1. Expect prices to stay high through at least the first half of 2022 as the supply chain continues working itself out AND
  2. Expect prices to drop for all markets sometime during the second half of the year as the industry as supply chains recover and overcapacity asserts – except for US buyers as the Senate passed the Forced Labor Bill and eventually, President Biden is certain to sign it into law
  3. Expect the bifacial exemption from the 201 tariffs in the US to stay in place until … it doesn’t, but don’t worry, it will be back
  4. Expect manufacturers to continue adding n-type cell capacity even as high purity polysilicon additions lag
  5. Expect TOPCon to be the n and p type capacity addition winner as manufacturers begin the switch from PERC
  6. Expect announcements of new cell manufacturing capacity in Europe and the US
  7. Don’t expect all of these announcements to come true
  8. Do expect India to start adding manufacturing capacity while keeping their tariff in place
  9. Expect many, many, many product and service offerings promising to streamline and speed up module and component deliveries – this is the product of 2022 BUT
  10. Expect most of them to be a waste of money – these new products will likely include:
    1. New charges to ensure delivery from the port to the project
    1. New insurance offerings to cover unexpected supply chain hiccups
    1. New charges from sellers to get the product ready to pass customs in the US and other countries (in the US, these will be specific to the Forced Labor Bans)
    1. New consulting products to help developers and manufacturers understand new rules – advice that is freely available already
    1. New packaging to reduce the time from crate to field
    1. Many many companies offering quality inspections
    1. Etc.
  11. Do expect governments to continue discussing and taking meaningful action on forced labor in Xinjiang and potentially cobalt mining in the Congo
  12. Expect more announcements about Perovskite technology but
  13. Do not expect more installations of Perovskite technology and
  14. Do expect governments to continue investing in Perovskite technology
  15. Dread further buildup of solar manufacturing capacity in China and the countries of Southeast Asia as it makes for a lopsided, unhealthy supply chain with little options when things go awry
  16. Dread continued quality erosion in polysilicon, wafer, cell, and module manufacturing as pilot scale production becomes a thing of the past and shortcuts become the new normal
  17. Expect pure play cell manufacturers such as Aiko and Tongwei to aggressively add cell capacity as others, such as Jinko Solar, Trina Solar, and Canadian Solar, add wafer and module assembly capacity, sending wafers out for processing to be returned as cells
  18. Expect bids on projects to remain high through mid-2022 before dropping like a stone towards the end of the year
  19. Dread the continued use of coal to provide electricity to solar manufacturing
  20. Hope for more distributed solar + storage so that countries can move away from the outdated and archaic utility scale business models
  21. Hope governments invest in updated, resilient, and renewable ready utility infrastructures because a flexible, resilient grid, renewable energy technologies, and storage are the keys to a future of energy independence

22. Expect the solar market to grow despite all obstacles and setbacks because the trend is clear and solar is finally big business

The CPUC is set to vote NEM 3.0 into law, changing the math for prospective residential solar customers. Commissioners will hold a hearing on December 20 and vote in January 2022. Once NEM 3.0 is adopted, the commission will choose a date to sunset NEM 2.0 and transition solar system owners to NEM 3.0.

Currently, solar system owners are compensated at retail rates for electricity generated by their systems. NEM 3.0 will compensate system owners at rates less than retail. The new rates have not been established. NEM 3.0 will change the payback time for system owners. Solar system owners will also pay a grid benefits charge and a monthly grid participation charge. The proposed charges are in the following tables.

