• Steven Bowen

Banning Xinjiang Polysilicon Is The Right Thing To Do & Won’t Impact Solar Growth

Banning Xinjiang Polysilicon Is The Right Thing To Do & Won’t Impact Solar Growth


Polysilicon prices are on fire – for many reasons

PV grade polysilicon prices have seen a vertiginous increase in the last 12months & are up 350% since ’20 lows. Germany’s Wacker Chemie AG is the only EU polysilicon producer – it’s shares have followed, up 132%.  Expectations of the Biden administration banning Xinjiang poly because of the alleged use of forced labour in its production are one reason recently sited.  Xinjiang produces about 45% of global pv grade polysilicon.

Many other drivers of solar grade pv inflation are also evident; some are transitory like trade bottle-necks, but many are longer term in nature, such as the substantial increase in solar capacity necessary to fulfil society’s growing energy requirements while decarbonizing electricity production to keep us on the path to 2030 milestones and ultimately net zero 2050. By ’30, new solar capacity is expected to grow by 630GWpa (vs ~700MW total base today) and increase overall 20 fold by 2050 (IEA).

Plenty of commentary – don’t believe the hype

We’re never lacking commentary about pv solar. And inflation is incredibly topical, so it’s no surprise that journalists are going nuts writing about polysilicon inflation. Given the staggering size of the 30 year capex program called the energy transition (our numbers: ~$120Trillion, or roughly one quarter of all capex), plenty of ink gets used on the unbearable costs of renewables.

The problem is – a lot of the commentary about polysilicon price inflation impacting growth is just plain bunk.  In a recent opinion piece, David Fickling from Bloomberg worried that poly prices could endanger the future solar power capacity we need for greening the grid.


Poly is far from the only input

Polysilicon is a hugely important part of solar power.  Without it, grid-scale PV wouldn’t be the cheapest cost energy source on the planet.  From 2010 to 2020 module prices/kwh declined by >80% (NREL). The incredible gains in efficiency of turning the power of the sun into electricity have a great deal to do with the improvement in availability and quality of high grade polysilicon.

But thankfully, poly is only one of many inputs in a solar module & module are themselves only part of the cost of an installed, scale pv solar site.  Poly makes up ~15% of the cost of the module and in-turn modules are less 20% of total installed cost (BNEF, NREL).

All else being equal…

Now a tripling of silicon price does indeed have a nasty impact on overall installed price; +600bps, all else being equal. And analysts are right to highlight directional changes in input prices (moving from being a deflationary input to an inflationary one) can have big impacts in models discounting decades into the future.  But thankfully all else is not apt to remain equal.

Here it’s worth looking at the role played by the rest of the PV system in that big -80% price decline.  From 2010 to 2019, about half of the decline in installed price was driven by inputs other than modules including inverters, BoS and R&D expense(IRENA,NREL).  Within modules, MIT found production scale was equally as important as poly to price deflation after 2010 – that makes sense since that was roughly when the German industry was wiped off the map due to massive investments in production by the Chinese government.


Paradigm Shift: Cost to Value

Installed equipment cost is far from the only value driver of PV however.  How the asset is used: efficiency and yield, are just as important in making the investment decision.  And this is where things are rapidly improving – it’s where the next leg in lower electricity cost is apt to come from. Through the implementation of complimentary technologies like diurnal storage and smart grids, peak-shaving & time shifting can be used to address the key challenge of renewables, a relatively low utilisation rate due to intermittency.


Matthias Schroder unsplash

Beating the Duck Curve – the next step in increased solar efficiency

Solar & wind power are by definition intermittent.  Solar cells are only productive when the sun is shining, and wind is seldom constant.  Adding variable generation to the grid will cause more volatility in wholesale electricity prices and as intermittent resources start to make up a larger and larger part of energy generation, they can lead to cannibalization of revenue & curtailment when it’s sunny or the wind is blowing. This is known as the duck curve. At the same time, the grid requires back-up capacity for the times when it isn’t and that requires substantial capex and lowers returns for investment in additional renewable capacity.  It also means keeping GHG emitting gas & coal plants on-line and that requires continued investment in the entire hydrocarbon value chain, which is neither good for the planet nor an efficient use of energy investment capital in a net-zero framework.


Smart Grids and Storage are likely to drive future increases in value of solar

Space and time are the energy traders key constraints in getting value from energy generation. Smart grid/interconnection & improving energy storage options allow solar to provide maximum value by providing both.  Excess generation during sunny and lower consumption times of the day can be time-shifted & volatility in pricing can by shaved by either routing the excess production to a different time zone or storing it for higher usage times at the end of the day.  Both increase the value of variable generation assets by increasing utilization periods and improving average pricing & they lower the grid investment requirements for back-up capacity from fossil fuel plants.

Lower battery costs and increasing efficiency are key reasons that battery storage can support a continued decline in levelized cost of energy for solar and wind.  Utility scale lithium-ion battery costs have fallen by ~87% in the last decade (BNEF).  IEA estimates are for the equivalent construction of 20 gigafactories pa for the next decade & global capacity rising from 160gigawatthours today to 6,600 by 2030 & they see energy density roughly doubling during that period.  By 2030, the IEA assumes combined renewables/storage installations will be commonplace globally & it’s worth noting in the most recent German renewable auction all 18 winning bids included solar + storage.

Xinjian Poly shoudn’t be bot for ethical reasons – don’t let anyone tell you the economics mean we have to

By-passing forced labour polysilicon will lead to higher costs for solar modules, but the net impact on installed costs will be manageable.  Far more important for solar growth is the increased value that can be had from what’s already the lowest cost way to electrification through new technologies in the grid.

Baptiste Riffard from unsplashed