Lower prices: Our main prediction was that, with specific exceptions like Henry Hub, 2025 would be a “year of falling prices” across the energy spectrum. Oil and gas benchmarks did indeed fall more or less as predicted, as did power prices across Europe, in broad correlation with gas. Solar and battery unit costs also continued their decline. 

While energy security and affordability were dominant criteria for energy financial flows, sustainability remains an indispensable factor. These three points of the classic “energy trilemma” increasingly overlapped. As the global energy system steadily but unmistakeably shifts from fossilised to electrified, smart capital is being allocated to sustainable technologies because they build greater resilience and long-term value. 

Solar: slightly over-estimated. We forecast 700 GW global capacity additions, “bringing total installed capacity towards the 3 TW mark” - slightly ambitious versus the actual 655 GW. Cumulative installed provisionally looks to have reached ~2.85 TW, so last year saw an impressive 30% increase in solar generation capacity. (For context, in 2017 the world installed around 100 GW of solar, so we have seen 6.5x growth in annual installations over just 9 years.) 

Batteries: slightly under. Predicting a breakthrough year for batteries was a no-brainer, with new projects getting consistently bigger and cheaper. Grid-scale and behind-grid battery energy storage systems (BESS) hit 315 GWh of installations (up ~50%), slightly above our 300 GWh expectation. Lithium-ion battery shipments rose by a similar rate to 2.28 TWh, of which 1.5 TWh was for EVs - the rest for stationary storage. So there is no shortage of capacity available for rapidly expanding new installations. BMI reports operational giga-scale projects increasing from 17 in 2024 to 46 in 2025 and as many as 150 more underway. We were however slightly over-optimistic on lithium-ion battery pack prices, which only reached $108/kWh rather than our $100/kWh target. 

Wind: very under. We forecast that global capacity additions would be 100 GW, for a total installed capacity of 1.2 TW. The Global Wind Energy Council has now confirmed actual installations almost certainly exceeded 150 GW, for a cumulative total of ~1.32 TW. Approximately 100 GW of this was in China, but other regions grew fast too, with European installations rising 30% to 22.5 GW. Wind is a more mature sector than solar or batteries, but at a global level it continues to grow at a very solid rate, with 2025 installations just under 9% up on 2024. 

Grids: slightly under. Our forecast of $450 billion investment was around $30bn too low. BNEF has reported a provisional actual of $479 billion. That 16% growth in 2025 partly reflects price inflation given ongoing supply chain shortages for key grid equipment. We expect to see continued double-digit growth in grid investments over the coming years, providing further tailwinds for critical equipment manufacturers.  

EVs: slightly under. Our global EV sales forecast of 20 million (with 22 million as an upside case) was marginally below the actual outturn of 20.7 million, (20% growth vs 2024). What is clear is that EV adoption (including battery electric vehicles and plug-in hybrids) is scaling fast across many markets. Not least in China, by far the world’s biggest car market, where they reached 54% of new sales at 12.9m units sold, up 18% on 2024. We take no great credit in correctly predicting that BYD would take global output top spot, amid red-hot competition from rival Chinese manufacturers like Geely, Chery, XPeng and even mobile-phone maker Xiaomi, and that Tesla would slump down the charts. European EV sales actually grew faster than China, at 33%, albeit off a much smaller base, while the Rest of World (ex-China, Europe and N. America) grew by nearly 50%. 

Gas & LNG: slightly over. We expected average 2025 realised prices of TTF €40/MWh, NBP 100p/therm, Henry Hub $3.40/MMBtu and JKM $13/MMBtu. Given how choppy and unpredictable gas markets can be, these were directionally right and proved reasonably accurate, with outturns ~10% lower than forecast. TTF averaged just over €36/MWh from a high of €58.26 in February down to a low of €26.51 in December - a fall of over 50% from peak to trough. Likewise, UK NBP averaged just under 90p/therm. We were also directionally correct on Henry Hub, which rose to average slightly higher than forecast at $3.64/MMBtu. Howver, gas prices proved less volatile than we expected, particularly over Q2-3. (For the uninitiated, gas prices are somewhat awkwardly priced in different units and currencies around the world, making them harder to compare. For more see our Price Explainer.) 

Oil: about right. We forecast that strong supply along with tepid Asian demand “should pull Brent from its early year highs around $80/bbl to below $60/bbl.” That is essentially what happened, with Brent trending down to a low of $58.86 by mid-December, before ending the year marginally above $60. If we had remembered to give an average Brent prediction, on that basis it would have been around $65/bbl. The actual Brent front-month average for 2025 was $68/bbl. We also forecast that global gasoline demand would peak in 2025 - this looks likely but has yet to be confirmed. 

Coal: slightly over. Our global demand forecast of 9 billion metric tonnes was slightly above the provisional total of 8.85 billion MT. In fact, in mid-December, the International Energy Agency (IEA) stated that 2025 is likely to prove the peak, before gradual annual declines over the remainder of this decade. CREA also reported that coal generation dipped in 2025 in both China and India, which strongly reinforces the IEA view. It now looks like coal demand will never reach 9 bn MT after all. 

Carbon: slightly under. Our EUA price forecast of €65-70/MT proved a little bearish compared to the actual average of ~€75/MT. 

AI: about right, albeit not that SMART. We predicted AI will get more energy efficient, mitigating fears around its energy demand. A month later, the Chinese DeepSeek AI model hit the headlines for doing exactly that. We also predicted AI will also help to optimise infrastructure performance and boost broader energy efficiency, but this looks like a longer-term dynamic. We also expected rapid growth in AI demand to drive new capacity investment across renewables, gas, nuclear and hybrid power projects. This certainly transpired, particularly in the US, where there was a strong correlation between electricity demand growth and solar capacity growth. However, AI remains a more marginal driver of new electricity capacity investment across the rest of the world versus industrial, heating, cooling and other core demand drivers.  

UK nuclear: half right, half wrong. We forecast Sizewell B’s 2035 decommissioning date to be extended and Hinkley Point C’s start up date both to get deferred. The former did not happen, but the latter was pushed back to 2031, only 2 days after our 2025 forecasts were published.  

UK power mix: slightly under. Our 265 TWh total consumption forecast was very slightly below the actual 267.64 TWh (a 1% increase over 2024) - based on data from Drax Electric Insights. Stable demand over 2025-26 after years of decline implies that electrification is now counterbalancing efficiency and deindustrialisation effects, so UK electricity consumption may now start trending back up. 

The breakdown by type proved broadly similar to 2024, with nuclear the most notable change. We were a little optimistic on wind versus gas output, and therefore also on emissions intensity, which barely changed at 120 gCO2e/kWh versus our 100g estimate. The final phaseout of coal in September 2024 made no material difference. 

Solar did increase, with 2025 being the UK’s sunniest year on record and more capacity in place, but remains only 6.4% of total supply.