Table 7 Grid Benefits Charge Recommendations
PartyPG&E CustomersSDG&E CustomersSCE Customers
Public Advocates Office364$7.66/kW$6.14/kW$5.76/kW
Joint Utilities365$14.13/kW$14.06/kW$10.24/kW
Table 8 Adopted Monthly Grid Participation Charge for Successor Tariff Customers
Customer SegmentPG&ESDG&ESCE
Residential$8.00/kW$8.00 kW$8.00 kW
Low-Income$0/kW$0/kW$0/kW
Nonresidential$0/kW$0/kW$0/kW

Comment: The CPUC is relying on low system price data that is not reflective of the average to shore up their assumption that payback times will not increase. The commission’s data does reflect higher prices for system components that are currently driving system prices up. The commission’s analysis fails to consider the effect of tariffs and bans on modules shipped from Xinjiang and how a supply shortage could affect the future price of components. NEM 3.0 will arrive just as system prices increase and inflation increases prices for groceries and other necessities, thus complicating the system buying decision. 

Commissioners also see inequity in how grid operation charges are spread among ratepayers and seem to assume that adding grid charges to the bills of solar system owners will level the playing field and potentially result in lower electricity rates for all ratepayers. 

Bluntly, are the commissioners even slightly familiar with the utilities they oversee? Because even a casual observer would assume that rates will not go down, and in fact, utilities will soon be bellying up to the bar for another increase. 

Meanwhile, climate change worsens while the commissioners continue enabling the polluters. 

Bernadette Del Chiaro, Executive Director, California Solar & Storage Association (CALSSA), said: “This is, frankly, not a sound public policy proposal. It penalizes consumers who invest in clean energy, which is insanity as the country continues to be pummeled by ever-present climate disasters. It rubber stamps the investor-owned utility’s gutting of NEM, cutting export credits 80% overnight. And it makes retroactive changes to 1.3 million early adopters moving the timeline up five years. This proposal would devastate tens of thousands of jobs and hundreds of small businesses. It is now up to the other four commissioners and the governor to course correct.”

Barry Cinnamon, CEO of Cinnamon Energy Systems, said: “PG&E has dumped a ton of coal in the stockings of all California customers who do not yet have solar. Going solar under NEM 3 will more than double the payback time for all new solar customers. Moreover, by retroactively reducing the grandfathering period by 5 years for existing solar customers they take away 25% of the value of all existing solar systems.

Congratulations, PG&E! With one fell swoop you manage to maximize your profits, reduce benefits for existing solar customers, double the payback for new customers and make it impossible for California to meet its zero carbon emissions goals.

The two departing CPUC commissioners who oversaw this potential destruction of the largest solar market in the U.S. are punishing current and future solar customers, and dooming California’s climate change efforts.”

Lesson: What the government giveth, the government will find an excuse to taketh away.

Shipment Progress 1997, 2004, 2009, 2020

1997 was the first year shipments topped 100 MWp and manufacturers looked to 100-MWp of capacity to provide economies of scale

In 2004, the industry topped 1-GWp in shipments and manufacturers pointed to 1-GWp of capacity as the economies of scale benchmark

In the mid-2000s, manufacturers in Taiwan and Europe planned to become pure play cell manufacturers

In 2009, with shipments nearing 10-GWp annually, manufacturers focused on 10-GWp of capacity to provide economies. Pure play cell manufacturing fell out of favor.

After 2020 – capacities of >30-GWp are the focus and pure play cell manufacturing is back!

Economies of Scale – the good, the bad, and the underutilized 1987, 2007, 2017, & 2021 Estimate

Unused capacity is a cost to manufacturers.

Shipments 1989 through 1999, all Technologies

In the early days of the PV industry, shipments of standard monocrystalline technology dominated. In the 1980s into the early 1990s, amorphous silicon performed well, losing share when the consumer-indoor application (watches, calculators) declined, and it failed to prove less expensive to manufacture.

Crystalline Shipments 1989 through 2009

In 1998, multicrystalline, which was less expensive to manufacture, began gaining share in the market.

Multicrystalline took the major share of shipments in 2003. At the same time, shipments from China began taking share from manufacturers in Europe, Japan, and the US.

Crystalline Shipments 2009 through 2019

Though it seems to be an overnight shift from multi to monocrystalline, the change began much earlier. In 2012, manufacturers began adding p-type mono and multicrystalline PERC capacity, primarily focusing on p-type mono-PERC.

