The “Real” Birth of the Solar Industry

...and the beginning of Solar 2.0.

January 29, 2013

Brad Matt­son is CEO of Solex­ant, a DBL port­fo­lio solar startup.  He is also a renowned entre­pre­neur from the semi­con­duc­tor indus­try. Start­ing at Applied Mate­ri­als, he later founded and scaled both Nov­el­lus and Matt­son Tech­nol­ogy, and to date is the only founder of two pub­licly traded semi­con­duc­tor equip­ment firms. Most recently, Mr. Matt­son was a Part­ner at Van­tage Point Ven­ture Part­ners work­ing in the clean­tech area, lead­ing the solar sec­tor, but also focus­ing on light­ing, trans­port and scal­ing. Mr. Matt­son holds twelve patents in semi­con­duc­tor tech­nol­ogy, and in 2005 he was listed as one of the top 50 most influ­en­tial peo­ple in the semi­con­duc­tor industry.

He’s penned a piece about what he calls “Solar 2.0,” or the “real” birth of the solar indus­try, which ran today on Green­tech Media’s Green­tech­So­lar page, and can be read (along with some great data visu­al­iza­tions) here.   The arti­cle is reprinted below.

As odd as it may sound, this is an excit­ing time to be in the Solar Indus­try.  It is an indus­try poised for mas­sive growth. Unfor­tu­nately, the doom and gloom sur­round­ing events like Solyn­dra and other fail­ures are mask­ing an epic pos­i­tive change in the energy indus­try.  The very cost and price reduc­tions that are caus­ing all the angst are also dri­ving up demand.  As we reach grid par­ity in an ever-​​increasing seg­ment of the global mar­ket, we are wit­ness­ing the true eco­nom­i­cally dri­ven birth of a global Solar Industry.

As we enter this Brave New world of what I refer to as Solar 2.0, there are sig­nif­i­cant oppor­tu­ni­ties where the U.S. has real com­pet­i­tive advan­tages and can suc­ceed.  This arti­cle will focus on the first part of what I hope will become an on-​​going dia­log on the topic of a “U.S. Roadmap for Suc­cess” in solar.  In this first part, I will sug­gest that we have passed the tip­ping point and solar is becom­ing the renew­able energy of choice globally.

First, let’s look at the cur­rent sit­u­a­tion.  Much of the cur­rent neg­a­tiv­ity is due to the impact of mas­sive over­ca­pac­ity on the mar­ket, but this over­ca­pac­ity is actu­ally a nor­mal part of grow­ing pains.  Sure, these grow­ing pains feel par­tic­u­larly ago­niz­ing, but they are not unprece­dented.  For many who were involved in the semi­con­duc­tor indus­try, you have seen this cycle before….in fact many times.   Shown below is one rep­re­sen­ta­tion of the Semi­con­duc­tor Mem­ory busi­ness cycle.  I per­son­ally went through at least 5 of these cycles.

The cycle can start when­ever there is over-​​enthusiasm for a mar­ket (Robust Mar­ket).  The over-​​enthusiasm can exist for many rea­sons, but my obser­va­tion is that it is par­tic­u­larly pro­nounced when a tech­nol­ogy or busi­ness area is deemed strate­gic by indus­try, by gov­ern­ments or both.  The strate­gic nature dri­ves capac­ity addi­tions (Cap­i­tal Equip­ment Expan­sion) over and above what might be seen as eco­nom­i­cally nec­es­sary.  As new entrants come in and exist­ing com­pa­nies expand in order to main­tain or increase mar­ket share, excess capac­ity is inevitable (Excess Capac­ity).  This over-​​capacity dri­ves prices down (Price Soft­en­ing) as com­pa­nies try to reduce inven­tory.  Com­pa­nies can­not main­tain prof­itabil­ity, and cash flow prob­lems slow down or halt the expan­sion (Weak Mar­ket).   Man­u­fac­tur­ers stop buy­ing equip­ment and some even go out of busi­ness (Cap­i­tal Equip­ment Con­trac­tion).  As a result, the sup­ply slows down, stops, or even reverses as capac­ity comes off-​​line (Lit­tle Capac­ity Added).

Now, in par­al­lel with these dynam­ics on the sup­ply side, the “Price Soft­en­ing” three steps back also has an impact on demand.  As with any com­mod­ity mar­ket, lower prices increase demand.  This is cer­tainly true in the energy indus­try.   As demand increases with sup­ply soft­en­ing, we begin to see prices firm up (Price Firm­ing and Sta­bil­ity).  As man­u­fac­tur­ers see prices sta­bi­liz­ing, prof­its come back and the mar­ket looks good again.  The greed over­comes the fear and com­pa­nies once again look at capac­ity expan­sion in order to gain mar­ket share.  Thus, the cycle starts over.

