Updated: Thirty-Year Solar PV Forecast

Since writing my original long-term PV forecast, I’ve been doing additional research into PV futures and technology trends, along with reviewing analysis from both large mfrs-utilities and smaller advocacy groups. Allow me to update my forecast:

By 2042, small photovoltaic (PV) electricity systems
will cost less than $0.90/watt, installed and commissioned,
which is equivalent to $0.02/kwh to $0.05/kwh,
depending on geographic location.

All costs expressed in 2012 dollars.

All per-watt prices expressed as DC nameplate rating of panel or inverter.

All “kwh” expressed as AC at the meter, after 0.73 derating from DC nameplate rating,
shown as an annual average from a fixed-mount pointing south.

Assuming 25-year warranted system life with 20% end-of-life PV panel derating.
There is evidence that end-of-life performance will improve with newer generation
panels, but this is not factored into my forecast. Moreover, most “25 year” systems
will remain in operation much longer
, making PV less expensive than this forecast.

Over the last 30 years, solar cell technology has halved its price-per-watt roughly every 7-8 years. Today, more capital than ever is being poured into PV research on all levels – academia, industry, government. Because of this and other strong market forces and indicators, it seems a reasonable assumption that PV technology will more or less follow its historical 30 year price trajectory into the next 30 years. If anything, affordability should accelerate. To keep a conservative forecast, I am assuming that PV panel price will halve every 12 years (not the historical 7 years) for the next 30 years, until reaching  a minimum cost-to-produce+profit of silicon, aluminum, glass. This gives us commodity solar panels at roughly $0.20/watt by 2042.

Today, DC-AC central inversion is roughly 50 cents-per-watt. The preferred DC-AC micro inversion technology is around 80 cents-per-watt with 25 year warranty. Texas Instruments (et al) are moving towards total integration of micro-inversion components, leaving just a few externals. And those externals (filters, flybacks, caps, switching transistors, etc.) are being life-optimized and scaled. As more major players move into the micro-inv business (China, Korea, India, etc.), look for a continuous drop in price per watt. Being very familiar with power supply design and high efficiency manufacturing, I have a high confidence that integration and scaling will push installed micro inversion to under $0.20/watt by 2042 (if not far sooner, perhaps 2025-2030). Energy Secretary Steven Chu is forecasting central inversion at $0.10/watt.

Commodity crystalline PV efficiency is around 15-16% today. PV research labs (Sharp, First Solar, etc.) have proven manufacturable efficiencies beyond 40%, but lab results have historically migrated slowly into real-world manufacturing. Historical 30-year PV efficiency has improved roughly 3% per year (1.03x per yr). To remain conservative, let’s assume that PV efficiencies will improve at just 2% per year over the next 30 years (keep in mind that, when looking at total PV system costs, “efficiency” is not nearly as important as cost-per-watt).

A 2% per-year PV efficiency improvement gives us 30% efficient panels by 2042, effectively halving the number of physical panels and mounting structures required per site — lowering freight, installation, and job management costs to roughly $0.50/watt. For new construction and re-roof jobs, new generations of “integrated thin-film solar roofing” will bring combined costs down even further, while solar permit processes become more streamlined (my solar permit app was about 12 pages plus formally drawn plan set – Germany’s solar permit app is 2 pages + simple written description). Innovative PV leasing and financing will continue to grow, allowing solar power installation with zero customer investment. And if a business or homeowner self-contracts (far simpler in 2030-2040), installation cost drops even more.

By 2042, this conservative forecast gives us:

Panels: $0.20/watt
Micro-Inversion: $0.20/watt
Install, Mgmt, Profit: $0.50/watt

Installed grid-tied PV systems at $0.90/watt puts us around $0.02/kwh in high sun areas (Las Vegas-Phoenix) and $0.05/kwh in weak solar areas (Seattle-Fairbanks) — all without rebates or other incentives. By 2042, the median U.S. cost to generate PV electricity becomes $0.035/kwh, which is roughly 1/3 the median U.S. rate of utility-delivered power. My research shows that U.S. coal-gas electric generation plants must sell electricity to the grid at roughly $0.06/kwh to remain profitable. The slope of this forecast suggests that median U.S. PV electricity may become cheaper to generate than median fossil electricity starting around 2030.

Our remaining “night time and cloudy day” electricity will be increasingly on-grid from other renewable sources, such as hydro, wind, geothermal, marine, biomass, compressed air batteries, “common dirt” batteries, nano-batteries, vanadium redox batteries, vortex engines, infrared nano-thermal, atmospheric differential, 24-hour PV hybrids, hydrogen, solar funnels, efficient catalyzers, and so forth. And I’m confident that surprisingly new and disruptive energy generation and storage technologies will enter the picture over the next 30 years – technologies we cannot yet imagine!

This historic and dramatic disconnect of electrical energy from fossils will change significantly the “power-utility company” model we know today. By 2042, on-site PV will no longer be receiving incentives, but will be paying a grid-use tax (grid maintenance tariff). The “grid dynamics” of renewable energy will require significant structural changes to the grid itself — no small challenge as power generation becomes increasingly distributed.