In 2018, manufacturers began shipping significant volumes of PERC, particularly monocrystalline.

A little Pricing History …

Since 2009 the accurate cost of PV cell and module manufacturing has been obscured by subsidies. Price is a market function and not a reliable indicator of manufacturing progress. Module prices are currently increasing.

Current Industry Cell Capacity

Current Module Assembly Capacity

Shipments 2011 to 2021 estimate

Once upon a time off-grid dominated

Off-grid (remote) solar deployment uses storage – in the past, lead acid batteries. Progress with storage has been slower than hoped as it was thought of as too expensive. Now, the future depends on it.

To truly become one of the industries of the future solar needs to:

1. Have a supply chain free of forced labor.

~45% of polysilicon is produced in Xinjiang

Metallurgical Silicon producer Hoshine serves clients outside of China

~60% of cobalt is manufactured in the Congo

2. Have a supply chain free of conventional energy.

Much of the world’s metallurgical silicon, polysilicon, wafers, cells, and modules are produced using coal as an energy source.

3. NOT provide electricity to coal and other dirty mining enterprises.

Solar installations are commonly used to provide electricity for mining – including coal

4. Have manufacturing in ALL countries with solar demand.

The industry provides jobs in science, engineering, manufacturing, and construction – jobs, jobs, jobs

5. Have a diversified workforce.

6. Be transparent about ALL the cost components of PV manufacturing – meaning ALL incentives, grants, loans, debt – everything from consumables to polysilicon to glass, aluminum, et al, so that the industry has a clear cost curve with which to track progress.

7. Accept that manufacturers need more than a 5%, 10%, 15% margin to run operations.

Forecast

The accelerated and conservative forecasts need a global manufacturing effort to come to pass and there must be sufficient capacity for all materials.

For storage to play a significant role, battery technologies need continued research effort – including into cobalt free battery technologies – and storage must play a significant role in a modernized grid.

“There are only two ways to live your life. One is as though nothing is a miracle. he other is as though everything is a miracle.” Albert Einstein

1997 was the first year that global shipments broke the 100-MWp mark, and the US had a 42% share. After breaking the 100-MWp mark, manufacturers began pointing to the economies of scale that 100-MWp manufacturing capacity would bring. Back in the day, US solar cell and module manufacturing was primarily supported by oil or energy companies such as Arco, Shell, BP, and even Enron. The Walton family (think Walmart) was an early investor in CdTe manufacturer First Solar. Then and now, the US government, via DoE grants and loans, provides support for startup solar manufacturing but little support for ongoing commercial solar cell and module assembly manufacturing.

 By 1999, driven by Japan’s strong residential rooftop incentive and government support,  manufacturing shifted to Japan. Then, in 2007, boosted by the FiT-driven market and government manufacturing incentives, manufacturers in Europe, primarily Germany, took the lead.

Accelerating demand for PV installations in Europe – including the new multi-megawatt utility-scale application – strained the supply of polysilicon available to the PV industry. As a result, prices for polysilicon, wafers, cells, and modules increased significantly, increasing costs and margins for manufacturers and buyers.

In 2009, having built sufficient manufacturing capacity, China’s manufacturers entered the market with an aggressive pricing strategy that drove prices down by 42% in one year. Manufacturers in other countries could not compete with the low margins China’s manufacturers were willing to accept, and so – one after another – they failed.

Manufacturers in China were able to take advantage of government loans, grants, free or inexpensive land, low energy costs, and low labor costs to dominate global shipments from 2011 to the present day.

Figure 1 offers average module prices from 2005 through an estimate for 2021. Figure 2 presents shipments from China and the rest of the world from 2007 through 2020.

Figure 2 presents shipments from China and the rest of the world from 2007 through 2020.

Why do PV manufacturers fail?

In a perfect world, countries with demand for solar installations would have sufficient manufacturing to meet that demand. There are no perfect worlds. Buyers seek less expensive products.  Meanwhile, manufacturing migrates to lower-cost areas with more manufacturing incentives. Governments, observing decreasing prices and assuming they represent progress, peg aid to further price decreases creating a landscape in which participants continue to chase goals that are, in reality, nothing but mirages.

