> It took the solar PV industry 68 years to reach 1 TW of installed capacity - from 1954-2022.
> It has taken only 2 years to reach the next TW (2022-2024), with the 2 TW milestone reached in recent weeks according to estimates calculated by the Global Solar Council and SolarPower Europe.
In 2023 we built 450 GW, we will be interesting to see where we land in 2024.
Either way we are looking at a complete reshaping of the global energy system before 2030.
In uncertain times, this is excellent and refreshing news, because it means it will be very hard for fossil interests to slow the avalanche of clean energy transition simply out of cost decline and manufacturing learning rate figures.
Too early to tell how quickly deployment velocity will increase (cost of capital, supply chain, labor constraints), but existing manufacturing capacity (~1.1TW) is built; that will not slow down. PV printer goes brrr.
Using GSC's 20% estimated capacity factor, you'd need 100TW of PV installed capacity to satisfy even current world marketed energy consumption of around 20TW, which would be 90 years at 1.1TW per year. But atmospheric carbon capture, not to mention other advancements, will probably require a significant amount of energy on top of that.
Exponential growth, by contrast, would raise these 2TW to the required 100TW in 20 years if we assume a 22% annual growth rate, 10 years if we assume 48%.
Think in systems. Yes, it is going to take some time to continue to deploy clean energy globally to meet total human energy consumption. But, as we hit economic inflection points, that will rapidly force coal and fossil gas out of the mix (unless someone with deep pockets is willing to light fiat on fire to keep uneconomic generation online). Once that fossil generation is demolished, it will be unlikely it comes back, locking in a clean path forward.
Pakistan comes to mind, where so much solar is coming online from individuals and businesses, utility demand is rapidly being destroyed (potentially stranding those utility fossil generation assets and their capital investment).
The sun is almost free, and batteries are getting cheap fast. Enough sunlight falls on the Earth in a few hours to power humanity for a year.
Interesting story - in pakistan, PV is an arbitrage/loophole.
Pakistan exporters are under capital controls, so they get a very bad exchange rate when they convert to ruppee. Instead, they buy panels (getting 100% worth of their dollars), which they ship into pakistan, and sell.
On the other side of the trade, home owners get a very generous feed in tarrif, due to pressure by wealthier constituents. This hits the local utility thrice (lost revenue, paying the feed in, AND massive duck curve)
One nitpick, though. The article says, "Pakistan's solar revolution is only possible because of the plummeting cost of solar PV modules, which have reduced in price by 90% in the last 15 years alone," but the correct number is 94%, from €1.67/Wp for crystalline panels in China in November 02009 http://web.archive.org/web/20121027002200/http://www.solarse... to €0.10/Wp today https://www.solarserver.de/photovoltaik-preis-pv-modul-preis.... So the article understates the actual price drop by almost a factor of 2.
> (potentially stranding those utility fossil generation assets and their capital investment).
This is not true. For most of Pakistan's power plants, the loans (denominated in US dollars) were given in exchange of sovereign guarantees by the State of Pakistan. So, the capitalists, in theory, don't have anything to worry about. The whole country could switch to renewables, and the State of Pakistan will still have to pay back on their investments. The reason for the sovereign guarantee is because Pakistan does not have a stable political landscape, so the free market risk is too high for such large investments.
In practice, the government is forcing local investors to renegotiate some of the loan terms or contracts to reduce its payment obligations [1]. When I say forcing, I mean, mafia style. It's not clear how much this will help.
Someone is taking a haircut on that debt, that is the stranding. Whether it’s Pakistan, the World Bank, or institutional investors, ¯\_(ツ)_/¯. They could forgive the debt instead, but humans gonna human.
Sovereign defaults are a thing, and sometimes necessary. It’s just numbers in a spreadsheet or database after all.
Institutions like World Bank, Asian Development Bank or IMF will not take any sort of cuts or deviations from the loan schedule.
International commercial investors won't also. They are also too many to negotiate with. Defaulting on those loans will send Pakistan's bonds and exchange rate into a tailspin.
That leaves other countries. US and Europe will not change loan terms either. So what has happened since 2018, is that Pakistan takes additional loans from China, UAE or Saudia Arabia to pay off other loans. If then Pakistan is unable to pay their loans backs, these countries reschedule payments into the future. But because of consistent refusal of the Pakistan state and elite to restructure and improve the economy has meant that even these countries have lost their appetite for help.
