Empirical knowledge of innovations can be gained from the IT sector, which abounds in new ideas and innovations. Looking at how quickly tablets and smartphones became commonplace in our daily lives, it seems that in only a decade we may expect to see our roads travelled by electric cars and electricity being generated using wind and solar energy. Let us take a look at photovoltaics. The world’s entire installed solar power capacity rose from less than 1 GW in 2000 to 39 GW in 2010 and 176 GW in 2014, an average exponential growth by nearly 45% annually. Every two years, the capacity is doubled. In accordance with Swanson’s law, photovoltaics are quickly progressing to grid parity, which is believed to open up the way for a massive switch to solar energy. The global capacity of wind turbines, which have a longer history, grew by 22% annually between 2005 and 2014, doubling every three and a half years, and has already achieved grid parity.
However, transposing the experiences of innovation from the sphere of IT to that of new power generation technologies should be done with extreme caution, especially with respect to their time course. The basic difference between the innovation cycles in IT and power generation is the economic life of the devices used. For smartphones, it is two years, whereas for solar panels and wind turbines − 20 or more years. If we wanted to give new phones to all smartphone owners in the world it would take us about two years. That is the time after which people replace their old smartphones with new models, and the industry is tuned to the replacement rate of 50%. However, if we wanted to provide all house owners with a new house, that would obviously take a lot longer − some 50 years. The world’s house-building capacity is set to the replacement rate of 2%. Seen from this perspective, solar panels and wind turbines are more like houses than smartphones, as only several percent of them will be replaced every year. Once they reach maturity, the solar and wind energy industries will also adapt their capacity to appropriate replacement rates. Currently, they are still building the capacities and their (net) contribution to the global energy supply is negative. When can we expect them to reach maturity?
As N.L. Cardozo, G. Lange and G.J. Kramer argue in the essay ‘The cradle of new energy technologies. Why we have solar cells but not yet nuclear fusion?,’ published in December 2015 in ‘The colours of energy. Essays on the future of energy and society’, it may take a while and may coincide with controlled nuclear fusion reaching maturity too. The authors arrived at this conclusion by studying the relatively simple mathematical model of innovative technologies reaching maturity, which uses the S-curve, widely discussed in the literature.
In a nutshell: every new energy technology goes through three phases of development. First is the exponential growth phase, during which installed capacities double every 3 to 4 years. This stage may last for several decades. From the economic point of view, it is an innovation scaling phase, which involves investing heavily in the new technology. In this phase, factories and dedicated machinery and equipment need to be built, infrastructure and supply chains of materials developed, raw materials mined and staff trained. During this time, the technology is brought from the laboratory to materiality. Its global industry becomes a new energy source (net producer) and starts to be visible on the radar of the world energy market.
The authors argue that a new technology with the target energy generation capacity equal to 10% of the global demand materialises when it meets 1% of the demand. By the time such capacity is reached, a significant industry has already been established. To illustrate this point: the worldwide investment in photovoltaics was USD 100bn in 2012, slightly more than 1% of the world’s expenditure on energy. Yet, the total contribution of photovoltaics to the global energy supply was still below 0.1%.
Reaching maturity affects the further trajectory of the new technology development. Its growth is no longer exponential, but linear, with similar capacities added every year. This phase also lasts a few decades and ends when the technology reaches its global capacity. A saturation phase follows and the installed base levels off.
Such a shape and characteristics of the S-curve lead to some interesting conclusions concerning innovative power generation technologies.
- One consequence of the exponential growth and capacity doubling at more or less regular intervals is that (globally) more energy is consumed than produced by the new technologies in the first growth phase.
- The (net) consumption of energy in the first phase of renewable energy development (which involves investing heavily) means that, globally, the technologies do not yet reduce carbon dioxide emissions or generate income (they do not reach grid parity).
- This first phase is entirely financed by the public (taxpayers).
- It is a necessary and costly investment, which precedes the return by several decades.
How big an investment are we talking about? The authors point out that for a new technology to reach 1% share in the global energy supply, between EUR 900bn and EUR 1,800bn need to be invested over a period of several decades. The above is true both for photovoltaics and nuclear fusion. The only difference is temporal distribution. Whereas the prototype solar cell was operational more than 60 years ago, the 500 MW ITER reactor is still under construction. If it develops successfully, nuclear fusion will come in a few big steps, entailing a huge risk of failure. Photovoltaics, on the other hand, will continue to develop in many small steps, each involving a relatively small and manageable risk. It is this risk profile rather than anything else that gives photovoltaics an advantage over nuclear fusion.
How long can the exponential growth phase (with large public investment) last? The authors calculated that a new energy technology that leaves the lab with the capacity of 10 MW of effective, year-averaged power must grow by a factor of 20,000 to reach the 1% of world demand mark. That corresponds to more than 14 doublings. Even when the capacity is doubled every three years, that still requires 40 years of exponential growth before the world can reap any rewards from this technology.
Can this period be shortened? The authors say it can. The mathematical property of exponential growth is that 50% of capacity is installed in the last few years before materiality is reached, and the process consumes 50% of the entire investment. Therefore, it is a good idea to speed up the earlier phase, when investment is still at a much lower level. The earlier in the development, the more time can be gained at a lower cost. There really is no good reason not to leapfrog a few development generations by taking higher risks. Those risks are tiny compared with the social and economic gains that can be achieved by reaching the productive phase earlier.
Even more effective, though less controllable, would be acceleration in the research phase prior to exponential growth. Research budgets are but fractions of the turnover in the industrial implementation phase.