By 2050-2060, the widespread economic shift to PV (et al) will likely leave fossil-nuke plants producing less than 40% of U.S. electricity (down from 85% today). And by 2070-2080, most of the industrialized world will be generating well over 80% of its electricity from renewable sources. Germany is now predicting 100% renewable by 2050, and California is mandating 33% renewable electricity by 2020 (realistically, I think California will achieve around 25% by 2020). In 2012, renewable energy sources (i.e., biomass, geothermal, solar, water, wind) accounted for 46% of new electrical generating capacity installed in the USA. Very good news!

Simplified graphs of my forecast:

Now the bad news. Cheap, clean, renewable electricity does NOT mean we will solve our energy problem by 2040-2060 (For reasons too lengthy for this post, I’m calling 2040-2060 “the hump”). The majority of world transportation is expected to run on fossil oil through 2040-2050. Fully 1/3 of U.S. energy demand is sourced from oil. One forecast puts 2032 as the year when hybrids+EVs outsell pure internal combustion automobiles, but globally there will still be over one-billion petrol-based vehicles and apps (cars, trucks, engines, aircraft, mfg stock, military, agriculture, derivative, etc.).

Fossil oil demand has more or less peaked in the Western world. Exxon predicts that U.S. oil demand will drop 15% (20M b/d to 17M b/d) by 2040. On the other hand, the non-OECD industrializing world (India, China, etc.) is predicted to nearly double its oil demand over the next three decades, requiring an increase of worldwide oil production from 90M b/d to nearly 120M b/d by 2040. (It’s been noted that increased U.S. immigration could also spark a new population boom, re-kindling U.S. oil demand into 2040.)

Many energy researchers (myself included) are not convinced that world oil production can (affordably) supply 120M b/d. As world oil demand ramps up, it’s not clear that emerging economies can support healthy economic growth while paying $150-200/bbl for oil (in 2012 dollars). Global industrialization occurred thanks to cheap, abundant oil ($20/bbl throughout most of the 20th century). Increasingly expensive oil will likely lead to growing economic stagnation and increasingly crippling global boom-bust cycles.

A wild-card in our energy future is natural gas. Gas provides 1/4 of total U.S. energy demand. Fracking (etc.) is opening up vast new U.S. gas resources. How much is there? Estimates vary, but it could be significant enough to provide relatively cheap local NG for decades, especially as PV and other renewables displace NG-generated electricity. Any long-term energy forecast should allow for lower cost gas to assume some of oil’s historical roles (vehicle fuel, feed stock, etc.). On the other hand, large U.S. LNG export ports now being built will force U.S. NG to compete on a global market, pushing prices up to global parity. My gut sense is that, in the long term, cheaper North American NG will help offset U.S. oil demand, but will have little impact on a global scale.

Energy giant Sasol is building a massive $20B NG-to-oil conversion plant (100,000 b/d) in Louisiana, betting that gas prices will stay low, and oil prices will keep rising. This project is the “…largest foreign manufacturing investment in the history of the United States.” I personally see North America becoming free from most OPEC oil by 2040, but domestic independence does not solve oil demand and depletion on a global level. Our domestic economy is inexorably linked to the global economy, and I suggest that oil will remain the #1 impediment to global growth for the next half century.

Looking past this conservative forecast, it’s my hope that PV achieves “median global grid parity” far sooner – say 2025 – and that the accelerating move to PV electricity will be a strong market signal towards rapidly prioritized electric mobility and storage research. Given the power of economics to change historical momenta, I would not be surprised to see pure electric vehicles by 2042 that outperform IC vehicles in every metric, including life-cost and range, with short charge times. For those requiring “fast fill” I would not be surprised to see the average 2042 “plug-in hybrid sedan” approaching 100 MPG, nor would I be surprised to see pure EVs with a small “emergency” fossil engine.

(Often overlooked in these discussions is the true social and environmental costs of carbon-based power. The complete cost of fossil energy is higher than its raw extraction and generation costs, even before considering any nightmarish greenhouse gas scenario. Depending on who you read, the social cost of atmospheric carbon is anywhere from $2 to $250 per ton. Let’s use the current U.S. Government estimate of $21/ton. We know that coal creates 2.1 pounds of CO2 per kwh. At $21/ton, coal’s social cost becomes $0.02/kwh. A growing number researchers say this cost should be closer to $0.20/kwh.)


2 thoughts on “Updated: Thirty-Year Solar PV Forecast

  1. Installation labor cost is already a third of the total cost of a panel.

    It’s unrealistic to assume more than a 2/3rds reduction in price ever.

    • There are two key reasons why installation costs will continue to drop, and both were addressed in my post. Today’s rooftop install cost is around $1.00/W. By 2040, rooftop PV install cost should be in the $0.50/W range, in today’s dollars. The only way this can’t happen is if PV efficiencies remain stuck at today’s ratings, and PV+roofing manufacturers fail to improve form-factors and integration methods.

      We know that (1) PV efficiencies have predictably improved for 50 years, and (2) labs are today showing remarkable progress in doubling or even tripling electrical efficiencies. I see little reason to think that we won’t double commodity PV efficiencies by 2040 (effectively halving the number of panels required per rooftop), along with many types of improvements in rooftop mounting and integration techniques, easily giving us $0.50/W install costs (in 2013 dollars). And as PV efficiencies continue to improve after 2040, installation costs will continue to fall, possibly nearing $0.25/W as we approach theoretical PV efficiency limits (think 1kW per panel).

Leave a Reply

Your email address will not be published. Required fields are marked *