In the year 2000, the International Energy Agency made a prediction that would come back to haunt it: by 2020, the world would have installed a grand total of 18 gigawatts of photovoltaic solar capacity. Seven years later, the forecast would be proven spectacularly wrong when roughly 18 gigawatts of solar capacity were installed in a single year alone. Ever since the agency was founded in 1974 to measure the world's energy systems and anticipate changes, the yearly World Energy Outlook has been a must-read document for policymakers the world over. Over the last two decades, however, the IEA has consistently failed to see the massive growth in renewable energy coming. Not only has the organization underestimated the take-up of solar and wind, but it has massively overstated the demand for coal and oil. Jenny Chase, head of solar analysis at BloombergNEF, says that, in fairness to the IEA, it wasn't alone. "When I got this job in 2005, I thought maybe one day solar will supply 1% of the world's electricity. Now it's 3%. Our official forecast is that it will be 23% by 2050, but that's completely underestimated," Chase says. "I see it as the limits of modelling. Most energy system models are, or were, set up to model minor changes to an energy system that is run on fossil fuel or nuclear. Every time you double producing capacity, you reduce the cost of PV solar by 28%. We've got to the point where solar is the cheapest source of energy in the world in most places."