Your daily dose of Photovoltaic Technology Developments and Solar News

News Feeds | Newsletter | Advertise | About Us | Subscriptions

Order your copy of Photovoltaics International Edition 1#

Conversion inefficiency: Whose solar manufacturing forecast should we believe?

22 September 2008 | By Tom Cheyney | Chip Shots

forecast1 Years of tracking semiconductor market prognostications have taught me a thing or two about statistics and the art of high-tech sector economic forecasting. One truism: the farther out the forecast, the more wobbly the numbers tend to be.

But at least the purveyors of predictive studies in the semiconductor sector have tightened up their ship somewhat, updating their forecasts on at least a quarterly--if not more frequent--basis. With a maturing market, the longer-term number crunching has at least a semblance of reality as well.

A couple of recent forecasts from the solar manufacturing sector, however, make me wonder whether someone is using that time-honored method of predicting market growth--the dartboard.

Although I didn't attend SEMI's PV luncheon in Santa Clara last week, I did get ahold of the three presentations: Oerlikon's Chris O'Brien on thin-film PV; a supply-and-demand perspective on solar's global prospects from Paula Mints of Navigant Consulting (who I'm told didn't actually present her info, but had someone stand in for her); and Linx Consulting's Mark Thirsk on the market for PV materials makers.

I skimmed through the presentations and found a few nuggets, although Mints' and O'Brien's info was stuff I'd already heard or seen several times over, in various incarnations. I found a bit of fresh material (pun intended) in Thirsk's info (whose total forecast for the solar manufacturing materials market--crystalline silicon, thin film, and moduling--is in the neighborhood of $5 billion by 2012).

Among his foils was a list of PV material needs and opportunities, which included:

Silicon recovery in sawing slurries (potential key resource saving; technical breakthrough required).
Diamond wire (cheaper wafering, less surface damage).
Improved texturization etches (higher cell efficiency).
Cleaning technology (improved line yield, better cell efficiency).
Patterned emitters and MWE (reduced cell shadowing, improved ease of moduling).
Noncontact metal deposition (reduced breakage, reduced materials cost).
Chamber cleaning gas (environmental concerns, reduced global warming potential [GWP]).
Backsheet alternative (UV protection sheet supply).

But it was a couple of graphs among O'Brien's slides that really got me shaking my head and wondering how accurately he threw his predictive darts.

The Oerlikon head of market development for North America has one pie chart showing the total solar market size reaching $150 billion by 2015, with thin films accounting for 52% of the total, crystalline silicon 33%, and concentrating solar power the remaining 15%.

That 12-figure number far exceeds any other forecast I've seen, and nowhere else have I heard his claim that more than half of all solar will be thin film in about seven years! A bigger piece of the pie, yes, since TFPV is the fastest growing segment of the industry and promises to continue to do so. But usually the percentages presented by other market researchers come in closer to 25-35% by the mid 2010s.

Since I didn't attend the lunch, perhaps he offered some explanation of this or an audience member posed a challenging question, but the inflated amount comes off as suspiciously self-serving for someone in the TFPV business.

But another chart has me reeling even more. O'Brien's figure depicting the TFPV equipment market growth outlook shows a compound annual growth rate of 59%, leading to the sector's worth reaching $13 billion by 2011. His 2008 number? About $4 billion.

Now let's compare those findings with those in the just-released NanoMarkets report, "The Future of Thin-Film and Organic Photovoltaics Manufacturing," and the same company's "Thin Film Photovoltaics Market: 2008 and Beyond" issued a couple of months back.

(Editor note: NanoMarkets latest thin film equipment sales forecast figures have been updated on a recent story here).

NanoMarkets forecasts that the entire TFPV market will be about $6.4 billion in 2011 in the latter report, and, in the newer document, that the total amount spent on TFPV equipment--captive and merchant--will reach about $1.4 billion in 2011, based on an installed capacity of 6.8 GW. (btw, NanoMarkets' 2012 TFPV capacity projection is 8.9 GW, which is not too far off from the 9.6 GW "projected capacity" forecasted in the new Prometheus Institute report.)

The firm's tool sector prediction for 2015 reaches $4.8 billion, based on an installed capacity of about 29 GW. The same year, NanoMarkets posits a TFPV market of more than $22 billion.

So my question is, whose numbers should we believe? O'Brien's numbers are incredibly optimistic compared to NanoMarkets'--or anyone else's I've seen. It's not just a billion here or there, but many billions in discrepancies between the two. I'm not saying that Nano has the magic forecasting formula, but the company's prognostications do seem to err, if anything, on the side of caution in this case.

Any prospective (and current) equipment supplier to the TFPV sector should take a hard look at every market forecast it can lay its eyes on before accepting any of them as reliable enough to base a business plan on. The smoke seems to be quite a bit thicker and the mirrors more carnival funhouse than usual compared with the forecasting foibles that I'm used to in the semiconductor space.