It appears to me that the cur­rent panic in solar might be ampli­fied some­what because this is our first real cycle, and peo­ple that haven’t expe­ri­enced this feel like “the sky is falling”.  As an indus­try, we do not yet under­stand that these are just nor­mal grow­ing pains.  We could, of course, learn from the past and not repeat these painful cycles, but that is unlikely.  Solar is a strate­gic indus­try, and large multi-​​national cor­po­ra­tions as well as entire gov­ern­ments get into the mix with poli­cies and incen­tives that drive what seems like irra­tional behav­ior.  That is still hap­pen­ing in China, so we are not through the bot­tom of the cycle yet.  Some capac­ity is com­ing offline and prices have firmed a lit­tle, but there is still sig­nif­i­cant excess capac­ity so we have another year or so before we get back to “Robust Mar­ket”.  The demand is there, but not yet enough, and the prof­itabil­ity needed to enter the next capac­ity expan­sion cycle doesn’t yet exist.

Still…do not despair!  It is this cycle, with its con­sis­tent and relent­less price decline that is now dri­ving the “real” birth of the solar indus­try.  Let’s look more closely at cost, as it is cost (and result­ing prices) that dri­ves indus­try growth.  You have prob­a­bly heard of the con­cepts of learn­ing curve and tip­ping point.  Let’s look at how each of those relates to costs and how that dri­ves the solar industry.

The chart below shows the solar price reduc­tion over time in the tra­di­tional “learn­ing curve” for­mat.  In this chart, mod­ules ASPs (aver­age sell­ing prices) came down from ~$26/​watt in 1980 to ~ 2.60/watt in 2009.  This is an amaz­ing 10X reduc­tion in about 30 years.  Break­ing it down to smaller incre­ments, it is a 16.6% reduc­tion for each 2X increase in cumu­la­tive volume.

But this is only the begin­ning of the story.  You can see in the chart the “hump” from 2005 to 2009.  This was a one-​​time indus­try sit­u­a­tion based on poly­sil­i­con pric­ing.  This tran­sient in poly pric­ing is a story all in itself, but suf­fice to say it was only a tem­po­rary diver­sion from the cost reduc­tion trend.  Once the poly short­age was rec­ti­fied, panel costs resumed their down­ward plunge, only this time things were different.

The dif­fer­ence was China.  China had been dab­bling in solar for a while, but by 2005 they were hav­ing a huge impact.  Much has been writ­ten on this, so no need to bela­bor it here, but China decided that solar was a strate­gic resource and embarked on an ambi­tious pro­gram push­ing pro­duc­tion capac­ity at sev­eral lev­els in the food chain. Even with sig­nif­i­cant over­ca­pac­ity in place in 2009, Chi­nese PV com­pa­nies con­tin­ued to add capac­ity in 2010 and 2011.  The resul­tant excess capac­ity has dri­ven prices below costs as sup­pli­ers try to clear out inven­tory to gen­er­ate enough cash to keep afloat.  The crazy price decline of Tier 1 mod­ule sup­pli­ers that started a few years ago is shown below.  Of course the decline didn’t stop, but con­tin­ued in 2010, 2011, and 2012.  Mod­ule prices may have sta­bi­lized some­what recently, but are well below $0.70/watt even for top tier suppliers.

Because solar is a com­mod­ity, as China (the low­est cost pro­ducer) goes, the whole world goes.   Over the last four years, the price of solar pan­els has dropped an amaz­ing 85%.  Not ½, not ¼, but almost 1/​8th of the price.  We essen­tially fell off the learn­ing curve.  This abrupt change in the curve was pointed out recently quite nicely by Dick Swan­son, founder of Sun­Power.  There are many ram­i­fi­ca­tions to this “once in an indus­try life-​​cycle” occur­rence, includ­ing the indus­try tip­ping point, resul­tant growth rate, com­pet­i­tive shifts, indus­try con­sol­i­da­tion, ver­ti­cal inte­gra­tion, eco­nomic and trade pol­icy, viable future suc­cess sce­nar­ios, and maybe even the com­pet­i­tive­ness of the U.S. econ­omy !  All of these top­ics are impor­tant and inter­est­ing, but the bal­ance of this arti­cle will focus pri­mar­ily on the tip­ping point and its impact on growth.

There is a fine book by Mal­colm Glad­well, “The Tip­ping Point: How Lit­tle Things Can Make a Big Dif­fer­ence”, that intro­duces the con­cept of tip­ping point.  We can only get into a some­what lim­ited analy­sis in this forum, but the short story is this writer believes we have passed the tip­ping point in solar.

