Can renewable deployment be exponential?

In recent months we have all had to learn quickly about exponential rates of change as COVID-19 spread throughout the world, but as has been seen with the virus itself, exponential patterns typically stop and some other pattern of change emerges. Some mechanism intervenes in the process, typically starving the exponential process of its means for reproduction.

With the urgency around rising levels of carbon dioxide in the atmosphere, there is a desire by many to see renewable energy grow exponentially and quickly back out fossil fuels. Four years ago Fortune magazine reported that American futurist Ray Kurzweil forecast the dominance of solar PV in a little over a decade. Fortune noted the following;

Kurzweil’s basic point, reported by Solar Power World, was that while solar is still tiny, it has begun to reliably double its market share every 2 years — today’s 2% share is up from 0.5% in 2012. Many analysts extend growth linearly from that sort of pattern, concluding we’ll see 0.5% annual growth in solar in the future, reaching 12% solar share in 20 years. But linear analysis ignores what Kurzweil calls the law of accelerating returns — that as new technologies get smaller and cheaper, their growth becomes exponential. So instead of looking at year over year growth in percentage terms, Kurzweil says we should look at the rate of growth — the fact that solar market share is doubling every 2 years. If the current 2% share doubles every two years, solar should have a 100% share of the market in 12 years.

There is no doubt that solar PV has been doubling on a near two year basis. In 2015 global deployment was 230 GW installed capacity and this year it should be approaching 800 GW, which isn’t quite doubling every two years, but is close. But can this rate of change persist? It would imply 10 TW installed capacity by 2030 and 120 TW by 2040, from which enough energy could be generated to about meet all global requirements (~900 EJ in 2040), assuming it could be stored and channeled to the necessary services at the right time and place.

A related issue emerged on Twitter recently when well-known climate and energy expert Glen Peters (@Peters_Glen) took issue with a paper written by two colleagues in the Shell scenario team and published in Nature over a decade ago. In a twelve-tweet discussion Glen challenged the findings;

Will energy technologies grow exponentially until they reach “materiality” & then have linear growth? A common perception, but what does history say? Short answer: Old technologies are growing exponentially (not nuclear), just slower due to crowding out effects! . . . . . The ability for renewables to *displace* fossil fuels (not add to existing supply) will dictate whether renewables can grow exponential beyond “materiality”. If fossil fuels are taken out of the energy system, there is more space for renewables to grow exponentially. There is probably no physical reason why we can’t have 100% renewables, just like we once had 100% bioenergy, 50%-50% bioenergy-coal, etc The only limit is the one in our heads . . . .

There are a number of points here, but the question of exponential growth keeps returning. This isn’t about the level in the energy mix that renewables finally reach, but about the time it might take to get there. Exponential growth, such as that proposed by Ray Kurzweil, ought to be able to deliver a 100% renewable energy system by 2040.

The original paper that Glen Peters refers to was written by Martin Haigh and Gert Jan Kramer, so I took the question to Martin for further discussion. He noted that much of the background work to their Nature paper concerned the supply industry constraints related to production of materials required for deployment, but they had only limited space in their article to go into the necessary detail with graphs and charts.

So in practice, ongoing exponential deployment of solar is unlikely to happen because of stranded solar PV production facilities. Let me explain using two simple scenarios. In both cases the goal is to supply global energy demand with solar PV (putting aside the technical issues related to actually doing this) and global demand is rising from the current level of about 600 EJ by 1% per annum. The basic model I have developed also assumes that a field installation lasts for 35 years and that newer cells have improved efficiency over older models.

In an exponential scenario, solar PV rises at the rate described above, but this must be supported by investment in manufacturing facilities. By the time global demand is met in 2040, the world would have PV manufacturing capacity of 18,000 GW per annum, doubling from 9000 GW per annum just three years earlier. Even building the production facilities at this rate is questionable. But most of this capacity would then become stranded assets as demand for new solar PV would collapse. There would be residual demand for the newest most efficient modules replacing older modules, but this wouldn’t keep the industry going. In reality, this scenario wouldn’t happen as investors would see the coming over capacity problem and back away, therefore starving the exponential process of its means of reproduction.

In an investment scenario, solar PV rises exponentially in the early years, but then starts to shift away from the exponential case as investment in PV cell manufacture tapers. The result is a much slower rise towards meeting demand, but the investors in manufacturing don’t end up with stranded assets. In this example, manufacturing levels out at around 3000 GW per annum, one sixth of the above. Once solar PV generation meets global demand, the industry remains roughly balanced as it caters for both growth in energy demand and replacement of older installation. For much of the deployment period, growth is linear, not exponential.

Exponential Scenario

Charts shown are for illustration purpose only

In the latter part of the century, some additional cell manufacturing capacity would be warranted, rising to around 4000 GW per annum due to the increasing replacement requirement. This would close the energy supply gap that is starting to appear after 2080.

But the outcome of the investment scenario is a 30 year delay in meeting the goal of 100% of global energy demand supplied by solar PV. This then challenges the other goal behind the requirement for very rapid deployment of new energy infrastructure; the need to limit warming to 1.5°C.