R&D spending patterns
There has been a direct correlation between the growth of the PV industry since 2007 and increased allocation of financial resources from leading PV module manufacturers towards R&D activities.
However, in 2012 that trend was reversed for the first time as the industry struggled to deal with the second consecutive year of chronic manufacturing overcapacity, declining ASPs and financial losses for all but one CdTe thin-film manufacturer.
With capital expenditures primarily relegated to maintenance, combined R&D spending at 12 of the major module suppliers decreased from the highs seen in 2011. R&D spending in 2012 declined to US$464 million, down from a peak of US$510.4 million in 2011.
Separately, the eleven c-Si producers within the featured group spent US$332 million on R&D activities in 2012, down from the US$369.9 million peak of 2011.
However, the overall declines are relatively mild. The total decline was only US$46 million, or a 9% decline year on year. The c-Si producers cut spending by only US$35 million, or a 9.5% decrease, year on year.
The mild decline suggests that despite financial woes, companies are increasingly aware that R&D spending is a critical part of future earnings and differentiated product offerings. It should also be noted that in the past, certain companies have said that they lowered spending due to the conclusion of specific R&D projects.
However, seven companies reduced spending in 2012 from 2011. These included, First Solar, Yingli Green, Trina Solar, Canadian Solar, REC Solar, ReneSola. With respect to Suntech, we have estimated that R&D spending declined from US$38.6 million in 2011 to US$19.2 million in 2012. The estimated decline is based on spending peaking in 2010 and the financial condition of the company over the 12-month period.
Since 2009, First Solar has topped the R&D spending ranks and has led by a wide margin compared to its rivals. However, First Solar curtailed spending in 2012 by around US$8 million compared to the previous year, which saw spending peak at US$140.5 million.
Yet the key to the overall spending declines were the relatively large cuts made by four of the seven companies that reduced spending – REC, Yingli Green, Trina Solar and Suntech.
Importantly, five companies actually increased R&D spending in 2012, including SunPower, HanwhaSolarOne, SolarWorld and Jinko Solar. Noteworthy was SunPower, which increased spending from US$57.8 million in 2011 to US$63.4 million in 2012. SunPower has been a consistent high-ranking spender throughout the period covered, one of only four companies to do so.
At the other end of the scale, Hanwha SolarOne increased spending from US$10.8 million in 2011 to US$14.5 million last year. However, the company remained in the lowest group of spenders (below US$20 million per annum), which included Canadian Solar and the leading R&D spending laggard, Jinko Solar.
Guidance on R&D spending in 2013 remains limited; although companies typically guide shipment and revenue forecasts, few actually provide guidance with respect to R&D expenditure. Currently REC and SolarWorld have indicated that 2013 spending would decline to US$22 million and US$52 million respectively.
R&D staffing patterns
There is a close correlation between R&D spending and staffing levels within R&D departments. Increased spending on R&D often includes costs with associated increased staffing levels.
It is clear that there was an explosion in R&D staffing levels amongst the major producers covered in this report, compared to small but noticeable increases between 2009 and 2010. However this is not noticeably mirrored in overall spending levels for the same period.
Dedicated R&D staffing levels increased from just over 1,500 in 2009 to just short of 3,500 in 2010, an increase of approximately 125% year on year.
It should be noted that although REC provides detailed employee data in its annual reports, the company does not break out employee levels dedicated to R&D. This is also true of First Solar.
Recently, REC explained to PV Tech that the reason it does not breakout R&D staffing levels separately is due to the close connection between engineers working in R&D with production line engineering staff. This could also account for why First Solar does not break out R&D staffing numbers.
Therefore our staffing numbers for both companies are estimated but based on R&D spending and staffing levels in the sector in generally, as well as research that provides indicators to R&D staffing levels.
Although spending leader, First Solar, has a strong R&D headcount, the c-Si manufacturers have started playing catch-up, pushing staffing numbers higher in recent years. Overall, R&D headcount for the leading manufacturers has jumped significantly at key times in recent years rather than growing at a steady rate.
Staffing levels have continued to increase since 2010, albeit at a significantly slower rate, reflecting the period of profitless prosperity.
However, the continued increase in R&D staffing levels throughout the last six years, belies the fact that five companies (Trina Solar, Jinko Solar, Hanwha SolarOne, ReneSola and SolarWorld) reduced their R&D headcount in 2012.