Using the US as an example, dearly departed manufacturers include Siemens Solar, BP Solar, AstroPower, Evergreen Solar (ribbon), United Solar (a-Si), and the un-dearly departed Solyndra. Over the years, there were acquisitions (BP Solar and Solarex) and (ARCO, Siemens, and Shell Solar).

In 2004, eager to take advantage of the accelerating market for solar, GE acquired bankrupt AstroPower only to shut it down in 2009. Observing the success of CdTe manufacturer First Solar, GE acquired Colorado-based CdTe startup PrimeStar in 2011, announcing it would build a 400-MWp manufacturing facility. In 2013, having ramped nothing, GE sold PrimeStar’s technology to First Solar. China-based Hanergy acquired CIGS startups Global Solar and MiaSole.

Manufacturers fail for various reasons, so the best answer to – why do manufacturers fail? – is, it depends. However, solar manufacturers typically fail, and in some cases fail to launch, because of unrealistic expectations about cost and price and a lack of government patience and support. Companies that begin by chasing an unrealistic goal will never catch it.

The Prescription is Time and Money – Lots of Time, Lots of Money

As long as the metric for progress is based on an artificial cost curve, it will be difficult to build solar manufacturing capacity outside of China and Southeast Asia.

While offering demand-side incentives and a smattering of state-by-state manufacturing incentives, the US has chosen a stick approach, using tariffs on imports from China to protect, basically, no one. The US also, as indicated, provides grants and loans to startup manufacturing through the DoE. Currently, the country has one large thin-film manufacturer and several gigawatts of module assembly for crystalline. This year, Senator Jon Ossoff, Democrat, Georgia, introduced the Solar Energy Manufacturing Act and successfully negotiated to include it in the upcoming Congressional Budget Measure. It could pass as stands, be reduced in scope, or excluded from the final budget.

India has chosen a stick and carrot approach using high tariffs to discourage imports while introducing a production-linked incentive to encourage manufacturing to locate in the country. India will need to keep its incentive in place long enough to support its new manufacturing base against competitively priced imports.

Across Europe,  no stick and a smattering of manufacturing incentives has done little to encourage domestic supply. The EU might consider a production incentive similar to India that stays in place long enough to support a fledgling industry. A Buy-European incentive would help to encourage domestic buying.

Countries looking to start up photovoltaic manufacturing will need patience and the willingness to support the effort for years. Though carrot and stick approaches are common, manufacturers usually pass the tariffs on to buyers, increasing cost. Maybe it’s time for a two-carrot approach – incentives for buying domestically produced cells and modules along with manufacturing incentives.

The solar industry has gotten a bad rap for always having its hand out. The conventional energy industry has had its hand out for decades. To build a domestic supply chain, governments might consider offering solar manufacturers the same embedded incentives available to oil, coal, and natural gas, incentives embedded in the energy structure to the degree that they become invisible to the public over time.

In 2021 a combination of disruptions in coal production, higher-priced imports of coal, and low availability of hydroelectric due to drought have led the country’s generators to ration electricity and to impose blackouts to preserve the grid.

Two-thirds of China’s electricity comes from coal, the rest from a combination of natural gas and renewables. China is the world’s largest importer of natural gas. In October, flooding in Shanxi, the country’s largest coal-producing region at a quarter of the country’s supply, led to a suspension of operations at 60 mines and halted work at close to 400 others.

Businesses – including polysilicon, wafer, cell, and module manufacturers – are, by government order, reducing hours and acceding to consumption limits. As a result, solar polysilicon, wafer, cell, and module manufacturers are reducing hours and idling equipment while facing mounting orders that they may not be able to fill.

Meanwhile, similar to California’s energy crisis in the early 2000s, the cost of imported coal and natural gas has spiked, leaving generators caught between the government-regulated price to consumers and the market-determined price.

