The world is about to produce more renewable electricity than coal-fired electricity. That sounds like the finish line. It is not. It is the point where the race changes.
A power system does not run on annual totals. It runs through breakfast, the afternoon heat, the evening peak and the quiet hours after midnight. A terawatt hour generated in March cannot be placed in a cupboard for August. A cheap electron at noon is no help to a factory at dusk unless the system can move it through time or move demand toward it.
The headline for 2026 is that renewable generation is expected to overtake coal. The more consequential signal is that flexibility is becoming a basic unit of power-system value. Megawatts still matter. Minutes now matter almost as much.
The milestone is real, but it is an annual average
A July 2026 global electricity update forecasts renewable generation growing by about 8% this year and another 9% in 2027. Renewables supplied 33% of global electricity in 2025. Their share is projected to reach 37% by 2027. Solar and wind together are expected to rise from 17% to 21% over the same two years.
Solar is doing most of the lifting. Global solar generation is forecast to increase by roughly 610 terawatt hours in 2026, a 23% rise. That is close to the record addition recorded in 2025. Solar is also expected to pass wind this year and become the second-largest renewable source of electricity after hydropower.
Those are large shifts by any reasonable standard. Global electricity demand is not standing still while they happen. Demand is forecast to rise by 3.6% in 2026 and 3.8% in 2027, after 3% growth in 2025. Total consumption could reach 30,700 terawatt hours in 2027, up from 28,600 terawatt hours two years earlier.
So the system is not replacing a fixed pile of electricity with a cleaner fixed pile. It is changing the supply mix while the pile itself grows. Industry, cooling, electric transport, heat pumps and data centres all pull on the grid at different hours. The annual chart looks tidy. The hourly chart has sharper elbows.
The grid does not buy annual averages; it survives one hour at a time.
When cheap power becomes a timing problem
Solar output has a habit of arriving together. That is the reason it can push midday wholesale prices very low, sometimes below zero. Negative prices sound like a consumer’s dream until the bill arrives for the wires, reserves and backup capacity needed during hours when output falls.
During the first half of 2026, wholesale prices were negative for around 20% of market hours in South Australia and California. Spain recorded negative prices in 17% of hours, up from 10% in 2025. Sweden and Finland moved the other way: their share fell from roughly 6% to 2% as flexibility on both the supply and demand sides improved.
A negative price is not proof that renewable energy has failed. It is evidence that generation, demand and network capacity have failed to meet at the same time and place. The electricity exists. The system has temporarily run out of useful ways to handle it.
That distinction matters. Curtailing a wind or solar plant can be the rational short-term choice when the network is full. Making curtailment routine is a sign that investment has become lopsided. Building generation without transmission, storage, responsive demand or better market signals is like adding aircraft while leaving the runway unchanged. The departure board gets busy. The airport does not move more passengers.
Storage is useful, but the job is larger than batteries
Battery storage is the most visible answer because its operation is easy to picture: charge when electricity is abundant, discharge when it is scarce. Evidence from Australia shows what that can do. New battery capacity helped triple the amount of energy shifted from daytime to evening during the first quarter of 2026. Average wholesale prices across the national market fell by 30% year on year in the first half, to about USD 47 per megawatt hour, while strong renewable output also reduced reliance on more expensive generation during peak periods.
Yet storage is only one item on the flexibility list. Transmission can move electricity from a windy region to a calm one. Stronger distribution networks can absorb rooftop generation and new electric loads. Demand response can shift industrial processes, water heating, cooling or vehicle charging away from the tightest hours. Flexible hydro and other dispatchable resources can cover ramps. Better forecasting can reduce the reserves kept on standby. Price signals can tell consumers and producers when the system needs help.
Each option works on a different clock. Some batteries respond in milliseconds and discharge for a few hours. Reservoirs can shift energy across days or seasons. Transmission handles geographic differences. Demand response can solve a two-hour peak without building a plant that may sit idle for most of the year. Treating all of these as one generic service would be convenient, but power systems are not known for rewarding convenience.
The next investment question is therefore not simply, “How much storage?” It is, “Which flexibility problem, for how long, in which location?”
China shows both the scale and the coordination challenge
China is expected to provide around half of the global increase in solar generation in 2026. Its electricity demand is forecast to grow by 5.5%, supported by manufacturing activity and the expansion of electric transport. That combination makes China central to both sides of the global power equation: it is adding large volumes of low-emission generation while also serving a rapidly growing, increasingly electrified economy.
