Hemlock Semiconductor forecasts Global Polysilicon Supply capacity will exceed Solar Demand until at least 2013.
“There will be a surplus of silicon supply for the next several years.”
 |
| From Intersolar North America |
Hemlock Semiconductor Group Vice President Gary Homan presented “The Silicon Supply Chain” just after CaliSolar at Intersolar North America last month.
Beginning with a quip, Mr. Homan said:
After that last presentation maybe I should just say we’re closing up shop because there’s no future for polysilicon. But I won’t do that.
The above slide, HSC’s Global Polysilicon Supply/Demand Forecast (Adjusted for Grid Parity Costs), captures Hemlock’s perspective on Polysilicon supply and solar and semiconductor end market demand through 2014.
If the demand for crystalline silicon solar modules grows by 50% per year through 2013, only then will polysilicon supply and demand achieve equilibrium assuming polysilicon production from incumbents and new entrants. At moderate 30% demand growth, equilibrium won’t even happen by 2014 for incumbents alone.
Using a zero based approach of silicon supply minus electronics industry use, Hemlock projects 173069 MT (Metric Tons) of polysilicon and other silicon feedstocks could be produced in 2013 to enable the manufacture of almost 26 GWp (GigaWatt-peak) of silicon solar modules implying 6.67 grams per Watt silicon utilization. These potential module production numbers exclude thin films including amorphous and silicon inventories.
Perhaps more to the point, the photovoltaic (PV) industry could have doubled silicon solar module production in 2009 to 10 GWp based on silicon supply. Exiting 2008, an estimated 400 to 700 MWp (MegaWatt-peak) of material was in inventory either as polysilicon or modules.
As Mr. Homan explained, Hemlock first tallied the public announcements of silicon capacity expansions by incumbent suppliers, new polysilicon producers, and Upgraded Metallurgical Silicon (UMG-Si) entrants. OCI Company Ltd. (was DC Chemical) and M.Setek Co., Ltd. were added to the polysilicon incumbents while Chinese suppliers going online this year remained in the “Others” category. Mr. Homan said some company announcements were discounted up front as “smoke and mirrors.”
Next, through the reality lens of the silicon costs required to achieve grid parity, Hemlock derated the capacity announcements from incumbents, new polysilicon, and UMG-Si entrants with decreasing factors. Mr. Homan doesn’t place much feedstock in UMG-Si even though Dow Corning, a joint venture partner in Hemlock with a 63.25 percent ownership stake, offers solar grade silicon metal products.
 |
| From Intersolar North America |
“Don’t lose sight of the goal.”
Regarding the silicon cost required to achieve grid parity, Gary Homan said:
There are some silicon manufacturers in this leading suppliers group that are operating at costs that are significantly below the grid parity numbers. So you should be happy that’s good news that this is going to be a reality long term. Grid parity is very achievable. The silicon portion of the end product is going to be less than 15% probably less than 10% as we look long term into the future based on polysilicon not UMG.
As the final takeaway, Mr. Homan said:
We can do it. It’s possible today to have crystalline based products out there that will meet or exceed grid parity in the very, very near future.
 |
| From Intersolar North America |
What is the history of silicon pricing?
Mr. Homan said, “My legal team has told me I cannot put numbers out there so I’ll let you use your imagination what these numbers are.”
The average silicon pricing is supposed to be on a relative basis. I postulate the left scale is in $20 increments based on 2001 polysilicon pricing in the mid $20’s per kilogram.
On the 2009 polysilicon price trend, Mr. Homan said:
At the end of the day, it is going to be lower than this by the time we get to the end of year. I think the challenge on pricing is don’t fool us again. If prices continue going down, it’s not so much Hemlock because we are committed to our expansions, but it is going to create this price level not just for polysilicon but all across the supply chain. Once you get below reinvestment levels, you stop investing, and you’ll end up in the same cycle.
With spot pricing coming down off ridiculous levels in 2008, excess capacity may drive polysilicon pricing back to historical $40 to $60 per kilogram levels perhaps even retesting the 2001 lows.
“Holy polysilicon, Batman!”
UMG-Si is trailing in the PV POLL.
Actually the silicon demand- supply and price is not driving down the module price as SoG silicon cost constitutes only 15% of the module as mentioned in your article. The demand – supply rather overstock of modules is causing the price down. The silicon plants are long life plants and one or two year of lower demand & price should not influence the decision making of polysilicon plant developers.
The actual problem is that the loan extending bankers who donot see the long term or life term benefit analysis of any silicon project. They are concerned with the current year or next year analysis and benefit. All bankers, government bodies, different institutions and public, agreed that the final source of power for mankind will be solar power but are short sighted and donot assist in development of associated industries of solar power . Mr Jana- +91 94334 00655
I am very impressed with this data. The investment in silicon is still higher than the return, however, when the initial start-up costs are covered the sky’s the limit. It is a shame that the bankers cannot look to the future and beyond the immetiate ROI. Investing in this type of development should have different ROI expectations
There is a rumor that Hemlock has production issues. Is that true?
Wow, that 1st chart turned out to be amazingly wrong.