IEA (2020), Clean Energy Innovation, IEA, Paris https://www.iea.org/reports/clean-energy-innovation, License: CC BY 4.0
- A cleaner and more resilient future energy system with net-zero emissions will require a wide range of technologies, some of which are still at an early stage of development. For these new technologies, innovation is an uncertain and competitive process: many ideas fall by the wayside. This report looks at how to manage uncertainty and expand the number of available and affordable clean energy technologies in support of net‑zero emissions on a timetable compatible with international energy and climate goals. It features new IEA modelling that highlights candidates – including electrification (supported by batteries), hydrogen (and its derived fuels), CO2 capture and bioenergy – that could speed up progress in long-distance transport and heavy industry, sectors that in most cases lack readily scalable low-carbon technologies today. For policy makers, it offers recommendations for action.
- Successful technology concepts eventually pass through four stages: prototype, demonstration, early adoption and maturity. Feedback between the stages means that technology options are always evolving. Size, consumer value and synergies with other technologies are all attributes that determine the speed with which technologies pass through the stages.
- The process of innovation involves a wide range of participants: governments, researchers, investors, entrepreneurs, corporations and civil society all play important roles in generating ideas for new or improved technologies and in improving and financing them right through to market entry and deployment. Innovation systems are complex and rest on four pillars: resource push, knowledge management, market pull and socio-political support.
- Governments have a particularly central and wide-ranging role to play that goes far beyond the provision of funds for R&D. They set overall national objectives and priorities and play a vital role in determining market expectations. They also have unique responsibilities for ensuring the flow of knowledge, investing in enabling infrastructure and facilitating major demonstration projects.
- The remarkable 70-year history of almost continuous cost reductions for solar PV illustrates how governments can effect change. At different stages, the US, German, Chinese and other governments used R&D and market-pull policies, including targets and revenue guarantees, to encourage investments all along the value chain that supported innovation and economies of scale. The way in which lithium-ion batteries have developed has showed similar patterns.
- The Covid-19 pandemic potentially brings about a major and unanticipated setback to clean energy innovation, and an IEA survey reveals that companies that are developing net-zero emissions technologies consider it likely that their R&D budgets will be reduced. The economic recovery plans now being developed on a large scale by a range of countries, however, provide an opportunity for governments to support clean energy innovation jobs and accelerate technology progress, at a time when the need for such innovation has never been greater.
A rapid shift to net-zero emissions of greenhouse gases is needed if we are to meet the energy-related Sustainable Development Goals of the United Nations, including by mitigating climate change in line with the Paris Agreement. This requires the use of a wide range of clean energy technologies. Some of these are well established; others are still at an early stage of development, or exist only as prototypes. Further technologies may emerge in due course from current research work. These energy technologies also offer the prospect of other benefits, including cleaner air and greater energy security as a result of, for example, improved electricity systems flexibility.
Success will not be easy or straightforward. It depends upon technological innovation, and this takes time: it has taken decades for solar photovoltaics and batteries to reach their current stage of development, for example. And not every technology that is developed will be successful; the evolution of existing and new technologies is inherently uncertain. But these points merely serve to underline the importance of finding ways to innovate that are successful in bringing about rapid change.
It is against this background that this Energy Technology Perspectives Special Report focuses on accelerating technology progress for a sustainable future. It emphasises that we are at a critical point, and it concludes with recommendations to help bring about real change.
Chapter 1 describes the steps involved in clean energy technology innovation and the role that government and other actors play throughout the innovation process, building on historical experience. It explains why strong and cohesive innovation systems are vital for clean energy transitions and looks at the risks and opportunities that may arise from the Covid-19 crisis.
Chapter 2 provides an overview of the status of clean energy technology innovation. It reviews the different resources that support innovation, from government and public sector funding for research and development (R&D) through to venture capital investment and also patents. It then assesses the potential impacts of the Covid-19 crisis on these different resources.
Chapter 3 looks at long-term clean energy technology innovation needs through the lens of the IEA Sustainable Development Scenario, which maps out a way to meet the key energy-related goals of the United Nations Sustainable Development Agenda, including by mitigating climate change in line with the Paris Agreement. The trajectory for emissions in the Sustainable Development Scenario is consistent with reaching global “net-zero” CO2 emissions by around 2070.