Buyers, already facing increased prices for components, higher shipping costs, shipping delays, and, in the US, cost increases and delays due to the WRO, can now expect more supply constraints. Further complicating supply chain concerns, a resurgence of COVID in the Southeast Asia manufacturing hubs of Vietnam, Malaysia, Indonesia, and Thailand, have led to production shutdowns.  

Comment: The world is still dealing with COVID realities, including manufacturing shutdowns and supply chain issues. As countries ramp activities, electricity demand is peaking just as efforts to ramp down the use of conventional energy collides with ramping up the use of renewables. As renewables are variable resources, storage is required to make the switch.

Even if storage were ready for a sustained rollout, global traditional transmission and distribution infrastructures are not resilient enough to support the change. Conservation and money are needed to move forward – governments need to be prepared and willing to pay.Lesson: In 1988, then-NASA scientist James Hansen testified before the US congress about the realities of climate change and what the world would face without acting. We didn’t act, and we’re facing the realities of extreme weather, droughts, higher prices, and scarcity.  Will government commit to the cost – history has not borne out a willingness to do so.

The simplest way to describe pilot-scale production is as an experiment, the point of which is to replicate a result and establish a consistent average.

Assumptions based on little data do not count. Until it is proven by methodically repeating the experiment, the guess remains a guess.

Traditionally, pilot-scale production has taken ten, shrinking to five years, to produce commercial technology. However, with accelerating demand, particularly for multi-megawatt installations, the timeline for establishing repeatability and assessing reliability has decreased, and reliance on rapid tests and the assumptions based on them has increased. Meanwhile, compressed margins have pushed manufacturers to streamline quality control measures, assuming rapid tests would pick up the slack.

Outsourcing, always a factor in the PV industry, has compounded problems, as module buyers cannot be sure whose cells are inside the module, no matter the brand, or, in some cases, know what manufacturer assembled the cells into the module.

As the time from pilot scale to commercial production has shortened, quality in the field has decreased, and failures have increased.
Pilot-scale production is a process undertaken to find a cell or module technology’s average performance in terms of reliability and efficiency. The process involves repeated testing to arrive at a stable average. As indicated earlier, pilot production takes two to five years for new technologies. New manufacturing facilities, even module assemblers, require a period of pilot-scale production. The addition of new equipment in an already functioning facility involves a period of pilot scale. Pilot-scale production includes tuning equipment and adjusting flow rates, adjusting consumable formulas (adhesives, for example),

For example, research and development into SunPower’s crystalline IBC (Interdigitated Back Contact) crystalline cell began in the 1970s at Stanford University. In 1975 research was published on IBC cells. In 1987 Ron Sinton of Sinton Instruments, winner of the 2014 Cherry Award, and a team at Stanford developed a 3 mm x 5 mm IBC cell with 28.3% conversion efficiency. This cell, which could not be soldered and was not stable, was a research step on the long innovative timeline from idea through commercialization.

During the early 2000s, US-based Applied Materials and Switzerland-based Oerlikon used their experience in manufacturing and flat-screen television production to develop multijunction, thin-film silicon-based solar panels (micromorph) manufacturing lines. Both companies assumed that turnkey manufacturing would shorten the time from pilot scale to commercial production. Neither company succeeded, and both abandoned their efforts.

Pilot-scale production is expensive and as margins have shrunk, so have pilot-scale timelines.

Pilot-scale production remains crucial to the future of the photovoltaic industry. Shortening the process and using engineering models and testing to estimate reliability and conversion efficiency leaves open the likelihood of poor quality modules and failures in the field.

