation or take on risk of performance, paying only for the electricity the systems generate. The proportion of US commercial PV installations using this financing method went from nearly none in 2005 to probably the majority in 2007. These types of financial innovations will continue to propel US market growth in the next few years.
FIGURE 2. US PV grid-tied module use 2006 (by application) Source: IREC and Larry Sherwood
JAPAN’S GROWTH SLOWS
In contrast to the dynamic growth of US PV demand and rapid growth of US thin film companies, the Japanese market seemed stagnant in 2006 (Table 6). PV cell production grew by only 11% in 2006 from 2005 as major Japanese producers found their access to vital polysilicon feedstocks limited by both rising prices and the physical limitations of domestic polysilicon producers. Both Kyocera and Sanyo managed to grow by over 20%, but Sharp’s flat production on a large share of the market kept overall growth down. Japan’s sole pure thin film producer, Kaneka, managed growth of over 40%, giving further evidence that thin film technologies are enjoying a market opportunity created by high polysilicon and module prices.
TABLE 6. Japanese PV cell production from 2000–2006 (MW DC). Source: PV News, April 2007
One of the most interesting aspects of the tepid Japanese growth has been the dearth of announcements for expansion of domestic polysilicon. Three of the established seven polysilicon producers worldwide – Tokuyama, Mitsubishi and Sumitomo – are Japanese. But while major new polysilicon production has been announced by their counterparts in Europe and the US, Japanese polysilicon producers have been very slow to announce new production plans. With the domestic PV support programmes from the Japanese government ending in 2005, perhaps the resulting market uncertainty has caused domestic manufacturers to pause and take stock of market strength before committing substantial capital to new capacity expansions.
New York: building-integrated thin film roof SCHOTT SOLAR
EUROPEAN GROWTH SPEEDS UP
In stark contrast to the tepid Japanese market, European cell production and installation has been growing wildly. European production of PV grew to 657 MW in 2006, up 40% from 2005. The growth in production was broad based, with a number of German companies like Q-Cells, SolarWorld, Schott and Ersol being joined by Norway’s Scancell, Belgium’s Photovoltech and Switzerland’s Solterra in growing overall production. Isofoton of Spain was also a strong player, but the company was building new capacity in 2006 and also entered into a JV to produce polysilicon in Spain and to ensure access to feedstocks (Table 7).
TABLE 7. European PV cell production from 2000 to 2006 (MW DC) Source: PV News, April 2007
As previously mentioned, the market in Europe was the strongest in the world in 2006, with Germany growing 50% to 1050 MW and Spain tripling to 70 MW in 2006. Italy was the only national market in Europe that experienced a decline in PV installations in 2006, as the country’s second PV subsidy programme spectacularly failed to launch. With new or revised programmes in Italy, France, and Greece in 2007, European growth may become more broad-based in the next few years, but substantial risks remain in the execution of those programmes.RoW GROWTH EXPLODES
While PV producers in the traditional PV-producing nations scrambled to procure necessary polysilicon feedstocks, often through signing multi-year and high-priced, fixed-supply contracts with substantial down payments, a set of emerging producers in China and Taiwan were driving prices further in both contract and spot polysilicon purchases. Led by China-based Suntech and Taiwan’s Motech, these producers challenged US, Japanese, and European producers for access to polysilicon, wafers and the production equipment necessary to convert it into cells and modules. Using capital that was raised in either domestic equity markets or in issues on London’s AIM exchange or the New York Stock Exchange, these companies aggressively deployed billions in capital to upstream producers and used the high market price environment to outbid for spot polysilicon supply.
TABLE 8. RoW PV cell production from 2000 to 2006 (MW DC) Source: PV News, April 2007
Specifically, Taiwanese producers doubled production in 2006 to 177.5 MW, while Chinese producers nearly tripled to 369.5 MW. Outside of these countries’ producers, the remaining RoW producers (Table 8) did not show substantial growth. However, recent announcements from multiple Indian conglomerates using thin film production equipment from Applied Materials indicate that India may yet re-emerge as a strong growth country in the next couple of years. Also, a number of rumoured deals in the Middle East, including the United Arab Emirates, suggest that well-funded energy exporters are looking to diversify their fuel mix into solar and PV technologies.