Chapter 4 discusses the opportunities and challenges arising from the Covid-19 crisis for clean energy technology innovation. It presents a Faster Innovation Case, which sets out what would be needed in terms of clean energy technology innovation to achieve net-zero carbon emissions by 2050. It also presents a Reduced Innovation Case, which sets out the risks and consequences of a delay in scaling-up key emerging clean energy technologies.
Chapter 5 concludes with recommendations for policy makers to boost clean energy technology innovation. It distinguishes near-term priorities from structural changes, and makes the case for immediate action in the light of the scale of the challenges we face and the long lead times involved.
This report treats technology innovation as the process of generating ideas for new products or production processes and guiding their development all the way from the lab to their mainstream diffusion into the market. At each stage of development there are funding risks, technical risks and market risks, which are influenced by various social and political factors. As a result, only a minority of products ever make it as far as mass market deployment in practice.
The innovation journey of any given technology is evolutionary. There are three main ways by which a technology evolves with experience to become better adapted to its environment, notably through improved costs and performance: 1) learning-by-researching; 2) learning-by-doing; and 3) economies of scale. As the technology is improved, it is more likely to be chosen by R&D funders and new users with different selection criteria. This creates a virtuous cycle and so‑called “increasing returns to adoption”. However, in the early stages, when costs are usually higher than those of competitors, these feedback loops are much weaker and it takes concerted, risky investments to access the first market opportunities. Both radical and incremental advances are vital to the process of innovation.
Choices about technology are made in an environment that is constantly changing, as companies, consumers, policies, competing technologies, infrastructure and social norms change. Technologies can become more attractive to users for a variety of reasons. These include changes in related technologies, consumer behaviour, policy and, sometimes, a change in the information available to users. Each of these variables can also change in ways that cause a technology to be overlooked in favour of alternatives, or lead to a technology that was previously rejected finding new market opportunities. Governments and private sector actors raise their chances of successful innovation by simultaneously addressing the improvement of technology, for example through research, and of the selection environment, for example though regulation, advertising or the development of new business models.
Successful innovation systems involve a wide range of actors with aligned interests and a wide variety of functions, each of which can be enhanced by public policy (Gallagher et al., 2012). These functions can be grouped under four headings. An innovation system will struggle to translate research into technological change without action under each of these headings. A sustained flow of R&D funding, a skilled workforce (e.g. researchers and engineers) and research infrastructure (laboratories, research institutes and universities) is required: these resources can come from private, public or even charitable sources, and can be directed to specific problems or basic research (resource push). It must be possible for knowledge arising to be exchanged easily between researchers, academia, companies, policy makers and international partners (knowledge management). The expected market value of the new product or service must be large enough to make the R&D risks worthwhile, and this is often a function of market rules and incentives established by legislation. If the market incentives are high, then much of the risk of developing a new idea can be borne by the private sector (market pull). And there needs to be broad socio-political support for the new product or service, despite potential opposition from those whose interests might be threatened (socio-political support).
- Technology: Any device, component of a device or process for its use that is dedicated to the production, storage and distribution of energy, or the provision of new or improved energy services or commodities to users. Where necessary for clarity, this report differentiates between “technology application” (e.g. renewable power), “technology type” (e.g. solar PV), “technology design” (e.g. perovskite cells) and “technology component” (e.g. smart inverters).
- Technology innovation: The process of improving the means of performing tasks through the practical application of science and knowledge, usually resulting in higher performing equipment as measured by, for example, energy efficiency, user friendliness or cost. This process includes learning-by-researching (R&D) and learning-by-doing, and their interaction with the technology innovation systems to which they contribute.
- Technology innovation system: The dynamic and evolving interactions of the tangible and intangible factors that determine each stage of the innovation process for a given technology. It comprises the innovators, users, institutions, financers, civil society actors and the perceptions, networks and rules that govern their actions.
- Learning-by-researching: The accumulation of knowledge by devoting R&D resources to the search for new ideas and their development into viable products and services, including prototypes and demonstration projects.
- Learning-by-doing: The accumulation of knowledge from direct experience of undertaking the activity through repetition, trial and feedback.