In August, an anonymous coalition, American Solar Manufacturers Against Chinese Circumvention or A-SMACC, filed a petition asking for an investigation into manufacturers accused of avoiding antidumping and countervailing duties against China via expansions into Southeast Asia. The companies named in the petition are Jinko Solar (Malaysia and Vietnam), LONGi (Malaysia and Vietnam), JA Solar (Malaysia), Canadian Solar (Thailand and Vietnam), Trina Solar (Thailand and Vietnam), Talesun (Thailand), Astroenrgy (Thailand), Sunergy (Vietnam), Boviet Solar (Vietnam), and GCL (Vietnam). Wiley, a DC-based law firm, filed the petition. A-SMACC released the following statement:

“For too long, obvious circumvention of antidumping and countervailing duties on Chinese solar products has hobbled the US industry, eviscerated our supply chains, and put our clean energy future at risk. It is time for America to lead in this critical sector.

While Chinese companies now almost exclusively export to the United States from Southeast Asia, the vast majority of manufacturing, research and development, and capital investment remain in China. In cases like this, the law is clear; the duties on Chinese solar products should be extended to circumventing entities.” 

SEIA, the US solar industry’s lobbying organization, responded by stating that the proposed tariffs and resulting price increases would devastate US solar industry demand causing developers to pause plans and manufacturers in Southeast Asia to cease importing products.

On Thursday, September 30, the US Commerce Department will consider whether to investigate or dismiss the petition. At that point, the members of A-SMACC may be made public.  

Who is behind A-SMACC?

The question solar industry participants should be asking is – who has something to gain from the petition? The US has no crystalline cell manufacturing. First Solar (CdTe) is the country’s only major cell manufacturer. The US has ~6-GWp of module assemble capacity available for crystalline cells and imports cells primarily from Southeast Asia and South Korea.

If First Solar and module assemblers buying from South Korea (primarily Hanwha Q-Cells and LG) are behind the petition, there is no good reason they should keep this a secret – other than avoiding an unflattering spotlight and many questions concerning their agenda.

There is always an agenda behind actions such as the one before Commerce.  When (if) the petitions are made public, their agenda will be clear.

Are China’s manufacturers shipping through Southeast Asia to avoid tariffs?

A-SMACC claims that manufacturers in China are using Southeast Asia as a passthrough to avoid tariffs — meaning that some value is added to the product before shipping it to its destination as a new product. 

Manufacturers in Southeast Asia have 19%, 57.3-GWp, of global capacity to produce cells and 23%, or 82.3-GWp, of global capacity to produce modules. The manufacturers named in the petition have 39.6-GWp of cell capacity in Southeast Asia, leaving other manufacturers with 17.7-GWp of cell capacity in the region.

Manufacturers expand to other countries for a variety of reasons. The most important reason for expansion off-shore is cost – lower manufacturing costs. Typically, a combination of incentives including low tax rates, grants, loans, favorable leases, and low-cost inputs including energy and labor are traditional reasons to locate manufacturing in a country. Other reasons to expand in other countries include favorable labor laws, more benevolent laws in general, nearness to supply lines or markets, and, of course, avoiding tariffs. Decisions about locating manufacturing are always complex.

A-SMACC’s claim that China’s manufacturers are using Southeast Asia as a passthrough to avoid tariffs would likely be challenging to prove. 

Would Additional Tariffs have a Catastrophic Impact on the US Solar Market?  

The US has no crystalline cell manufacturing to protect, one major thin-film manufacturer, and about 6-GWp of module assembly capacity. The module assembly capacity relies on imported cells. Maxeon announced plans for cell manufacturing in the US– but announcements are not actions. It will take several years for the US to establish crystalline cell manufacturing and a combination of manufacturing and buyer incentives. Module assembly can be established faster but relies on imported cells.

The US is an import market. Meeting the country’s solar deployment goals requires imports.

If the Commerce Department investigates and rules in favor of the A-SMACC petition, it would have no choice under the law and would have to impose tariffs

Tariffs act as a tax on buyers. The proposed tariffs would hit US market participants already reeling from higher shipping costs and delays, higher costs for inputs, proposed changes to the ITC that add complexity and cost, and the WRO, which is currently stalling cells and modules at the ports. Additional costs would not shut the market down, but it probably would lead some developers to rethink plans.