It is not so much that solar is already eco­nom­i­cally com­pet­i­tive with tra­di­tional sources of energy, although it is in many, many cases.  It is more that the momen­tum has shifted so far and so fast, that it is unlikely that any other renew­able energy source will be able to catch up.  To put a sharper focus on the issues just look at the pos­si­ble sce­nar­ios for 30–50 years from now:

1.  There is lit­tle chance we can con­tinue to use fos­sil fuel to sat­isfy our grow­ing energy needs.

2.  Hydro is great, but will be lim­ited by geography.

3.  Wind is great, but blows at night when we don’t need it, and grid stor­age is really tough.

4.  Nuclear is…..well even if it makes sense no one wants it in their back yard.  You can’t site it.

5.  Solar is clean, abun­dant, works at any scale, is eas­ily dis­trib­uted, and now it is cheap.

I’m sure the quick sum­mary above does not do jus­tice to the world of renew­able energy, but from the 30,000 ft. level, it looks like solar is really ready to take off.

Let’s look in more detail at that last state­ment….”now it is cheap”.  It is the price of solar and its com­par­i­son to the exist­ing cost of elec­tric­ity that dri­ves indus­try growth. This is the con­cept of “grid par­ity” where:  cost of solar elec­tric­ity  =  cost of grid electricity

Attached is the sum­mary of a study done by NREL to deter­mine when and where we hit grid par­ity in the United States.  It should be noted that this chart is using the retail price of elec­tric­ity.  That is, the prices that you or I pay at the meter at our home or the price busi­nesses pay at their fac­to­ries.  This retail price, of course, is much higher than the cost to gen­er­ate elec­tric­ity.  For exam­ple, in Monte Sereno, CA  I pay $0.28/kwh for elec­tric­ity.  This elec­tric­ity prob­a­bly only costs $0.08–0.10/kwh to gen­er­ate.  Of course, the elec­tric­ity gen­er­a­tors are typ­i­cally many miles from my home, so there is also the cost of trans­mis­sion and dis­tri­b­u­tion… and of course PG&E has to make a profit.

Some would say we have not reached grid par­ity until we hit the whole­sale, not the retail, cost of elec­tric­ity.  I dis­agree.  It just depends on who the cus­tomer is.  As a pay­ing cus­tomer for grid elec­tric­ity (from PG&E), I can tell you that we are at grid par­ity at my house.  I am a believer that the future of U.S. energy inde­pen­dence will come through more dis­trib­uted (solar on roof tops) gen­er­a­tion.  That bat­tle between dis­trib­uted and cen­tral­ized gen­er­a­tion is yet another impor­tant topic that we don’t have time to cover here.

So, with the caveat that we are talk­ing about retail, we can look at the NREL data.  The fig­ure on the left shows the sta­tus of the United States in 2008.  The dark red areas are where we are already at retail grid par­ity.  You can see it is a fairly small por­tion of the map on the left.  But, this level of grid par­ity is based on an installed cost of solar of about $8.00/watt, or roughly the cost in 2008.

The chart on the right shows the antic­i­pated areas of grid par­ity if the install price of solar is low­ered to $3.50/watt.  This is an amaz­ing increase in cov­er­age.  So…when will we get to $3.50/watt?  The answer is we are already there in some states and will be there in many oth­ers soon.

Let’s try to trans­late that into GW demand.  Some great work has been done by Richard Keiser at Keiser Ana­lyt­ics. He went through each state, coun­try and even area code in the U.S. to deter­mine cost and demand for elec­tric­ity at each price point.  There are of course some sim­pli­fy­ing assump­tions that need to be made to trans­late installed cost into the LCOE (lev­elized cost of elec­tric­ity), but with­out get­ting into com­pli­cated LCOE dis­cus­sions, it seems that rea­son­able assump­tions have been made.  Fig­ures are given with and with­out the gov­ern­ment ITC sub­sidy (Invest­ment tax credit).  Here are the esti­mates made of demand at each level of installed cost:


  Installed                With                   Without

  Cost ($/​w)           ITC (GW)             ITC (GW)

$5.50                        3 0

$5.00                        5 1

$4.50                       20 1

$4.00                       33 2

$3.50                       99 4

$3.00                      308 20

$2.50                      493 53

As you can see, at the $3.50/watt install cost, where much of the map above goes red, there is almost 100GW of demand in the U.S.  To put that in per­spec­tive the U.S had a record break­ing year in 2012 for solar instal­la­tions with ~ 3GW installed.  Obvi­ously, we have just scratched the sur­face.  Of course this depends on main­tain­ing the invest­ment tax credit.  But even with no gov­ern­ment sub­sidy at all, we just need to get to $2.50/watt installed cost to see more than 50GW in demand.