- Economies of scale: Cost advantages reaped in manufacturing and installation when fixed and variable costs rise more slowly than the number of units of output. It is associated with mass production of similar goods as well as the use of larger equipment, such as pipelines for which material needs do not scale linearly with throughput. Though much rarer, diseconomies of scale have also been seen (Coulomb and Neuhoff, 2006).
- Forgetting-by-not-doing: Interruptions in production or use of a technology that cause accumulated knowledge to be lost and lead to higher unit costs for the next unit put into service after the interruption.
- Spillovers: Positive externalities of learning-by-doing or learning-by-researching that increase the rate of innovation in an area that was not the target of the original innovative activity. Spillovers can be considered to be “free” inputs to parallel innovation ecosystems, related by geography or scientific proximity. Knowledge spillovers refer to the incorporation of new principles, e.g. the adoption of breakthroughs in semiconductor manufacturing by those producing solar PV. Application spillovers refer to the adoption of a technology in a new application only once it has been refined through innovation targeted at a separate, original application, e.g. the adoption of lithium-ion (Li-ion) batteries in vehicles after their development for consumer goods.
- Public goods market failure: The private sector has limited incentive to produce knowledge if firms cannot fully exploit the returns on their investment because that knowledge is easily available to others. Patents and public spending on R&D are in part a response to this market failure.
- Materiality: A threshold above which a technology is considered to have sufficient market share for its impact on supply chains to be “material”, defined in this report as 1% of national stock in a given sector. Beyond this threshold, most technologies are sufficiently mature in their design, production and familiarity for the next stage of deployment to be more straightforward.
Innovation processes are rarely linear, and no technology passes all the way from idea to market without being modified. Their trajectories are influenced by feedback loops and spillovers between technologies at different stages of maturity and in different applications, and often involve setbacks and redesign. It is nevertheless worth considering the four distinct stages through which all successful technologies eventually pass because each stage has different characteristics and requirements. These stages are relevant to all the different levels of technology definition – type, design, component – but are most applicable to technology designs.
Prototype: Following its initial definition, a new concept is developed into a design and then a prototype for a new device, a new configuration of existing devices or a new component to improve a product on the market. The probability of success at this stage is low, but the costs per project are also generally low.
Demonstration: The first examples of the new technology are introduced onto a given market at the size of a single full-scale commercial unit. The purpose is to show that the technology is effective and to reduce the perception of risk for financiers: potential customers will generally not consider a new product until it is shown to work at a profitable scale and cost. Demonstration involves more time, cost and risk than the prototype stage. This phase is often referred to as the “valley of death”, especially for large-scale, tangible technologies.
Early adoption: At this stage, there is still a high cost and performance gap compared with existing technologies, but the technology is used by customers who want to try it out or need it for a particular purpose. This period represents a continuation of the “valley of death,” and in many cases revenue from early niche markets doesn’t cover costs. In cases where governments see a broader social, environmental or economic benefit from its wider diffusion, they may help, for example through discretionary procurement or financial support. Operating in a commercial environment means, however, that more of the costs and risks can be borne by the private sector, with competition driving down costs and encouraging refinements. As the number of niches grows, the technology arrives at a material share of 1% or more of the addressable market.
Maturity: As deployment progresses beyond materiality to maturity, the product moves into the mainstream for new purchases and may even start to compete with the stock of existing assets, leading to early retirement of those assets and driving even faster diffusion. Incremental learning-by-doing continues during this stage, as feedback from engineers and users stimulates new ideas for more radical enhancements to be prototyped. Although there may still be some cost or performance gaps at the beginning of this stage, a dominant design has become accepted and the risks are generally familiar enough for private investors to bear.
Throughout the early adoption and maturity stages, innovation continues to improve the technology. In some cases, significant discontinuous improvements occur long after mainstream diffusion into the market has started, as for example with as Li-ion batteries for electric vehicles. In other cases, technologies reach a point where only very incremental changes are expected from the ongoing learning processes, as for example with large hydropower plants.
At each stage of the energy innovation journey, public and private sector actors, including not‑for-profit research institutions and funders, play essential roles. For all actors, competition is a major driver of energy innovation. Firms of all sizes, including state-owned enterprises, have incentives to refresh their offering to customers to increase market share and to avoid losing out to competitors with cheaper or better performing products. Investment funds seek new companies that can deliver the highest returns and help the funds compete for more capital. Countries also often compete to secure investment and market share for companies and workers in their countries. The same is true for subnational governments, which are playing an increasingly important role in reshaping urban energy systems.