Since 2015 average module prices reliably decreased – year, after year – convincing buyers, and the industry as a whole, that an increase not just unlikely but impossible.

From 2015 through 2020, average module prices decreased by a compound average of 9%. Figure 1 presents average module prices from 2015 through 2020.

Figure 1: Average Module Prices 2015 through 2020

Manufacturing in the solar industry has been low margin for decades, yet average prices either decreased or remained flat for over ten years. Manufacturers in China and Southeast Asia seemed to accept low margins, thus creating an uncompetitive landscape for manufacturers in other regions and countries.

In 2021, buyers find themselves faced with module price increases. The current pricing situation has some aspects in common with the mid-2000s. During these years, Germany’s Feed-in-Tariff law encouraged demand in the country to accelerate. Other countries in Europe rapidly followed suit, creating, almost overnight, a gigawatt level market for solar deployment.

The feed-in tariffs were intended to drive demand – and they worked even better than envisioned. Though proponents expected accelerated demand, sufficient thought was not given to how the speed of market acceleration would impact upstream participants such as polysilicon suppliers.

Previous to the FiT, the photovoltaic industry had primarily survived on scrap. The polysilicon industry’s largest customer was the semiconductor industry. Feed-in-tariffs marked the end of this paradigm and beginning in 2004, polysilicon prices spiked, rising in some cases to $400/kilogram on the spot market with contracts spiking almost overnight from $20 to $30/kilogram to >$60/kilogram. Wafer, cell, and module prices also increased with buyers captive to the increase. Polysilicon, ingot, wafer, and cell producers funded capacity growth via long-term contracts. New cell and module manufacturers sprang up almost overnight, each believing that prices would not decrease. Then, in 2009 PV manufacturers in China had sufficient capacity to enter and, using aggressive pricing strategies, rapidly drove manufacturers in other countries out of business.

Figure 2 presents average module prices from 2000 through 2010.

Figure 2: Average Module Prices 2000 through 2010

As with the earlier period, demand for solar is accelerating, driving a need for more polysilicon. Unfortunately, in 2020 accidents in several facilities in China took significant capacity off-line.  Repairs took time because of the pandemic and because, well, repairs take time.

Glass supplies were also impacted. Due to overcapacity, China’s government has controlled glass supply for years, allowing additional capacity, only replacement capacity. As the demand for bifacial modules accelerated, the available supply of solar glass was unable to keep up, and prices rose. Mid-2021, glass supply constraints have eased, but prices have stayed high.


China has also been experiencing a coal shortage, which has driven the price of energy up in the country. Much of China’s solar production is powered by coal, and the shortage has led to higher energy prices for manufactures from polysilicon downstream to modules.

Back to polysilicon, capacity additions take over a year. Though additional capacity is planned and being constructed, it will not come online rapidly nor – for the sake of quality – should it. And, as wafer and cell manufacturers are adding n-type capacity, a higher grade of polysilicon is required.

Also, 45% of China’s polysilicon capacity is in Xinjiang. The US has banned materials from Xinjiang and will see supply constraints. If other countries and regions take similar actions, the industry would see a down year in shipments and construction activity. Finally, China’s manufacturers may have discovered that healthy margins are a good thing and may be unwilling to offer premium products at low margins in the future.

Figure 3 offers average polysilicon costs and prices from 2018 through 2023, along with the current price.

Figure 3: Polysilicon Prices 2018 through 2023

Module pricing is likely to be volatile in 2021, with potential spikes and little relief for buyers. As prices for wood, steel, and aluminum are also high, there is high potential for construction delays, with some developers choosing to wait out the situation.

Buyers should be aware that prices may not decrease for some time – then again, all it takes is for one multi-gigawatt manufacturer to break ranks and drop prices. It’s a pricing game of chicken. Developers worldwide can only delay projects so long while manufacturers concentrated in China and Southeast Asia hold all of the supply cards.

Figure 4 presents average module prices.