Now, just to clar­ify what is pos­si­ble, let’s look at Ger­many.  In Ger­many dur­ing 2011 the aver­age installed cost was only $2.25/watt.  They passed $3.50/watt a long time ago and passed $2.50/watt as well.  It is true that Ger­many has about 10X more solar installed than the U.S. and is there­fore fur­ther down the learn­ing curve, but can we be that much fur­ther behind?  We have access to the same pan­els, invert­ers, and all other com­po­nents needed to install a sys­tem, so we should have the same mate­ri­als costs.  Germany’s labor rates are actu­ally higher than in the U.S. so we should have equal or lower labor costs.  So, what is the dif­fer­ence ?  Why aren’t we below $2.50/watt already?

To be hon­est, when it comes to solar, Ger­many actu­ally has a coop­er­a­tive gov­ern­ment that has a more intel­li­gent, ratio­nal energy pol­icy than the U.S.  And I am not just talk­ing about the FIT (Feed –in-​​Tariff) that stim­u­lates demand, although that is rel­e­vant.  The FIT affects demand, not costs.  I am talk­ing about the stream­lin­ing of the bureau­cracy that allows solar per­mit­ting to be done in days or hours, as opposed to weeks or months.  The dif­fer­ence in these “soft costs” between Ger­many and the United States is ridicu­lous, and it accounts for almost all of the infla­tion of our installed cost.

Call me crazy for believ­ing in our gov­ern­ment, but I truly believe we will fix this paper­work prob­lem.   You have to hand it to Steven Chu, our Sec­re­tary of Energy.  He is a Nobel Prize win­ning, world-​​class sci­en­tist, and even he saw that the soft costs are a killer.  Through the DOE, an agency more known for excit­ing tech­ni­cal pro­grams, he insti­tuted a pro­gram to reduce soft costs.  Bravo Steve!

We have yet to see the ben­e­fits of these soft cost reduc­tions, but the fol­low­ing data from GTM research sup­ports that we are def­i­nitely track­ing towards Germany’s costs.  While the aver­age res­i­den­tial cost in the U.S. is still shock­ingly high (almost 2X Ger­many), this data shows it is drop­ping fast, about 5–10% per quar­ter recently.  This trans­lates into about a 30% yearly reduc­tion.  That would achieve the $3.50 price on the NREL chart some­time next year.   I expect we might see it even sooner.

The suc­cess of the SolarCity IPO is high­light­ing this “down­stream” part of the solar value chain, and we are see­ing bil­lions of dol­lars going into the instal­la­tion, or “solar power plant”, side of the busi­ness.   Look no fur­ther than War­ren Buf­fett, con­sid­ered by many the wis­est investor of our time.  He has just invested $5B in solar power plants.  With this much money com­ing into the down­stream part of the busi­ness, it is inevitable that com­pe­ti­tion increases just like it did in the upstream (panel) seg­ment years ago.  There are few bar­ri­ers to entry in the instal­la­tion busi­ness, so with lit­tle dif­fer­en­ti­a­tion there will be an intense focus to drive the costs out of every part of the instal­la­tion process includ­ing soft costs.

So, with a com­bi­na­tion of focus from the DOE, War­ren Buf­fet, and IPOs like Solar City, the soft costs will come down.  The U.S. should not only get to the $3.50/watt instal­la­tion cost and there­fore “cre­ate” the red map on the right soon, but we should be able to catch up with Ger­many and get to $2.50/watt which would drive over 50GW of demand even with no gov­ern­ment incentives!

Pulling this all together, when review­ing the solar indus­try, sure we see tremen­dous over­ca­pac­ity and cor­po­rate fail­ures left and right.  If we look at this with some per­spec­tive, though, we see that the cur­rent events are part of a cycle.  The same over­ca­pac­ity that cre­ates a sense of “doom and gloom” or makes us feel that the “sky is falling” is also dri­ving down prices at an his­toric rate.  In fact, the drop is so fast that it looks like we are falling off of the learn­ing curve.  Sure, there is still some cost reduc­tion to be done in the down­stream part of the busi­ness, but that is hap­pen­ing now.  Over­all, there is a strong feel­ing of momen­tum.  There is so much money, peo­ple, sup­port­ive data, pol­icy, even entire coun­tries behind solar, that I feel it has become a fait accom­pli.  We have passed the tip­ping point and we are wit­ness­ing the “true” birth of the Solar Indus­try.  These are very excit­ing times.

In the com­ing months, I hope to engage in a dia­log with those who might share some of this opti­mism to dis­cuss a roadmap for that suc­cess, and how the U.S. has real com­pet­i­tive advan­tages and can, must and will suc­ceed in this Brave New World of Solar 2.0.