Has the clean energy transition stalled and what should the UK government do?

Despite the gloomy news reports on backtracking from net zero targets while wildfires rage in parts of Europe, innovation, declining costs and growing deployment of renewables and batteries provide much to be positive about. The UK may face particular energy infrastructure challenges but Dimitri Zenghelis urges the Government and its new ministers to act on the facts, not the headlines.

Recently it has become fashionable to be pessimistic about humanity’s ability to reduce greenhouse gas emissions and wean itself off fossil fuels. Emissions continue to climb, net zero targets have been loosened in the EU and jettisoned in the US by President Trump, banks have withdrawn from initiatives like the Net-Zero Banking Alliance, and the fossil fuel industry is talking up new reserves.

The science of climate change remains impartial

Of course, the underlying science and laws of physics behind climate change remain unchanged, and this summer is bringing this home in the northern hemisphere. At this point in late July 2026, a run of successive unprecedented heatwaves have left Europe tinder-dry. Fires have caused over a dozen fatalities in Spain and forced more than 320,000 people from their homes across Spain and France alone (see Figure 1). Political rhetoric can loosen a target but it cannot dislodge a heat dome.

Figure 1. Europe: active fires and recently burnt areas, 27 July 2026

Source: https://forest-fire.emergency.copernicus.eu/ using MODIS, VIIRS and SENTINEL3 data

The pace of innovation is remarkable

Contrary to the pessimistic headlines, the energy transition has moved faster than anyone expected because the economic logic is undeniable. In 2025 renewables overtook coal in global electricity generation for the first time in the modern era, with solar, wind, hydro and other renewables together contributing more than one-third of global generation. Electric vehicle sales reached 21 million units worldwide in 2025, with one in four new cars sold globally now electric (in China, it’s more than half). Even the EU, following a stagnant 2024, saw sales rise 30%, overtaking China as the fastest-growing major market.

For all the fog, the acceleration in electrification and renewables is hiding in plain sight, obscured only by political noise – but it was ever thus. Oil is a fungible, storable commodity; no competent oil executive will talk up the prospects for renewables, since that would hit the price of oil today.

Far from slowing down, the drivers of the clean revolution are still gathering pace without significant policy impetus. Two enduring wars, both involving major fossil fuel producers, have only sharpened the drive towards secure, efficient alternatives with energy independence and climate action now pointing in the same direction as efficiency. China and India, the world’s largest coal burners, saw fossil generation fall in 2025 for the first time this century, as record clean power additions pushed low-carbon growth above demand growth. A global peaking in fossil fuel use is probably a year or two away.

AI’s energy problem is real, but not the one the headlines suggest

AI and automation, with their immense appetite for data centres and compute, are often cast as a threat to emissions targets. The honest picture is more mixed.

Water is a genuine constraint, though the popular framing of it is often wrong. Unpurified water consumed upstream by the gas, nuclear and hydro plants generating a data centre’s electricity is roughly 12 times the water used directly for on-site cooling. In the US data-centre water use is a mere 0.3% of public water supply, but that average hides real local strain: roughly two-thirds of new hyperscale campuses built since 2022 sit in high water-stress counties.

Data centres are still substantially fossil-powered. Globally, around 56% of the energy consumed by data centres comes from fossil fuels – roughly 30% coal and 26% natural gas – with renewables at only 27% and nuclear at 15%. This is mostly a legacy of the incumbent system, though coal-plant closures are being delayed and new gas turbines are being built to meet AI load.

However, the likelihood is that AI will accelerate rather than derail electrification and renewables in the long run. Compute costs are converging on energy costs and here China, with cheap, increasingly renewable electricity and a commanding lead in solar, wind and battery manufacturing, has a head start over America.

Renewable generation for data centres is expected to meet between half and two-thirds of the sector’s anticipated demand growth to 2035, but the dynamics of innovation suggest this understates the potential. Battery storage costs fell 45% in 2025 alone, with deployment up 46%, precisely the cost curve that lets renewables absorb new AI-driven demand rather than cede ground to gas and coal. Accelerated deployment will move the cost of renewables faster than the cost of gas.