One of Germany’s growing number of large ground-mounted installations SHARP
POLYSILICON STILL DOMINATES, BUT FOR HOW LONG?
Clearly, the dominant issue in 2006 was the limits to growth that available polysilicon supplies created. Even when available inventories and sub-prime silicon supplies were drained, the underlying demand was not met, and module and polysilicon prices had to rise to clear the market. Despite the shortage, polysilicon-based technologies still comprised 93% of the market, down from 94% in the prior year. The thin film technologies gained a small amount of market share for the fourth year in a row, having risen from 4% in 2003 to nearly 7% in 2006 (Table 9).
TABLE 9. PV cell production by technology from 2003 to 2006. Source: PV News, May 2007. Source: PV News, April 2007
We have looked at the capacity and projected production growth for both polysilicon and thin film based technologies and created forecasts through 2010 based on some combination of announcement, anticipations and our understanding of the powerful obstacles to getting new production built and on-line. Our polysilicon forecast shows a substantial amount of new supply coming on-line in 2008 and 2009 as plants begun in 2006 are completed. After taking into consideration potential new entrants, improvements in existing plant efficiencies, and inventory effects, we estimate that nearly 11 GW of cell production could occur in 2010 (Table 10).
TABLE 10. Total polysilicon supply (with inventory and overage effects) until 2010. Source: PV News, May 2007
Thin film production is also expected to grow rapidly in the next few years – to at least 2.5 GW in 2010 – as three forces combine to provide market opportunity. First, existing producers like Unisolar, First Solar and Kaneka will rapidly ramp their technologies with no equipment or feedstock obstacles. Second, new standardized amorphous silicon production equipment adapted from the digital display deposition industry will be sold by companies like Applied Materials and Oerlikon to customers with local distribution channels in Europe, India and the US. Third, scores of well funded and technically competent companies, with technologies ranging from CIGS to cadmium telluride to dye-sensitized (Table 11), will come to market with product, and while each has a low probability of success, only a couple of successes could substantially add to the thin film momentum.
TABLE 11. Projected and potential thin film PV production until 2010. Source: PV News, June 2007
With all of this supply available to be produced in 2010, the biggest question facing the PV industry now is how much demand will exist and what prices will need to be to equilibrate supply and demand through 2010. Most demand through 2010 is based on policy programmes that currently exist and are being rolled out. Our initial estimates suggest that module demand could range from 4.5-8.5 GW in 2010, depending on a wide range of market specific factors. This would be equivalent to between 5.4 GW and 10 GW of cell production for comparison with the supply numbers above.
The result of such supply and demand outcomes would be overcapacity in 2010, ranging from 15% to over 50%, and a price response to match. We have forecast a drop in prices of 20% in 2007 from Q4 2006, another 20% price drop in 2008, and 10% drops in both 2009 and 2010. These price drops mean that module prices would be under $2.00/W in 2010 for crystalline silicon and under $1.50/W for thin films.CONCLUSION
Already, 2007 has shown that prices are beginning to resume their historic slide, and that the worst of the polysilicon shortage is likely over. The PV market will continue to grow, and political and social momentum for the technology will likely do so as well. We continue to believe that the PV industry is only a few years into a multi-decade bull market, and that PV will continue to grow to become a dominant component of our global energy architecture. In order to do so, new financing and delivery models will need to emerge, new products must be developed to meet new applications, and prices will have to come down across the board with very unpredictable implications for suppliers and customers. However, two things are certain about the PV market: it will continue to grow and it will be fascinating to watch how fast it will change and adapt to obstacles and opportunities.
Travis Bradford is founder of the Prometheus Institute
Paul Maycock is a solar energy consultant and Managing Director of PV Energy Systems inc.