Solar power prices have been dropping faster than people expected, even faster than experts expected, and even faster than bullish experts expected.
I wrote about this general point a couple of years ago for an article for The Economist Group. The question I was supposed to explore was whether we needed solar technology breakthroughs in order for solar to take over the electricity generation market. Various solar experts emphasized to me that known, predictable, incremental improvements in solar PV technology and manufacturing would keep bringing down the cost of solar PV — at an incredible clip. The standout point was from Jenny Chase of Bloomberg New Energy Finance (BNEF), who highlighted that her team forecast a solar module* price drop from 62¢/watt in 2015 to 21¢/watt by 2040. That was with an assumption of no technological breakthroughs.
During a Zayed Future Energy Prize review committee meeting in September (I’m on the committee), I brought this up in one of the discussions we were having. Other analysts from BNEF who were there noted, “Actually, they’ve come down much quicker than we expected.” The average had already dropped to approximately 40¢/watt. They showed me a graph of the more recent cost drops as friendly proof. Astounding — even BNEF’s relatively “bullish” projections were far too conservative.
These BNEF analysts recommended I talk to Jenny Chase again, since she’s top of the world for tracking the solar industry, these prices, and what’s behind them. Incidentally, I was giving a presentation at Intersolar ME a couple of days later where I ended up standing right next to her at an evening rooftop reception that I nearly skipped. We had a nice long chat about the interesting history of BNEF, the disruptive solar energy industry, Donald Trump (ugh), and these surprising cost drops. I asked Jenny to provided some follow-up statements on email so that we could have the points in a nice format for this article. Here’s a deeper look at what happened, as well as some broader points of interest:
To be honest, over the past two years we have revised our experience curve down (again) — it looks like the learning rate for every doubling of cumulative manufactured crystalline silicon is more like 28% than the previously estimated 26%. PERC, diamond wire saws, and general fierce competition have been very effective in bring down prices. I suspect we’ll see more stable pricing in the next two years — we expect that, if the average module sold to a utility scale project in 2017 was about 35 cents, it will fall to 32 cents in 2018 and 31 in 2019. But of course we haven’t been very good at predictions in the past!
I think that the faster-than-expected drop has had an impact on record auction bids, but not the obvious one. Bidders at these extremely low (sub $40/MWh) prices aren’t going by costs they have already scoped out, as was evidenced this year by the extreme surprise of Indian developers about a small firming of the module price. The bidders are making a gamble on both module prices and EPC costs, because if they don’t, they won’t win projects. Historically it’s been a good gamble because modules (and EPC) have got cheaper faster than expected, so a price that looks terrible when bid has often turned out profitable when built.
However, they are speculation, and some of these projects may turn out not to be build-able (as found in some of the Brazilian projects bid in 2014, which have been cancelled despite falling tech costs due to the weaker Brazilian currency). The people who bid them will be long gone to their next job. I don’t think it is a terrible thing if these projects aren’t built, but governments may have to re-tender some of them. A few may also be built on the cheap and not perform to specifications, which again will not be the problem of the people who made the bid.
These are topics we’ve discussed before and will discuss further in coming articles, but the bottom-line point about solar panel prices coming down faster than expected is at the core of many solar articles here on CleanTechnica.
I recently discussed this matter in Abu Dhabi with Thierry Lepercq, Executive Vice President of Engie in charge of Research, Technology and Innovation, and separately with Michael Liebreich, founder of BNEF. One point that launched some interesting discussions: the unexpectedly low price of solar meant that new solar was starting to get competitive with electricity generation from existing fossil and nuclear power plants. It was just a couple of years ago that new solar was getting cost-competitive with new fossil power plants. That’s a big enough change, but when solar and wind get cheaper than electricity from 20-year-old, 10-year-old, and even 1-year-old power plants … things get interesting, and messy. This is a different kind of disruptive.
With “incremental” improvements to solar cell and solar panel technology, and improvements to how solar cells and solar panels are produced, costs have been coming down at an impressive rate and are putting solar at that point. This decreases the need, usefulness, and potential of any “breakthroughs” in solar technology, since solar is already getting to the point where it will beat already built and operating coal, nuclear, and natural gas power — but breakthroughs might be on the horizon as well. I’ll come back to that in a future article, but as a short teaser, here’s Adnan Z. Amin, Director-General of the International Renewable Energy Agency (IRENA), talking about one potential leap forward:
Getting back to the actual price of solar modules today — the topic of this story, after all — Jenny just sent me this update: “We have pegged the typical factory-gate module price at the end of 2017 at about 33 cents per W, meaning a price around 37 cents per W in most markets. About 98GW was installed in 2017.”
Remember, that’s after a “bullish” 2015 projection that solar modules would drop from 62¢/watt in 2015 to 21¢/watt by 2040.
The short story is what Royal Dutch Shell acknowledged in 2013: Solar power is going to dominate the energy sector this century. The question is just how quickly it will bring down costs and take over the world. Yet again, the answer seems to be: quicker than people thought.
*Chinese multicrystalline silicon solar modules, the most common sort.
Nope, Cobalt’s Not A Problem — The EV Revolution Will March On (coming later today)