Figure 4: Average Module Prices

The Biden Administration has big plans for fighting climate change in the US, aggressive infrastructure plans to combat it, and big solar manufacturing and deployment plans. Participants along the value chain are optimistic. Developers are gearing up. Residential and small commercial installers are confident. In California, mandates for solar on new residential buildings and soon for solar +storage on new commercial and multi-dwelling residential buildings offer a template for accelerating the move away from conventional energy. The US has the potential to be a 30-GWp market – but the real potential is much higher. The US could be an >50-GWp annual market for solar deployment.

Currently, the US has over 20-GWp of annual demand, 2-GWp of thin-film cell capacity, and an additional 5-GWp of module assembly capacity for imported crystalline cells – in other words, the US does not have the capacity to serve its market. The US solar market is fragile without sufficient domestic cell manufacturing, and participants have little control over module supply and price. A shock to the supply chain would likely stall – at least temporarily – market growth, potentially taking that >50-GWp of potential down to the low teens.

In late 2020 accidents in several polysilicon facilities in China took significant capacity off-line. Repairs took time because of the pandemic and because, well, repairs take time.

Due to overcapacity, China’s government has controlled glass supply for years, allowing additional capacity, only replacement capacity. As the demand for bifacial modules accelerated, the available supply of solar glass could not keep up, and prices rose.

Meanwhile, continuing into 2021, shipping costs increased, and a semiconductor shortage affected inverter and tracker manufacturers who were also experiencing rising costs – after years of absorbing margins – began passing higher costs on to customers.

Developers, used to years of price declines, at first, tried to wait it out.

Meanwhile, the situation in the Uyghur Autonomous Region of Xinjiang, China, finally caught the attention of politicians.

In anticipation of the US and other countries acting, on June 10, China’s central government passed a law forbidding Chinese companies from participating in audits of their materials.

Then, on June 24, Homeland Security ordered US Customs and Border Protection (CBP) to issue a Withhold Release Order (WRO) detaining in customs metallurgical silicon produced by Hoshine Silicon Industry Co., Ltd., and its subsidiaries for the use of forced labor in its manufacturing facilities.  Hoshine Silicon is the largest metallurgical silicon supplier globally. Its customers are polysilicon producers, including Germany-based Wacker, South Korea-based OCI, Daqo, GCL, Jiangsu Zhongneng, Asia Silicon, Xinjiang GCL, Xinte, and East Hope.

On July 14, the US Senate passed a bill banning goods from Xinjiang. The bill has gone to the House.

US Customs and Border Control has begun holding cells and modules at the border, delaying projects, and increasing costs and anxiety for developers and installers.

The US Trolley Car Problem

The US faces a moral and ethical dilemma. On the one hand, after years of endless discussions and delayed action on climate change, governments must act decisively and quickly. But, on the other hand, forced labor cannot be ignored or compromised away.

The English philosopher Philippa Foot developed the Trolly Problem in 1967. An out-of-control trolly car is barreling down the tracks towards five trapped people. There is no way to stop the trolly and save the five people, but if you throw a switch, the trolly will instead barrel down a track where only one person is trapped. So, the choice is five or one.

There is no bargaining with climate change. Unaddressed, it will continue barreling down the tracks.

There is no compromising with forced labor, forced sterilization, forced re-education, separation of parents from children, or labor camps designed to contain the Uyghurs.

Nor is there any denying that forced labor played a part in the low cell, module, and system prices that the solar industry has enjoyed for many years.

The political and timing is right for US solar industry growth to accelerate beyond anyone’s forecast. Behavior has changed, and there is pull from end-users. Solar has moved into the apolitical arena with proponents on the left and the right.

Unfortunately, as indicated, the US does not have sufficient cell manufacturing to meet its demand, and it will take years to build it. There is sufficient supply unrelated to forced labor in Xinjiang for the US, but it will be more expensive, and there will be periods of scarcity. Growth will come at a higher cost – but not a higher moral cost. Because, again, there is no compromising with forced labor.