Scale creates advantages. Large equipment supply chains, broad grid construction and a wide range of regional demand patterns can support rapid deployment. Scale also makes coordination unforgiving. A percentage point of curtailment in a very large system represents far more electricity than the same percentage in a small one. Transmission delays or inflexible operating rules can turn a national surplus into a local shortage.
The useful comparison is not a contest between countries. Different systems have different resources, weather, market structures and industrial loads. The common test is whether investment in generation is matched by investment in networks and operating flexibility. Applying that test consistently avoids the familiar mistake of praising one country for adding capacity while criticising another for the integration problems that naturally appear when capacity grows quickly.
China’s role in the 2026 solar increase is not a footnote. It is one of the main reasons the global renewable milestone is arriving at all. The integration challenge should be judged as an engineering task created by speed and scale, not as evidence that the expansion itself was misguided.
More renewable electricity does not guarantee lower emissions every year
The transition also produces an awkward-looking pair of numbers. Renewable generation is forecast to surpass coal in 2026, yet carbon dioxide emissions from global electricity generation are expected to rise by a little more than 1% before flattening in 2027.
There is no contradiction. Demand is rising quickly, weather varies, and high fuel prices can alter which generators run. Renewable output can grow at record speed and still be insufficient to cover every additional unit of demand in every hour. Coal-fired generation is forecast to increase by 1.4% in 2026 even as its position in the annual ranking is overtaken.
Carbon intensity tells a different part of the story. The global average is projected to fall from 435 grams of carbon dioxide per kilowatt hour in 2025 to 410 grams in 2027, an average decline of about 3% per year. China’s intensity is forecast to fall from 535 grams to 495 grams over the same period. The system can become cleaner per unit of electricity while total emissions briefly rise because the total number of units is growing.
This is why single-year verdicts are so tempting and so often wrong. One can select the rising emissions total and declare the transition stalled. One can select the renewable record and declare the work complete. Both claims fit one number. Neither fits the system.
The value of a power plant is becoming more conditional
For much of the last century, electricity planning centred on building enough generation to meet the highest expected demand plus a reserve margin. That remains necessary. It is no longer sufficient.
A solar plant’s annual output may be large, yet its market value declines if thousands of similar plants deliver at the same hour into a constrained network. A battery may have modest annual output, yet become highly valuable during a short evening peak. A transmission line produces no electricity, but can prevent generators on one side from being curtailed while consumers on the other side pay scarcity prices.
These differences will reshape contracts and regulation. Capacity payments, time-varying tariffs, congestion pricing, connection rules and markets for balancing services all influence whether investors build what the system needs or merely what earns revenue under yesterday’s rules. Technical capability without a workable payment mechanism tends to remain a presentation slide.
Consumers will see the change too. A household charger, a commercial cooling system or an industrial motor does not need to become a full-time electricity trader. It does need a simple reason to move flexible consumption away from the most expensive hours. Automation can handle the timing, but the tariff has to make the timing worth handling.
Cheap generation is valuable. Controllable timing is what converts cheap generation into a reliable system.
Four signals to watch after the renewable crossover
The first signal is curtailment. Rising curtailment can be acceptable during rapid construction, but persistent increases show that networks and flexibility are falling behind generation.
The second is the spread between the cheapest and most expensive hours. A wide spread creates an opportunity for storage and responsive demand. If the spread stays wide after large flexibility investments, the system may have a network bottleneck rather than a simple shortage of batteries.
The third is the frequency of negative prices. A few negative hours can encourage useful behaviour. A large and rising share suggests that market rules, connection queues or demand incentives need attention.
The fourth is carbon intensity, not only total emissions. Total emissions reflect both the cleanliness of supply and the size of demand. Intensity reveals whether each unit of electricity is becoming cleaner, even in a growing economy.
None of these indicators can stand alone. Together they show whether a power system is merely installing equipment or actually learning to use it.
The next bottleneck has already arrived
Renewables overtaking coal is a historic change. It deserves the headline. But the operating problem waiting underneath the headline is already visible in negative prices, curtailment and violent swings between midday abundance and evening scarcity.
The next phase will reward grids that can move electricity across distance, shift it across hours and persuade demand to meet supply halfway. It will punish systems that count annual megawatt hours and ignore the clock.
The power transition is no longer only a contest to build the cheapest generator. It is a contest to make every hour work.