Suntech is estimated to have reduced its R&D headcount in 2012, due to headcount peaking in 2010 and the estimated reduction in R&D spending due to its financial position over the last 12-months.
There were five companies (Yingli Green, First Solar, Canadian Solar, JA Solar and SunPower) that increased R&D-specific staffing levels in 2012.
Noteworthy is the fact that the massive increase in R&D headcount at Yingli Green in 2010 was not another anomaly. Yingli Green has continued to increase staffing levels, which reached a record high for the company and the sector in 2012 at 1,464.
At the opposite end of the scale, the laggard in R&D headcount is harder to reveal, notably due to the inconsistent staffing levels of some companies over the period covered. Although Jinko Solar is one of the laggards, so too has been JA Solar.
However, if we take a laggard to mean a company employing fewer than 100 in R&D activities in the last three years, then none could be called a laggard in 2012. However, JA Solar and SolarWorld would have been classified as laggards in previous years. The company with the lowest R&D headcount at the end of 2012 was actually Hanwha SolarOne, with 104.
Six companies in 2012 (Hanwha SolarOne, JA Solar, SolarWorld, Jinko Solar, ReneSola and Canadian Solar) had headcounts of between 100 and 200.
R&D spending rankings and analysis
Leading in R&D spending rankings has historically had little correlation with the same companies ranked on a revenue, shipment or production level basis. Only First Solar has achieved a market leadership position within a given year while holding the leadership position in R&D spending. At the other end of the scale, JinkoSolar has seen its market ranking rise rapidly over the last few years and currently is market ranked seventh but has been a very low spender on R&D.
At the lower laggard level, Canadian Solar, Jinko Solar, and Hanwha SolarOne have consistently under-spent compared to rivals yet are market ranked 5, 7 and 10 respectively.
However, in respect to JinkoSolar and Hanwha SolarOne, both companies increased spending in 2012, yet JinkoSolar more than doubled spending. In respect to Hanwha SolarOne the company is fast-tracking its assimilation of former Q-Cells cell and module technology in a ‘copy smart’ strategy, which could mean spending continues to rise modestly over the next few years.
Another trend highlight is the fact that three manufacturers (SolarWorld, ReneSola and REC) have not been in the top 10 market rankings yet have consistently been positioned highly in the R&D rankings.
Although REC had led the R&D rankings ahead of First Solar, the company only dropped to near middle table last year. Primarily, REC had spent heavily on developing its FBR polysilicon and wafer technology.
Wafer technology investments at ReneSola also dominated R&D activities and are expected to continue but also include the development of its Virtus wafer/cell and module offerings.
In the case of SolarWorld, recent high spending activity has been focused on monocrystalline wafer technology as well as the development of high-performance PERC cell technology. SolarWorld expects to commercialise both activities in 2013, which could explain its guidance that R&D spending would decline in 2013.
There is a certain sense of relief in being able to correlate First Solar’s and SunPower’s consistently high spending strategies with their cell and module conversion efficiency records.
SunPower recently launched its ‘X-Series’ module that employs its latest ‘Maxeon Gen 3’ solar cell technology, which has conversion efficiencies of over 24%. The modules have efficiencies of 21.5% and above.
First Solar has recently reported a range of record CdTe thin-film cell and module efficiencies both in the lab and at the manufacturing level. Already this year, R&D test cells produced at First Solar’s Perrysburg, Ohio factory and R&D centre have been verified by NREL to have reached a new record cell conversion efficiency of 18.7%.
The company has also been focused on bringing higher cell and module efficiencies from the lab to fab at a faster pace than ever before.
However, the news in February 2013 that it would acquire US-based monocrystalline solar cell start-up, TetraSun, shook the industry. Having exited R&D into competitive CIGS thin-film technology and rejuvenated its commitment to CdTe technology, First Solar has plans to start “tentative” production of TetraSun’s copper-based monocrystalline cells in the second half of 2014.
It would therefore be reasonable to predict that First Solar’s R&D spending levels will continue to be at high levels for several more years to come. Overall, R&D spending for the majority of the group has been closely tied to new product introductions but with an emphasis on development of lower cost processing.
According to the latest edition of the International Technology Roadmap for PV, significant changes, especially in the field of solar cell design and related move to N-type mono wafers and a switch to copper metallisation starting in 2015, are predicted. Indeed a whole suite of evolutionary changes from wafers, materials, cells and modules are expected that should require and have already forced some leading companies to continue to invest in R&D activities, a trend that is expected to continue in the mid-term.