Probably https://news.ycombinator.com/item?id=42266903 provides some useful clues about where the inflection point will be. 2 terawatts is about 1/64000 of the 128000 terawatts of ground-level insolation. Probably people will not cover literally every square meter of the planet with solar panels, but why would they stop much short of that goal?
In Shandong they're already building a 1-gigawatt solar farm offshore, 20 years earlier than I expected people to start doing that: https://www.youtube.com/watch?v=l_MlFOizMBo
The expected capacity factor for the Shandong project is 20.3%.
The peak generating capacity is off the charts and most of the largest plants are situated next to pumped hydro to give storage.
You've probably heard that solar installation is exponential and that it's now cheaper than nuclear even with storage and transmissions costs factored in and this is a great way to see it and understand it. https://reneweconomy.com.au/csiro-says-wind-and-solar-much-c...
Thank you for calling it a "milestone". I try not to have language peeves, but "barrier" is not a substitute for "milestone". There has to be at least some tenuous reason to think that the limitation might be difficult to exceed, to merit the term "barrier".
As useful context, world marketed energy consumption is about 20 terawatts, of which about 5 terawatts is China. Ground-level terrestrial insolation is about 128000 terawatts.
In this press release, the Global Solar Council is estimating an average 20% capacity factor, which means 2 terawatts of installed capacity amounts to 0.4 terawatts of average power. 20% seems like a broadly reasonable capacity factor; https://en.wikipedia.org/wiki/Capacity_factor#Capacity_facto... shows that from 02013 to 02018 the PV capacity factor in the US varied between 25.1% and 26.1%.
However, yesterday I was looking at CREA's statistics on China https://energyandcleanair.org/wp/wp-content/uploads/2024/08/.... Their plot seems to show 40 TWh generated from solar in June. 40 TWh per month works out to 55 GW. If we assume 670 GW of solar generation capacity was installed as of June (of the current 790 GW in https://www.nea.gov.cn/2024-11/22/c_1310787342.htm, which has grown 48% over the last 12 months) that's an 8% capacity factor, which is implausibly low—when I've done the stats for previous years in China, the capacity factor was about 10%, which was already very low.
Am I calculating this wrong? Are there better numbers on China's solar capacity factor? Why would it be so low? Will it improve if, e.g., more HVDC transmission lines or battery storage are built?
It's understandable that financing is tricky in the face of such rapid change. It seems likely that most solar farms built two years ago will owe too much money to ever be profitable due to competition from much cheaper PV modules—either in newly-built solar farms or in difficult-to-stop distributed generation. As toomuchtodo points out in https://news.ycombinator.com/item?id=42267156, this is already a significant factor in Pakistan.
As I remember from the times we did solar installs in Tanzania, 8-10% is about right for PV in real life. It's the 20+% figures look like creative accounting.
It varies a lot by geography. 10% is right for Germany and the Netherlands. Tanzania is pretty cloudy, isn't it? Almost half rain forest and moist forest? Though I'd've expected you'd put the solar farms on the savannah.
>If we assume 670 GW of solar generation capacity was installed in June
Was installed /as of/ June. (sorry I was double checking your workings and got confused)
Assuming the figures aren't typos.
1) What is counted as generation capacity? a 'finished' solar farm that is waiting for paper work / connection to the grid could be included in the figures.
2) What about solar that doesn't see the grid? the panels on my roof might be included as installed capacity, but not show in generation figures.
edit:
3) another commenter mentioned solar with pumped hydro. Is that classed as pumped hydro in the generation figures?
Exponentials are always hard to grasp.
> It took the solar PV industry 68 years to reach 1 TW of installed capacity - from 1954-2022.
> It has taken only 2 years to reach the next TW (2022-2024), with the 2 TW milestone reached in recent weeks according to estimates calculated by the Global Solar Council and SolarPower Europe.
In 2023 we built 450 GW, we will be interesting to see where we land in 2024.
Either way we are looking at a complete reshaping of the global energy system before 2030.
In uncertain times, this is excellent and refreshing news, because it means it will be very hard for fossil interests to slow the avalanche of clean energy transition simply out of cost decline and manufacturing learning rate figures.
Too early to tell how quickly deployment velocity will increase (cost of capital, supply chain, labor constraints), but existing manufacturing capacity (~1.1TW) is built; that will not slow down. PV printer goes brrr.
Using GSC's 20% estimated capacity factor, you'd need 100TW of PV installed capacity to satisfy even current world marketed energy consumption of around 20TW, which would be 90 years at 1.1TW per year. But atmospheric carbon capture, not to mention other advancements, will probably require a significant amount of energy on top of that.
