A trio of Wall Street analysts offer their takeaways after a first day of meetings and sessions at Solar Power International (SPI). Common themes: Optimism for US and China demand, prices still falling, no capacity expansions, and a telling example of supply-chain pushback on one major solar firm.
October 19, 2011 – A trio of Wall Street analysts offer their takeaways after a first day of meetings and sessions at Solar Power International (SPI). Common themes: Optimism for US and China demand, prices still falling, no capacity expansions, and a telling example of supply-chain pushback on one major solar firm.
Companies are optimistic about US demand in 2012, and think current estimates are too conservative, reports Credit Suisse’s Satya Kumar. 3Q11 demand was “very weak,” although German demand has been picking up in October. “Solar companies are still relatively cautious about demand trends in Europe but more bullish about demand in the US and China,” agreed Deutsche Bank’s Vishal Shah. “Most solar companies expect US demand to take-off.” Chinese demand continues to improve, heading toward 1.5GW in 2011 and 3-5GW in 2012, Kumar says.
Purchases are being postponed because of pricing declines, says Kumar, citing multiple companies as sources. They think pricing declines will flatten out in 2012 and eventually drive demand elasticity. A -40% dropoff for panels hasn’t yet been enough to “meaningfully” pick up end-demand in Europe (financing has been a factor in Italy).
So where and when is the trough, from which the industry can expect that demand elasticity? Not until companies cut production through late 4Q11 and into 1Q12 — and only when more consolidation occurs among Tier 1 and Tier 2-3 companies, something that “has not yet started to occur,” Kumar says.
Pricing visibility is notoriously short-term (“on a week by week basis,” says Kumar) and not an overall gauge (Tier 1 Chinese companies only have visibility for a small percentage of contracts, says Shah). That said, the analysts offer the following insights:
– Polysilicon: Poly prices have “approached” $35/kg (Tier 1 non-Chinese suppliers), Kumar says, and most poly-Si contracts are being negotiated down with some exceptions (e.g. Wacker) holding the line due to prepayments and more favorable mix. (Yingli Green Energy lists a $35-$40 range, says Kumar.) He still claims poly spot prices will touch $25/kg by year’s end, as production winds down in late November due to a seasonal demand dropoff.
Deutsche Bank’s Shah thinks poly pricing bends to $30/kg by year’s end and $25/kg in 1Q12, and maybe even lower & faster given midstream firms’ production shutdowns and procurement stoppages. Meanwhile, GCL Poly has now developed fluidized bed reactor (FBR) polysilicon production that could take its costs down to $15/kg, points out Citi’s Tim Arcuri. (And thus MEMC, he translates, needs to wring its profits out of its SunEdison business, because it “is still unlikely to make any money selling solar wafers.”)
– Wafers: Prices are now $0.35-$0.40/W, with GCL holding the floor, Kumar says. Conversion cost from poly to wafer is ~$0.20/W, which means companies are “at best near cash break even (assuming 3-4c/watt of depreciation for wafering) at current poly prices,” he writes. And some cell & panel makers with internal wafering capacity “are very close to shutting down their production for wafer making,” he says.
– Modules: Tier 1 Chinese modules are now selling at ~$1.00-$1.10/W, says Shah. Assuming continued drops in polysilicon prices (as much as another $10/kg within the next six months), those module prices could decline another $0.05-$0.10. Arcuri says module prices are $1.00-$1.05/W and “only 10%-15% gross margin at best” for Tier-1 producers (and probably lower for Trina and Suntech).
Companies are bracing for “a difficult 3-6 month period” yet with “refreshingly realistic” expectations, Arcuri noted, with most of the focus is still on supplies. “The vast majority of producers appear to have still over-produced demand in CQ3 meaning inventories are apt to rise again,” he writes. So far in 4Q11 some Tier 1 module makers (e.g. Trina Solar and Suntech) have started cutting back production, but others (Yingli) “are still running flat-out hoping for an end-of-year surge from Germany and, to a lesser degree, the US,” Arcuri writes.
Inventory is still high all across the value chain and will increase in 3Q11, Shah says. Yingli, cited by Kumar, thinks 70-75 days sales outstanding (DSO) “is reasonable.”
First Solar’s bigger channel partners “are pushing back hard on volume commitments in exchange for significant price concessions,” Arcuri says, and FSLR is staring at the possibility of $0.75-$0.80/W merchant pricing for 1Q12 or risk “European channel partners outright walking away from volume commitments.” Note that also suggests FSLR’s “core EPS” could swing to negative for a couple of quarters (though its “headline EPS” including its EPC pipeline will stay well in the black).
Tracking capacity (more to the point, whether there’s now a great excess of it) is an inexact science. Kumar laments that “there is massive excess capacities for wafer/cell/panel that is hard to track,” citing as examples several non-US listed companies that are on Asian markets have significant capacities, e.g. Eging (1GW) and Hareon (2GW cell). Capacity expansions by major players are at a full stop, says Kumar.
Nevertheless, the signs seem to be clear to all three analysts: companies in the mid-stream supply chain “are likely to shut down production in late Q4 and have completely stopped new capacity expansion,” says Shah. “Generally, capex plans have been placed on hold until end market conditions improve,” Shah says. For now, firms prefer to reduce working capital vs. reducing capacity utilization (at least not yet). In fact, says Kumar, Trina Solar, Yingli Green Energy, and “a Tier-1 cell maker” are running at or near full utilizations now, and that cell maker is responding to “rush orders finally coming from Germany.” (Deutsche Bank’s Shah claims “a 4Q11 demand rush from Germany has not yet materialized.”)
Yingli won’t be adding “meaningful” capacity for the next year, but the company is reiterating its 1.7GW nameplate capacity guidance for 2011, and says it can actually produce 130% of that amount, according to Kumar. The company sees 15%-20% cost reductions each year through 2013 for its n-type mono Panda technology, vs. only 2%-5%/year through 2015 for multi.
Kumar specifically pointed to First Solar’s new module (FS3-387) with calculated 12.15% efficiency, higher than the company’s reported averages. Meanwhile, BYD is expanding capacity to 1GW with a new higher-efficiency process — reminiscent of Suntech’s Pluto architecture, he says — generating 17.4% average cell efficiencies.