Harnessing the Sun

Introduction

For most of solar's history, the argument against it was cost. Twenty-five years ago, a watt of solar cost about $5 – its power more than ten times the price of coal's. A quarter of a century is a long time for a manufactured technology growing at double digits. Today that watt costs around 10 cents and the IEA now calls solar the cheapest electricity in history.

But what use is the world's cheapest electricity if it vanishes at sunset? Having beaten every expectation on cost, solar now faces a harder question: reliability. Demand does not set with the sun. The engineer's answer is storage: batteries to catch the day's surplus and serve it back at night. Silos for surplus sunshine. And since batteries have fallen in cost just as solar did, what was once an engineer's answer is becoming an economic one too. So how far can batteries go, at what cost, and where?

We built a model to find out. We divided the world into 5,000 regions, each about the size of Belgium, and in each calculated the solar and batteries needed to serve a constant 1 MW load in every hour of the year, and what that costs.

This is a sceptic's test. The model includes no demand shifting, no grid imports, no wind, no hydro and no backup. Just solar, batteries and a flat load. In reality, power systems have far more options: demand varies by hour, grids trade across regions, customers shift some use and the grid already has other generating technologies. If solar and batteries can hold a flat load alone, a real system can only do better.

The results show that solar plus batteries can already deliver high-uptime power across much of the world, at competitive cost, especially in the places where demand is growing fastest. What follows are nine findings.

1. Sunlight is abundant everywhere

The total potential of solar is enormous. The sun supplies the Earth with thousands of times more energy than humanity uses. Even after applying limits on usable land, build density and conversion efficiency, solar potential remains far larger than local electricity demand.

The map to the right measures each region's constrained annual solar potential as a multiple of its local electricity demand. More than nine in ten people live in regions where that potential is at least ten times local electricity demand; for over half of humanity it is more than a hundred times. In absolute terms, the scale is striking: more than 95% of the global population lives in regions where local solar potential per person is over twice current U.S. electricity demand per person, about 12 MWh/year. For roughly two-thirds of humanity, local solar potential per person is larger than Saudi Arabia’s oil production per person.

Globally, technically suitable land alone could generate around 185 times today's electricity consumption, and still about 125 times after stricter land-use exclusions such as excluding farmland and nature reserves (the "Policy" toggle).

This abundance is central to the rest of the analysis. In effectively every region on Earth, the question is not whether enough solar energy can be harvested over the course of a year; the annual potential is vast. The real constraint is timing: making that energy available in the hours when it is needed. Solve that timing problem, and the usable solar resource becomes unstoppable.

2. Batteries make all the difference

Once solar is cheap, overbuilding becomes the natural first lever to raise uptime. That works for a long time. More capacity raises annual generation, improves output in weaker hours and seasons, and increases the share of demand solar can serve.

But overbuilding solar eventually runs into limits. Additional panels keep adding energy, but less of it arrives in the hours that set uptime. In our simulations, even a twenty-fold overbuild never carries a solar-only system past about 47% of a flat load, anywhere on Earth. Annual generation becomes abundant while reliability stays capped by timing.

Batteries are the game changer. They turn curtailed solar output into usable supply and move solar from a resource that can cover a large share of energy demand into one that can cover a large share of hours. Watch the map as the battery grows from 0 to 24 MWh: uptime fills in, evenings first, then nights, then the weak-sun days within a week. Regions home to more than 90% of the world's population can reach 80% uptime, large parts of the world move into the 80–90% range, and favourable regions push past 95%, without any other backup.

This is why the relevant unit of analysis is not solar alone, but a solar-battery system: cheap generation paired with enough storage to shift it into the hours when it is needed.

3. Batteries work because most solar variability is daily

Batteries have such a large effect because, across most of the world, the main solar balancing challenge is daily rather than seasonal. That makes short-duration storage unusually powerful: a few hours of storage can absorb a large share of otherwise curtailed daytime output and shift it into the evening and night-time hours that determine uptime. Technically, most of the world can reach 80% uptime with enough solar and daily storage.

Seasonality still matters, especially at higher latitudes. It explains much of the difficulty of reaching very high uptime in places with weak winter output, such as the UK. But across most populated regions, the largest and most predictable mismatch is the day-night cycle: solar produces during the day, while a flat load continues after dark. Between the tropics, where variability is overwhelmingly diurnal, solar plus batteries can reach 95–99% uptime without major difficulty.

The hourly profiles make this clear. In the weekly animation, yellow shows solar used as it is produced, while purple shows the same sunlight shifted through storage and supplied after dark. Changing seasons reveals the seasonal swing, but the dominant pattern remains the daily surplus and deficit between daylight and night. Batteries are well matched to that pattern, which is why they lift uptime so sharply across much of the world.

4. Solar plus batteries are cheap exactly where most people live

So far, the argument has been mostly about physics. Now we can turn to the economics.

When one runs the cost numbers using today’s solar and battery costs from the IEA and BNEF, one can see that solar-plus-storage is cheapest exactly in the places where most people live. Around 80% of the global population can get 80% uptime solar-plus-storage power for less than $100/MWh, and half can get it for below about $80/MWh. That is already competitive with average fossil generation costs today, typically around $100/MWh.

This is not especially surprising. Most people tend to live in places with moderate climates and decent sunlight, not in the darkest and most seasonal parts of the world. Those same conditions also make solar easier to use: strong output, less extreme seasonality, and less need for long-duration storage.