The role of governments is particularly crucial. It encompasses educating people, funding R&D, providing network infrastructure, protecting intellectual property, supporting exporters, buying new products, helping small and medium-sized enterprises, shaping public values, and setting the overall regulatory framework for markets and finance (Hekkert et al., 2007; Bergek et al., 2008; Kim and Wilson, 2019; Grubler et al., 2012; Roberts and Geels, 2018). The essential justification for public intervention in innovation is that new ideas and technologies are undersupplied by the market – the so-called public goods market failure that leads companies to prioritise expenditures from which profits are more certain. In particular, radical new concepts, or “disruptive” technologies, often arising from basic scientific research, are rarely supplied by incumbent companies, which tend to focus on incremental improvements to their existing technology portfolio (OECD, 2015). Disruptive technologies can be of particular importance in relation to social or environmental outcomes that are desired by governments but have low market value.
While there is legitimate concern that public sector R&D might “crowd out” corporate incentives, the evidence suggests that the productivity of corporate research is increasingly dependent on ideas arising from publicly funded R&D (Fleming et al., 2019). Public funding for energy R&D may well stimulate more private sector spending, not less (Nemet and Kammen, 2007).
A mechanistic description of how governments fill gaps left by the private sector underplays their ability to make things happen. They have in the past used their powers to set incentives for, and work with, the private sector to deliver desirable outcomes: examples include space exploration, vaccines and nuclear power. It is increasingly recognised that many of the biggest clean energy technology challenges could benefit from a “mission-oriented” approach (Díaz Anadón, 2012; Mazzucato, 2018). Support for industrial clusters, strategic use of public procurement and investment in enabling infrastructure could all play a part in such an approach, increasing the probability of innovation success.
The innovation story of solar PV illustrates how concerted government action can steer and accelerate technology development while harnessing the advantages of private sector leadership (Box 1.2). This brings out the importance of government support from R&D through to the scaling-up of demand in successive niche markets, starting with the highest value and simplest applications. It also brings out how global the process of innovation can be: governments in several different countries played an important part in bringing solar PV from the laboratory to the market, responding to external events in ways that increased the chances of solar PV successfully moving along the learning curve. Importantly, although there were times when demand growth for solar PV dipped in individual countries in response to policy changes, the global market continued to grow as it was reliant on different national incentives around world.
The development of Li-ion shows some similarities (Kittner, Lill and Kammen, 2017), with Canada, the People’s Republic of China (hereafter “China”), Japan, Norway and the United States all playing a role. R&D efforts appear to have been important drivers of cost reduction, for example through the development of new cathode materials with improved specific capacity and higher share of utilised charge capacity.1 While falling cathode and anode material prices played a role, R&D also enabled the use of lower cost metals instead of cobalt.
The stories of PV and Li-ion innovation are far from finished. PV patenting activity remained far higher in 2017 than at any time before 2005, and Li-ion patents have not yet peaked: successful new components and designs are likely to make an appearance in the coming years. Their history to date, however, underlines the importance of R&D at the start of the innovation journey, and the key role of governments around the world in helping major new technologies achieve success.
Selected examples of the different roles of the main actors at each stage of the energy innovation process
|Stage||Private sector||Government||Investor community||Civil society|
|Across all stages from prototype to maturity||
Governments were critical in bringing solar PV from the laboratory to the market, stimulating early adoption and spurring continuing innovation, but no single country was instrumental.
The first demonstrations of PV cells were made in the 1950s in the United States by Bell Labs, which was granted the right to spend a certain share of AT&T and Western Electric’s operating budget on risky and basic R&D as part of its government-regulated telecommunications license. US dominance of the technology persisted through the 1970s under the supervision of the National Aeronautics and Space Administration (NASA), which had sizeable public R&D funds, and which began using PV in satellites and shuttles. The oil shocks of the 1970s spurred Japan and the United States to increase their public funding for PV research in a quest for more secure energy sources. In the United States, companies were spun off from government-regulated laboratories and found niche business opportunities for PV. In Japan, companies like Sharp were helped by the government to build production facilities and they too found market niches.
