Accurate technology cost assumptions are a cornerstone of effective energy transition planning. While we have been blessed with rapid cost declines in renewable energy technologies, most energy models and policy analyses continue to rely on outdated assumptions.
A new study by researchers from the University of Canterbury, LUT University, the German Aerospace Center (DLR), and the International Renewable Energy Agency (IRENA), “Are we too pessimistic? Cost projections for solar photovoltaics, wind power, and batteries are over-estimating actual costs globally”, published in Applied Energy, systematically reviewed 40 studies and 150 long-term scenarios for renewable energy technologies. It compared their projected costs (all inflation-adjusted to 2023 USD) with real-world market data and found a consistent trend: cost projections are too pessimistic.
Solar PV and battery costs are dropping faster than a rock
For solar PV, the backbone of the energy transition, the gap between projected and real costs is particularly striking. Most studies estimate that utility-scale PV will cost between $160-630 per kW by 2050. However, today’s global average is already around $500 per kW, and can be even lower as documented by the latest Trends Report from IEA-PVPS. The future, it seems, arrived decades early. Despite various attempts in the past to point out the risk of underestimating renewable energy technologies and, in particular, solar PV, such as by Victoria et al., Jaxa-Rozen and Trutnevyte, and Xiao et al., this issue persists across energy scenarios.
This cost overestimation matters. In many energy system models, these inflated costs make solar PV look less competitive than fossil fuels, leading to more conservative policy pathways and underinvestment in grid integration and storage. The same is true for battery energy storage, where today’s costs are already lower than most 2030 projections.
Why projections remain conservative
The issue lies less with technology performance and more with the assumptions underpinning the models. Many cost projections still rely on outdated assumptions and inconsistent regional factors, such as learning rates and soft costs, which can misrepresent the true potential for CAPEX reduction.
A common limitation is the use of uniform discount rates across countries. This overlooks region-specific financing conditions, risk profiles, and soft-cost structures that heavily influence the actual economics of renewable energy projects. In reality, the cost of capital for a solar PV plant in Germany differs substantially from that in India or Chile, yet many models still assume a single global cost of capital value. Challenges in energy scenarios may be less about identifying the right cost of capital for the present, and more about how to best project it into the future, often decades ahead, as has been debated by researchers.
While learning rates for solar PV and batteries tend to remain relatively stable over time, the number of cumulative doublings, and therefore the scale of cost reductions, is often set too conservatively in forecasting exercises. Such simplifications can distort levelized cost of electricity (LCOE) estimates, sometimes making renewable energy options appear more expensive than they will likely be, and at other times overlooking conditions where costs could fall even lower (particularly, in high-deployment, low-risk markets). The result is a systematic bias that narrows the expected cost trajectory and underplays the real pace of technological progress.
Evidence from updated datasets
Encouragingly, newer datasets are catching up. The U.S. National Renewable Energy Laboratory’s (NREL) Annual Technology Baseline, for instance, has cut its 2050 cost projection for utility-scale PV by over 50% between the 2015 and 2024 editions. Yet even these improved estimates still lag behind actual market trends.
A similar pattern can be observed in onshore and offshore wind power. While projections acknowledge declining costs, they often underestimate the steepness of that decline. Offshore wind power in particular remains volatile, with higher capital costs due to infrastructure and supply chain complexity, yet its long-term learning potential remains strong given its considerable deployment potential along densely populated coastlines worldwide.
Image: Sustainable Energy Research Group, University of Canterbury
Implications for energy system modelling
Cost projections are not just academic exercises; they directly shape how national and international models plan and prioritize future power systems and determine where billions in investment are directed.
When models rely on conservative assumptions, renewable energy appears less competitive, and complementary technologies such as green hydrogen, power-to-X fuels, chemicals, and materials, and electric mobility seem less viable. This, in turn, can mislead policymakers and investors' decision-making toward higher-cost, slower transition pathways.
In practice, underestimating future cost reductions does not simply misrepresent technological potential; it can translate into billions of dollars in avoidable system costs. By delaying renewable energy deployment, countries risk locking in fossil fuel-based infrastructure, missing opportunities for cheaper clean energy, and increasing long-term energy prices for consumers.
Outdated cost assumptions also risk delaying the deployment of technologies that are already cost-competitive today. In other words, they not only make the transition slower, but also more expensive.
A call to refresh assumptions
The analysis shows that real-world innovation consistently outpaces projections. Solar PV modules, batteries, and related technologies are improving faster, becoming cheaper, and scaling up more rapidly than expected. To ensure models and policies remain credible, cost databases must be updated frequently and transparently, incorporating region-specific discount rates and empirically grounded learning curves.
Otherwise, we risk designing a future that is already obsolete. Markets are ready to build tomorrow’s energy system today, and cost forecasts must reflect that momentum. The energy transition is not waiting for overly slow projections to catch up. The message is clear: models must evolve as fast as the technologies they seek to predict.
Looking ahead
The pace of cost declines in solar PV, wind power, and batteries keeps surpassing expectations, driven by global deployment, maturing supply chains, and improving financial conditions. The question is no longer whether these technologies will get cheaper but how fast they will get cheaper and whether the applied models and policies can keep up.
To avoid locking in outdated assumptions, cost databases must evolve as quickly as the technologies they describe: updated frequently, adapted for region-specific conditions, and reflecting the real-world learning already evident in global markets.
Conservative estimates may feel cautious, but in practice, they risk delaying investment, slowing electrification, and raising long-term system costs, becoming a costly hesitation. Accurate, dynamic assumptions are essential to plan a future that matches the speed of innovation. The energy transition is already here, as documented by more than 90% of all newly installed power capacity being contributed by renewable energy technologies.
Authors: Hadi Vatankhah Ghadim, Jannik Haas, Belén Silva Cardenas, Dominik Keiner, and Christian Breyer
This article is part of a monthly column by the LUT University.
Research at LUT University encompasses various analyses related to power, heat, transport, industry, desalination, and negative CO2 emission options. Power-to-X research is a core topic at the university, integrated into the focus areas of Planetary Resources, Business and Society, Digital Revolution, and Energy Transition. Solar energy plays a key role in all research aspects.
The views and opinions expressed in this article are the author’s own, and do not necessarily reflect those held by pv magazine.
This content is protected by copyright and may not be reused. If you want to cooperate with us and would like to reuse some of our content, please contact: editors@pv-magazine.com.
Agree that our models must evolve as quickly as the technologies they aim to predict