Exponential growth, by contrast, would raise these 2TW to the required 100TW in 20 years if we assume a 22% annual growth rate, 10 years if we assume 48%.
Think in systems. Yes, it is going to take some time to continue to deploy clean energy globally to meet total human energy consumption. But, as we hit economic inflection points, that will rapidly force coal and fossil gas out of the mix (unless someone with deep pockets is willing to light fiat on fire to keep uneconomic generation online). Once that fossil generation is demolished, it will be unlikely it comes back, locking in a clean path forward.
Pakistan comes to mind, where so much solar is coming online from individuals and businesses, utility demand is rapidly being destroyed (potentially stranding those utility fossil generation assets and their capital investment).
The sun is almost free, and batteries are getting cheap fast. Enough sunlight falls on the Earth in a few hours to power humanity for a year.
https://www.dw.com/en/pakistan-solar-power-renewable-energy-...
https://oilprice.com/Alternative-Energy/Solar-Energy/The-Uni...
> Pakistan comes to mind,
Interesting story - in pakistan, PV is an arbitrage/loophole.
Pakistan exporters are under capital controls, so they get a very bad exchange rate when they convert to ruppee. Instead, they buy panels (getting 100% worth of their dollars), which they ship into pakistan, and sell.
On the other side of the trade, home owners get a very generous feed in tarrif, due to pressure by wealthier constituents. This hits the local utility thrice (lost revenue, paying the feed in, AND massive duck curve)
Sigh. Humans will human...
(PV is still great and we need more of it)
Agreed! Thank you very much for the DW article; I hadn't seen that, though I did mention that dynamic in https://news.ycombinator.com/item?id=42266903.
One nitpick, though. The article says, "Pakistan's solar revolution is only possible because of the plummeting cost of solar PV modules, which have reduced in price by 90% in the last 15 years alone," but the correct number is 94%, from €1.67/Wp for crystalline panels in China in November 02009 http://web.archive.org/web/20121027002200/http://www.solarse... to €0.10/Wp today https://www.solarserver.de/photovoltaik-preis-pv-modul-preis.... So the article understates the actual price drop by almost a factor of 2.
> (potentially stranding those utility fossil generation assets and their capital investment).
This is not true. For most of Pakistan's power plants, the loans (denominated in US dollars) were given in exchange of sovereign guarantees by the State of Pakistan. So, the capitalists, in theory, don't have anything to worry about. The whole country could switch to renewables, and the State of Pakistan will still have to pay back on their investments. The reason for the sovereign guarantee is because Pakistan does not have a stable political landscape, so the free market risk is too high for such large investments.
In practice, the government is forcing local investors to renegotiate some of the loan terms or contracts to reduce its payment obligations [1]. When I say forcing, I mean, mafia style. It's not clear how much this will help.
[1] https://www.brecorder.com/news/40331398
Also see my comment here https://news.ycombinator.com/item?id=42268194
If you are a local resident, is this true?
Someone is taking a haircut on that debt, that is the stranding. Whether it’s Pakistan, the World Bank, or institutional investors, ¯\_(ツ)_/¯. They could forgive the debt instead, but humans gonna human.
Sovereign defaults are a thing, and sometimes necessary. It’s just numbers in a spreadsheet or database after all.
Institutions like World Bank, Asian Development Bank or IMF will not take any sort of cuts or deviations from the loan schedule.
International commercial investors won't also. They are also too many to negotiate with. Defaulting on those loans will send Pakistan's bonds and exchange rate into a tailspin.
That leaves other countries. US and Europe will not change loan terms either. So what has happened since 2018, is that Pakistan takes additional loans from China, UAE or Saudia Arabia to pay off other loans. If then Pakistan is unable to pay their loans backs, these countries reschedule payments into the future. But because of consistent refusal of the Pakistan state and elite to restructure and improve the economy has meant that even these countries have lost their appetite for help.
They’re almost always S-curves; but no one knows where the inflection point will be.
Probably https://news.ycombinator.com/item?id=42266903 provides some useful clues about where the inflection point will be. 2 terawatts is about 1/64000 of the 128000 terawatts of ground-level insolation. Probably people will not cover literally every square meter of the planet with solar panels, but why would they stop much short of that goal?
In Shandong they're already building a 1-gigawatt solar farm offshore, 20 years earlier than I expected people to start doing that: https://www.youtube.com/watch?v=l_MlFOizMBo
The expected capacity factor for the Shandong project is 20.3%.
It's fun to go to https://en.wikipedia.org/wiki/List_of_photovoltaic_power_sta... and look at the map view of these solar plants. The first link is roughly 10x20km in size, next to a big dam to hold pumped hydro.
