Fossil fuel use could peak by 2020 as renewables become cheaper, if policy follows

According to fossil fuel companies, the world will continue to rely on their products for decades. They even have sophisticated scenarios, outlooks and modelling to prove it.

What if they are wrong? New analysis published today by the Carbon Tracker Initiative and Imperial College’s Grantham Institute suggests global demand for fossil fuels could peak by 2020. The power sector would see the most dramatic change, becoming virtually fossil-free by 2040.

This fossil fuel “demand destruction” would be hugely disruptive for incumbent industry business models, the report’s authors say. Carbon Brief has two graphs to summarise their findings.

Changing assumptions

Oil majors, including Shell and BP, have long published outlooks showing why they think the need for oil and gas will keep growing. Last week, for instance, the annual BP Energy Outlook argued oil demand would continue to grow, despite the rise of electric cars.

Coal firms have also relied on such outlooks to justify new investment. However, these outlooks are frequently criticised for underestimating the pace of change in the energy industry or for downplaying the likelihood of climate action. The International Energy Agency (IEA) has been in the firing line too, with renewable forecasts that keep falling short.

At the heart of the problem lie complex models of the global energy system. These models are set up with a long list of assumptions: How quickly will economies expand? How many people will inhabit the world in 2050? What will it cost to build a wind turbine or solar panel in 2035?

If you change these assumptions, you can radically alter the models’ outlook for energy.

Peak fossil

It’s one thing to criticise the likes of BP, Shell or the IEA for their assumptions. Carbon Tracker and Grantham have gone a step further, modelling the impact on fossil fuel demand of using different assumptions instead.

Their stand-out finding is that global fossil fuel demand could peak by 2020, if clean technologies get cheaper more quickly than expected, and if climate policy is ramped up.

Historical global fossil fuel use and the outlook to 2050 (shaded areas) if climate policy is ramped up and costs for solar and electric vehicles fall quickly. Source: Expect the Unexpected, Carbon Tracker Initiative and Grantham Institute; BP Statistical Review of World Energy 2016. Chart by Carbon Brief using Highcharts.

In this scenario, coal use would peak by 2020 and then fall to 60% below 2012 levels by 2050. Oil demand would also peak by 2020, ending up 8% lower in 2050. Gas demand would continue growing until 2030, but it too would then peak and decline.

It’s worth emphasising that the new study includes a range of scenarios, each with different assumptions that are laid bare in the report. The authors say this transparency is vital if energy outlooks – and the business cases that rely on them – are to be interpreted and understood.

The scenario in the chart above assumes solar will cost $1.1-1.7 per watt by 2020 and $0.6-0.9/W by 2030. For context, in the US, utility-scale solar costs have already declined from $4.46/W in 2009 to $1.42/W in early 2016. Some forecasts see these costs falling below $1/W by 2020.

The scenario assumes battery electric cars, vans and buses will be as cheap as, or cheaper than, standard models by 2020. For comparison, Bloomberg New Energy Finance says electric vehicles (EVs) will be cheapest by 2022.

The 2020 cost assumed in the scenario is about the same as the current cost the recently-launched Chevy Bolt, says Dan The US associate professor in the Department of Civil and Environmental Engineering at Rice University. He notes the Energy Information Administration recently made “dramatic” downwards revisions to its forecast for EV costs.

The scenario says the figures are in line with the most up to date information and “low end market projections”. These are different to the assumptions behind most models, but still plausible. “Our lower cost assumptions may prove to be conservative,” the study notes.

It represents “strong” climate policies using a global carbon price of $50 per tonne of CO2, rising by 5% per year. Existing climate commitments would be roughly equivalent to a lower level of $30/tCO2, the study says. Note that models typically use carbon pricing as a proxy for the impact of all other climate policies, which are hard to model directly.

Power shift

The most dramatic shifts in this cheaper clean technology and stronger climate policy scenario would be seen in the power sector, as the chart below shows. (Note that it only shows coal, gas, wind and solar).

Global Electricity generation from coal, gas, wind and solar to 2050, under a scenario with stronger climate policy and cheaper solar. The chart does not include hydro, nuclear or other fuels. Source: Expect the Unexpected, Carbon Tracker Initiative and Grantham Institute. Chart by Carbon Brief using Highcharts.

Under these assumptions, coal use in the power sector would already have peaked and would fall two-fifths by 2030. It would then fall to zero by 2040. Gas use would increase to 2020 and then fall rapidly, declining three-quarters by 2040.

The study finds virtually no role for coal or gas with carbon capture and storage (CCS) in the power sector, because renewables are assumed to be cheaper. CCS could still be important for industry, the study suggests.


Energy firms have been criticised for being slow to adjust their outlooks in response to changing circumstances. They say that the world will continue to rely on coal, oil and gas for decades to come and that the energy sector, like an oil tanker, is slow to change course.

Today’s research suggests that they could be wrong, and that global demand for fossil fuels could soon start to decline. It does so by using different, but still plausible assumptions about the future path of technology costs and climate policy.

You can explore a range of assumptions, and the implications for fossil fuel demand, in an interactive tool developed by the team.