Throughout the 1980s and 1990s, PV for electricity production was uncompetitive except for off‑grid customers with a willingness to pay a high price for small amounts of power. Suppliers in the United States, then Japan and then Germany were, however, able to scale-up as a result of government procurement and incentive policies in these countries. As the potential became more apparent to researchers in more countries, R&D funding increased, the number of patents accelerated and costs fell. Of particular significance in helping to create a market were government feed-in tariff programmes, first in Germany in the 1990s, then in Italy, Spain, the United States, China and India by the 2010s. These programmes, backed by rising deployment targets, targeted grid-connected systems and provided the guaranteed scale-up needed for global supply chains. At this point, patenting peaked and the market consolidated around a dominant design.
Even though the development of solar PV to this point took around 60 years, progress would almost certainly have been slower if these countries – and others not mentioned here – had not shared the responsibility for these innovation stages (Gallagher, 2014; Nemet, 2019).
It is possible to assess the relative importance of resource push and market-pull measures by estimating the contribution to cost reductions made by different technical elements and allocating them to their high-level drivers using generalised assumptions (Kavlak, McNerney and Trancik, 2018). Technical improvements attributable to market pull measures – a combination of learning-by-researching by the private sector, learning-by-doing (repeated routine manufacturing activity) and economies of scale – are estimated to have contributed two-thirds of the cost reductions in producing solar PV panels between 1980 and 2012.2 While economies of scale contributed only around 22% of cost reductions over the entire period, they grew greatly in importance after 2001.3 It is likely that silicon prices, wafer area and factory design all benefited from developments in the semiconductor sector, indicating the importance of spillovers. Overall, this suggests that around 60% of the cost reductions arose from R&D, both public and privately funded. The incentives for R&D may have been particularly strong prior to 2001, during a period when competition between crystalline and thin film PV technologies created uncertainty about which design would dominate. The market leading technology in terms of share of global production changed twice in less than a decade before economies of scale in mass production of the more efficient crystalline technology in China reduced thin film’s market share to below 10% (Hoppmann, 2018).
The Covid-19 pandemic has delivered a brutal shock to countries around the world. By mid‑May 2020, around one-third of the global population was under full or partial lockdown. Assuming that containment measures are gradually phased out during the second half of the year, the global economy is expected to contract by at least by 3% in 2020; this would be the largest economic dip since the global depression of the 1930s (IEA, 2020). If outbreaks and containment measures last longer, there is a significant risk that the global economy could shrink by as much as 6%, with GDP contracting in nearly every country in 2020. Some low-income countries face particular pressures in dealing with the pandemic and its fallout.
As described earlier in this chapter, technology innovation is a driver of structural change. New technologies outcompete older ways of doing things and bring new services to society. This process attracts investment at each stage – from governments; high-yield, voluntary contribution funds; and, ultimately, cautious institutional investors. The evidence suggests that clean energy technology innovation brings particular economic benefits, as well as being essential for the creation of a more sustainable energy system. One study of the automotive sector finds that clean energy innovation is more productive in terms of its ability to stimulate knock-on inventions than innovation activity directed to incumbent technologies (Aghion et al., 2016). While the macro relationship between jobs and R&D expenditures is complicated, other studies suggest that R&D that supports new high-tech products is correlated with increased employment (Calvino and Virgillito, 2017). Clean energy innovation can also generate good value for taxpayers: reviews of six public clean energy R&D programmes in the United States found a return on investment of 27% since 1975, and a benefit-to-cost ratio of 33:1 (11:1 at a 7% discount rate; Dowd, 2017).
Worldwide, some 300 million full-time jobs could be lost as a result of Covid-19, and nearly 450 million companies are facing the risk of serious disruption. Clean energy innovation is labour intensive: we conservatively estimate that over 750 000 people are currently employed in energy R&D around the world, representing 1.5% of the approximately 40 billion workers in the global energy system, with half of these jobs being in China, Japan, the United States, France and Germany. If these workers are lost to the sector, it will be hard to build up the expertise associated with them again: it takes many years to acquire the specialist skills and experiences necessary to identify technology needs, formulate improved concepts and build the teams to test them. Our ability to meet the major energy challenges ahead – to develop the first zero-emissions flights or the next generation of solar panels, for example – will be enhanced if the numbers of those working on energy R&D are maintained and indeed increased.
