by Phil LoPiccolo, Editor-in-Chief, Solid-State Technology
The cost to produce electricity with solar photovoltaics (PV) technology could reach grid parity with traditional utility electricity generation in five to eight years, after which the industry would take off on its own merits without the need for incentives — and that’s a conservative estimate, according to Stephen O’Rourke, Managing Director, Deutsche Bank Securities, speaking at a SEMI New England “Wafers to Wall Street” Breakfast Forum (Sept. 19 near Boston).
To understand the viability of the PV industry, it’s instructive to take a long-term view, O’Rourke said — “not the typical one- to two-quarter time horizon, but more like a 10- to 20-year perspective.” Also, rather than looking only at the solar PV industry, a close cousin of the semiconductor industry, he added, one must also look at the overall energy industry and, in particular, the electricity industry.
The key factor, O’Rourke noted, is that the US solar PV electricity-generation cost curve for known solar PV cell technologies is expected to drop, from what is roughly 25¢-30¢/kWh today to about 10¢/kWh by 2020, transitioning through what is commonly considered grid price parity (~15¢/kWh) within five to eight years. Conversely, the average retail price of conventionally generated US electricity will rise, from 8.6¢/kWh in 2006 to potentially more than 20¢/kWh by 2020, depending on cost CAGR projections ranging from 4%-7% (see figure, above). At these rates, cost “convergence” will begin within the next five to eight years, he predicted.
And this may be a conservative forecast, given that it is based on expected improvements in technologies that either exist today or will go into production in the next few years, O’Rourke explained. These include both average and high-efficiency crystalline silicon (c-Si), which currently dominates the market; thin-film copper-indium-gallium-selenium (CIGS), which will likely be produced in the 2008-2009 time frame; several varieties of thin-film amorphous silicon (α-Si), including a novel dual-junction variant that could go into production later this year; and thin-film cadium-telluride (CdTe) on glass, which may have the best overall cost profile.
What is unaccounted for in the convergence curves is the emergence of breakthrough technologies. “When you look at potential game-changing technologies coming down the pike in the next five to seven years by several companies, mostly in the US — including next-generation materials, quantum materials, multiple excitation generation, and the re-emergence of concentrator photovoltaics — the solar PV cost-reduction curve can be pulled down even further, which means grid parity will be achieved sooner and, in turn, demand will take off.”
The outlook from Wall Street
All the companies playing in the solar PV industry — from semiconductor to PV manufacturers and from materials to equipment suppliers — are positioning themselves for the explosive growth phase, which starts in five to seven years, O’Rourke said. Thus, we’re marching toward a temporary oversupply first and then rapid growth afterwards.
The PV industry’s initial growth phase (2003-2008) has seen huge growth in financing activity, with many companies vying to go public with competing technologies, and stock performance in aggregate has been strong (see figure, below). But from 2009-2012 there could well be a period of shakeout and consolidation when all stocks will tumble, even the best-positioned ones, O’Rourke predicted. “Some companies will not exist when we come out the other side, and while we can identify some long-term winners today, some that will lead this industry down the road, are probably unknown today.”
In the early throes of this shakeout (2008-2009) the corporate financing boom could drop sharply, but from 2009-2012 finance activity will rebound, partly from merger and acquisition activity, O’Rourke said. After that, grid parity will be achieved, and the industry will take off on its own economic merits, without the need for incentives. “When that happens, we will see an enormous build-out of capacity, even greater corporate financing activity, and, with the departure of many weaker companies, stock profiles will improve. Eventually, perhaps beyond 2021, he estimated, we will witness the cyclical growth of a more mature industry.
Everything about the PV industry so far has been about selling modules and systems. But in the long term, it’s not about chasing a module or systems industry, but rather competing with an electricity industry that is currently $300 billion domestically and approaching $1 trillion globally, O’Rourke contended. With that in mind, he said, the key is to think about selling electricity as a distributed utility and competing on retail cost.
A distributed model makes sense because it’s not practical for centralized utilities to invest in solar PV today. The model they use is central generation of electricity, and the marginal cost to add, say, new coal-fired plant production is 1-2¢/kWh, plus the utilities add about 3-8¢ in distribution costs, so that retail costs on average are <10¢/kWh. Therefore, solar PV, even at 15¢/kWh, doesn't look in any way as attractive as a central-generation product for electricity, O'Rourke said. However, as a distributed electricity-generation source that will reach convergence with traditional energy generation -- and then cost less -- solar PV makes a lot of sense for competing on a retail cost basis with the utilities, he argued.
O’Rourke noted that many ask about the technical limitations of the PV industry