Market for MW-scale PV O&M and asset management to surpass 133 GW worldwide by end-2015
A November 2015 GTM Research report expected the global fleet served by MW-scale PV O&M and asset management to pass 133 GW in 2015 and exceed 488 GW by 2020.
Does solar use too much land, never repay its energy, fail when clouds pass, or leave toxic waste? Common PV claims checked against 2016 evidence, with a later update.
By 2016 solar PV had moved from a niche technology to one of the fastest-growing sources of electricity, and with that growth came a steady stream of claims, both for and against it. Some were outdated, some were exaggerated and some pointed to real engineering problems. This explainer takes five of the most common claims and tests them against the evidence published around 2016.
This was one of the oldest objections, and by 2016 it was clearly out of date. The measure that answers it is the energy payback time (EPBT): how long a system must operate to generate the energy that was used to manufacture, transport and install it.
Fraunhofer ISE’s Photovoltaics Report of November 2016 summarised the evidence:
The improvement came largely from using less silicon: material use for silicon cells had fallen from around 16 g/Wp to less than 6 g/Wp over the previous decade, thanks to higher efficiencies and thinner wafers. Because EPBT depends mainly on how much sunlight a system receives, systems in sunnier locations repay their energy faster than the same systems in Northern Europe.
Verdict in 2016: false. PV systems repaid their manufacturing energy many times over.
Solar is a low power-density source, so this claim has a factual basis, but the scale is often overstated. The most detailed study available in 2016 was NREL’s 2013 survey of US solar plants (covering 72% of US utility-scale PV and concentrating solar capacity installed or under construction in late 2012). It found:
A worked conversion helps with local scale. Using 1 acre = 0.4047 ha:
Land for a 50 MWac plant ≈ 50 × 8.9 acres = 445 acres
≈ 445 × 0.4047 ≈ 180 hectares
That is a substantial site, but much PV needs no new land at all: rooftops, car parks and other structures host a large share of installations. Land at ground-mounted plants can also be shared with grazing or kept as grassland.
Verdict in 2016: partly true. Large solar farms need significant land, but less than often claimed, and rooftop PV uses none.
PV output follows the sun: none at night, reduced under cloud, and varying by season. This is a genuine engineering issue. The key measure is the capacity factor, the ratio of actual energy produced to what the system would produce running at full rated power all year.
Fraunhofer ISE’s figures for Germany give a worked example: about 38.4 TWh was generated by PV in 2015 from about 39.6 GWp installed.
Capacity factor = 38.4 TWh / (39.6 GW × 8,760 h)
= 38.4 / 346.9 ≈ 0.11, or about 11%
Yet PV still supplied about 7% of Germany’s electricity demand in 2015. Variability does not make solar useless; it means grids must plan for it, using accurate forecasting, flexible generation, interconnection, demand response and, increasingly, storage. The integration task grows as solar’s share of supply rises.
Verdict in 2016: the variability is real, but it is a grid-planning problem, not a reason to dismiss PV.
The first global projection of PV waste was published in June 2016 by IRENA and IEA PVPS. Its model ran two scenarios:
| Scenario | Waste by end-2016 | By 2030 | By 2050 (cumulative) |
|---|---|---|---|
| Regular loss | 43,500 t | 1.7 million t | about 60 million t |
| Early loss | 250,000 t | 8 million t | 78 million t |
The report noted that annual PV waste in 2014 was around a thousandth of the 41.8 million tonnes of global e-waste that year, though by 2050 annual PV waste could exceed 10% of that 2014 e-waste figure. Modules are mostly glass, aluminium and polymers, with smaller amounts of valuable and hazardous materials. The report estimated that technically recoverable materials could be worth up to USD 450 million by 2030 and more than USD 15 billion by 2050, and recommended PV-specific waste regulation, building on the European Union’s experience.
Verdict in 2016: a real but manageable future issue, best addressed by recycling rules before volumes peak.
By 2016 cost had fallen dramatically. Fraunhofer ISE reported that a typical 10–100 kWp rooftop system in Germany cost around €14,000 per kWp in 1990 but about €1,270 per kWp at the end of 2015, a net price fall of about 90% over 25 years. Its experience curve showed module prices falling about 23% with each doubling of cumulative production. Whether solar is cheaper than alternatives depends on local sunlight, financing and grid tariffs, but the claim that it was inherently too expensive no longer held.
The fair summary in 2016 was that solar PV was a mature, energy-positive technology whose real challenges lay in integration and end-of-life management, not in whether it worked. Quality also mattered: poorly made modules could undermine all of the above, as described in our restored article on module quality problems. The scale of deployment by then is shown in the 177 GW global milestone for 2014.
Later editions of Fraunhofer ISE’s report showed further improvement in energy payback. Its July 2021 release reported payback times of 1 to 1.3 years in Europe for rooftop systems using 19.9%-efficient silicon modules made in China, ranging worldwide from 0.44 years in India to 1.42 years in Canada. The questions about land, integration and recycling have shifted from whether to deploy solar to how to do it well at much larger scale. In Malaysia, that scale-up began with the large-scale solar and net metering programmes described in Malaysia to install 1,250 MW of solar PV by 2020.