The peak generating capacity is off the charts and most of the largest plants are situated next to pumped hydro to give storage.
You've probably heard that solar installation is exponential and that it's now cheaper than nuclear even with storage and transmissions costs factored in and this is a great way to see it and understand it. https://reneweconomy.com.au/csiro-says-wind-and-solar-much-c...
Thank you for calling it a "milestone". I try not to have language peeves, but "barrier" is not a substitute for "milestone". There has to be at least some tenuous reason to think that the limitation might be difficult to exceed, to merit the term "barrier".
Or it could just be straight propaganda from the oil cos, as they are so fond of doing
As useful context, world marketed energy consumption is about 20 terawatts, of which about 5 terawatts is China. Ground-level terrestrial insolation is about 128000 terawatts.
In this press release, the Global Solar Council is estimating an average 20% capacity factor, which means 2 terawatts of installed capacity amounts to 0.4 terawatts of average power. 20% seems like a broadly reasonable capacity factor; https://en.wikipedia.org/wiki/Capacity_factor#Capacity_facto... shows that from 02013 to 02018 the PV capacity factor in the US varied between 25.1% and 26.1%.
However, yesterday I was looking at CREA's statistics on China https://energyandcleanair.org/wp/wp-content/uploads/2024/08/.... Their plot seems to show 40 TWh generated from solar in June. 40 TWh per month works out to 55 GW. If we assume 670 GW of solar generation capacity was installed as of June (of the current 790 GW in https://www.nea.gov.cn/2024-11/22/c_1310787342.htm, which has grown 48% over the last 12 months) that's an 8% capacity factor, which is implausibly low—when I've done the stats for previous years in China, the capacity factor was about 10%, which was already very low.
Am I calculating this wrong? Are there better numbers on China's solar capacity factor? Why would it be so low? Will it improve if, e.g., more HVDC transmission lines or battery storage are built?
I note that some recent Chinese-built solar farms abroad like the Gobustan farm in Azerbaijan (60km from CoP29) https://www.youtube.com/watch?v=qIJA5hHHawA are using single-axis sun trackers, which you'd expect to raise the capacity factor significantly over the 29% achieved by utility-scale solar in California, which is almost entirely fixed in place and thus unable to track the sun. [Edit: pfdietz says this isn't true! https://news.ycombinator.com/item?id=42266903 https://emp.lbl.gov/sites/default/files/2024-10/Utility%20Sc...]
Module prices have dropped about 70% (more than a factor of 3) in the last two years, to €0.10 per peak watt https://www.solarserver.de/photovoltaik-preis-pv-modul-preis... since the apparent failure of the price-fixing cartel announced at Davos five years ago https://www.reuters.com/article/us-davos-meeting-solar-gcl-i...
It's understandable that financing is tricky in the face of such rapid change. It seems likely that most solar farms built two years ago will owe too much money to ever be profitable due to competition from much cheaper PV modules—either in newly-built solar farms or in difficult-to-stop distributed generation. As toomuchtodo points out in https://news.ycombinator.com/item?id=42267156, this is already a significant factor in Pakistan.
> utility-scale solar in California, which is almost entirely fixed in place and thus unable to track the sun.
This is not true at all. Most utility scale PV in California uses trackers, I believe single axis.
https://emp.lbl.gov/sites/default/files/2024-10/Utility%20Sc...
https://emp.lbl.gov/technology-trends
Thank you very much for the correction!
As I remember from the times we did solar installs in Tanzania, 8-10% is about right for PV in real life. It's the 20+% figures look like creative accounting.
It varies a lot by geography. 10% is right for Germany and the Netherlands. Tanzania is pretty cloudy, isn't it? Almost half rain forest and moist forest? Though I'd've expected you'd put the solar farms on the savannah.
No, it was savannah in the middle of nowhere and mostly dry and sunny.
>If we assume 670 GW of solar generation capacity was installed in June
Was installed /as of/ June. (sorry I was double checking your workings and got confused)
Assuming the figures aren't typos.
1) What is counted as generation capacity? a 'finished' solar farm that is waiting for paper work / connection to the grid could be included in the figures.
2) What about solar that doesn't see the grid? the panels on my roof might be included as installed capacity, but not show in generation figures.
edit: 3) another commenter mentioned solar with pumped hydro. Is that classed as pumped hydro in the generation figures?
Thank you for the correction! I didn't spot the ambiguity.
All three of your questions are good ones, and I don't know the answers.