There are several ways in which clean energy innovation jobs and outputs are threatened by the Covid-19 pandemic. These include pressures on public and private budgets, a riskier environment for clean energy venture capital and disrupted global supply chains (see Chapter 2). Public R&D is expected to hold up better than private R&D, and there is a reasonable chance that the governments of major economies will seek to boost innovation funding as a response to the crisis. Companies face lower revenue and a lack of cash flow for capital investments to meet near-term growth targets, but there is little sign of those who have made commitments to reduce their emissions intensity and test new energy technologies seeking to back away from those commitments. For a rapid assessment of the likely impacts of Covid-19 on their ability to support innovation towards longer term goals, we surveyed industrial contacts in May 2020. Responses indicated no change in long-term commitments and an expectation that R&D budgets would be resilient, but overall sentiment about the impact on the full range of innovation activities was gloomy (Box 1.3).
The second half of 2020 presents a unique opportunity to double down on clean energy innovation. While near-term responses to the crisis have understandably focused on mitigating health, employment and liquidity risks, attention is now turning to the speed of the recovery, the creation of new jobs and the future shape of the economy. New players with new ideas aiming to displace high-carbon producers and to scale-up quickly may find a supportive environment if they are able to enter the market at the right moment. Economic stimulus plans now being proposed in countries around the world offer a once-in-a-generation opportunity to boost clean energy technology innovation. Many of the sectors that are critical to achieving net-zero emissions have investment cycles of many decades, so there is no time to lose.
In May 2020, the IEA contacted a number of large companies that are active in the development of technologies expected to play a significant role in the achievement of net-zero emissions, focusing on four specific technology areas: 1) direct electrification; 2) hydrogen; 3) carbon capture, utilisation and storage (CCUS); and 4) digitalisation. We included end-user companies outside the energy sector, including companies from the iron and steel, cement and chemicals sectors. The 28 companies that responded represent nearly 1.5 million employees worldwide.
The responses indicate serious disquiet among experts about keeping their innovation pipelines flowing over the next couple of years. Most respondents think it is at least “somewhat likely” that all elements of their R&D, demonstration and deployment strategies will be affected. Companies that are prioritising technologies for the electrification of energy demand, especially those in heavy industry, consider it likely that their R&D budgets will be considerably or significantly reduced. Companies pursuing CCUS consider it very likely that public budgets and grants for these technologies will be more uncertain and possibly reduced.
Expected impacts of Covid-19 on clean energy innovation from corporate experts in May 2020
|Reduced R&D budget|
|Uncertainty about public R&D budgets and grants|
|Reduced number of staff related to R&D activities|
|Reduced support to energy start-ups|
|Demonstration projects: Reduced CAPEX|
|Demonstration projects: Delays|
|Reduced participation in JVs for new technology development|
|Slowdown in adoption of recently commercialised technologies|
|Disruptions in supply chains hindering energy R&D|
|Likelihood:||Very likely:||Likely:||Somewhat likely:||Unlikely:|
Taking account of perceptions of risk in terms of both the magnitude of the impact and its likelihood, the highest levels of unease are focused on the demonstration and early adoption stages of the innovation process. There is unease in particular about the stability of public R&D funds, which are generally sought by corporations for testing in the field; the ability to execute large-scale demonstration projects; the resilience of collaborations; and a slowdown in adoption of new clean energy technologies. While these results take account of firm size, the sample of responses shows more concern among smaller firms, with larger firms indicating a higher expectation of avoiding significant cuts to R&D budgets.
A positive message from many respondents was that their strategic priorities for clean energy technology development will not change. Respondents also expressed little change in their appetite for risk-taking in their priority technology areas. If the flow of funding can be maintained and policies are supportive of growing demand for the technologies, then major companies seem likely to be ready to continue to support innovation.
Ziegler, M. and J.E. Trancik (2020), Personal communication on 1 March 2020, Massachusetts Institute of Technology.
Including module efficiency, non-silicon materials costs, wafer area and yield.
Including some non-silicon materials costs, plant size and silicon price.
Ziegler, M. and J.E. Trancik (2020), Personal communication on 1 March 2020, Massachusetts Institute of Technology.
Including module efficiency, non-silicon materials costs, wafer area and yield.
Including some non-silicon materials costs, plant size and silicon price.