Ideas beat extraction on cost

This is because fossil fuel extraction faces diminishing returns: more extraction means exploiting more remote reserves employing more exotic techniques, while manufacturing enjoys increasing returns, as innovation continually cuts costs.

Falling costs drive more deployment, which drives costs down further: a reinforcing loop that creates economies of scale and tipping points. Wright’s Law holds that every doubling of deployment of modular, scalable, replicable technologies cuts unit costs for solar PV, wind and batteries by 20–40% – the kind of feedback that, like a microphone before a speaker, produces change faster than outdated narratives can track.

The impact is profound. The cost of final energy from solar PV has fallen by 99.9% since it was first deployed commercially in satellites in the 1950s (see Figure 2). Fossil fuel costs, by contrast, have barely shifted, not because we burn these fuels less efficiently, but because an extractive system relies on costly labour and logistics at every cycle: once burned, more must be dug up, transported, refined and burned again.

Figure 2. The long-run cost of energy

Source: Way et al. (2022)

Even intermittency is becoming manageable. Battery storage is doubling each year and turning cheaper than fossil backup, while efficiency gains, demand-response management, diversified grids and storage options (batteries, pumped storage, hydrogen) further ease the challenge.

It is no wonder the world now invests twice as much in clean energy for electricity, transport and buildings as in fossil fuels. Around 90% of global electricity-generation investment goes into renewables. The new system’s capital expenditure-heavy intensity, and the need for grid expansion, is no different from the upfront investment railways, electricity, the internet or AI itself required. These headline numbers are the cost of improving the system’s productivity, not a resource cost to curtail.

The opportunity from electrification is greater still than that from renewables alone. End-use products – batteries, electric vehicles, heat pumps – carry higher margins than solar or wind generation itself. China, now accounts for almost one-third of the world’s electricity capacity (the US only 15%) and powers almost one-third of its energy use through electricity, versus about 23% in Europe and 21% in the US.

Fortunately, the electrotech and AI revolutions are complementary and reinforcing. Knowledge-based energy aligns naturally with AI: smart controls improve grid efficiency, optimise dispatch and help manage intermittency, while clean lab solar manufacturing shares the silicon wafer fabrication that underpinned the chip industry. A rooftop solar panel has more in common with a laptop than a boiler. Whoever builds the cheapest clean electrons builds the cheapest AI. It is this real race, not political rhetoric, that is deciding the pace of the energy transition.

Challenges for the UK

The UK’s challenges and structural problems remain greater than most. Following the arguably too-rapid closure of domestic fossil supply, there is currently no low levelized cost renewable option available to pull wholesale prices sharply lower. Offshore wind remains relatively expensive, and onshore wind and solar have been supply-constrained. Solar is less cost-effective in cloudy north-west Europe; gas prices are volatile day to day, but on current learning curves solar can be expected over the coming years to outcompete complex liquified natural gas with its liquefaction, transport, storage and regasification costs, and – once installed – a renewable asset generates cheap power for decades, reducing exposure to volatile import prices.

System costs remain high in a crowded country with strict planning laws, where erecting pylons is not straightforward, compounded by underuse of batteries and smart-meter demand response (both would ease pressure on grid expansion), capacity payments and curtailment (paying operators not to supply). The risk is a vicious spiral: costs rise, demand growth is disincentivised, and the case for grid investment weakens in turn.

But the clean transition is not stalling: it’s compounding. Every heatwave makes the case for urgency; every supply crunch makes the case for resilience and every fall in the cost of solar and battery deployment makes the case for energy economics. Electrification and renewables will define the energy system of the 21st century, and their costs will keep falling. The strategic choice is stark: invest now in the grids, storage and generation this future demands, or lock into a fossil infrastructure that is obsolete and stranded. The Government must act on physics, wars and cost curves and not headlines.