The main caveat is finance. Many sunny regions face higher borrowing costs, which raises the delivered cost of electricity. But even after accounting for local costs of capital, the core result remains: for much of the world’s population, solar-plus-storage is already a low-cost source of reliable power.

5. Solar plus batteries are cheap where electricity is missing

Around 760 million people still lack access to electricity. Their current options are no electricity, kerosene, diesel generators, battery charging shops, or waiting years for grid extension.

Many of these communities, shown in red on the map, are in regions with strong sunlight, mild seasonality, and low solar-plus-storage costs. In other words, many people still without electricity live in exactly the places where solar and batteries can already provide some of the cheapest power on Earth. 89% of people without access live where that 80%-uptime cost is under $80/MWh; and effectively 100% of people live in regions where costs are under $100/MWh.

The advantage is not only cost, but proximity. Local solar and batteries can generate, store, and deliver electricity close to demand, reducing the need to extend centralised infrastructure across long distances before service can begin.

The sun reaches every village in the world every day. The technologies to harvest and store its energy have now become cheap enough to matter, and can provide affordable electricity for all.

6. Solar plus batteries can already beat many grids on uptime

The problem is much larger than the 760 million people without electricity access. Another 2 billion people have a grid connection, but not reliable power. Their electricity is not available around the clock, and for about half of them, uptime is below 60%. That means daily outages, load shedding, voltage swings, damaged appliances, lost business hours, and expensive backup power.

For these people, the benchmark for solar plus battery reliability is not 100%, but the availability of the grid they have access to today. In many places, 80–90% uptime from solar plus batteries would be a major improvement, especially if it comes with lower delivered costs. The left chart counts the people at each level of grid reliability; the right chart shows what it costs to match that reliability with solar and storage instead. It shows how nine out of ten people could be served more reliably by solar and storage at a price at or below $70/MWh.

That makes unreliable-grid markets the natural starting point for scale. In these places, solar plus storage can deliver more reliable power at competitive cost, creating the demand base needed to drive further deployment, learning, and cost reduction. And this is already happening: solar and battery uptake is accelerating across emerging markets, including India, Pakistan, and Kenya. The transition is not waiting for perfect economics against perfect grids. It is starting where the existing system fails.

7. Getting from 90–95% to 100% needn't be expensive

Solar plus batteries can already get close to firm power in many regions. But “firm” needs to be put in context: no power technology runs at 100% uptime. Nuclear plants are among the most consistently used, but globally they typically operate at around 80–85% capacity utilisation. Coal fleets are often closer to 50–60%, while gas plants, especially when used for flexibility and peaking, can be closer to 30–40%.

Even when conventional plants are pushed hard, they are not available all the time. Typical technical availability is roughly 80–90% for nuclear and coal, and 85–95% for modern gas plants, depending on age, maintenance, fuel supply and operating regime. The relevant benchmark for solar-plus-storage is therefore not perfection, but whether it can approach the reliability that conventional plants actually deliver in practice.

On that benchmark, getting solar plus batteries to 90–95% is not the hard part. The expensive part is forcing them to cover every last hour on their own. The final few percent of hours are rare, but difficult to serve. Meeting them with only solar and batteries means building extra panels and storage that sit unused for most of the year.

That is why costs rise sharply at the very end. Moving from 95% to 99% uptime might add around $15/MWh, or roughly 20%, and in harder regions $30/MWh or more. Moving from 99% to 100% can cost more than three or four times as much again.

A real power system does not need to solve those final hours with solar and batteries alone. Once they provide 90–95% of annual supply, the remaining hours can be covered by a low-cost firming resource: a gas turbine, diesel generator, hydro, interconnection, demand response, or another dispatchable technology.

The simplest example is a gas turbine or diesel generator used only as backup. Keeping it available might cost around $5–10 per hour, whether or not it runs. When it does run, the variable cost might be around $80–200 per hour, mostly for fuel and maintenance. But if it runs only 10% of the year, that variable cost averages to just $8–20 per total hour.

That is why a modest amount of backup can turn 90–95% solar-plus-storage supply into a fully reliable system without dramatically increasing cost. Solar and batteries provide almost all of the energy; the backup resource only covers the small tail of difficult hours when they fall short.

8. Planned fossil capacity is at risk

Electricity demand is rising around the world. Many countries are still planning new baseload coal and gas capacity to supply that demand. The lion’s share of this planned capacity is located in regions where solar plus batteries are already cheap.

Of the roughly 850 GW of planned coal and gas capacity in the GEM dataset, around 590 GW, or 70%, is planned in regions where solar plus batteries can already deliver 80% uptime power for less than $100/MWh. The chart stacks that planned capacity against the cost of the solar alternative in the same place. These are often regions with fast demand growth, strong solar resources, growing populations and weak or incomplete grids.

This does not imply a simple one-for-one replacement of every planned plant. Power systems need capacity, flexibility, reserves and grid infrastructure. But the default assumption that rising electricity demand requires new fossil capacity is increasingly outdated. Instead of a baseload coal plant, a solar plus battery site at 75% uptime will already be materially cheaper. Or instead of an expensive baseload gas plant, a cheaper back-up plant plus solar and battery for the bulk of power is likely to be a better fit.

Planners now need to compare new coal and gas plants against solar-plus-storage systems that can provide high-uptime electricity at competitive cost. In many growth markets, that comparison will show that solar plus battery beats almost all pure-play fossil projects on cost today.

9. Falling costs turn affordable into irresistible

All of the analysis so far uses today's costs. But solar and batteries are manufactured technologies, and their costs are still falling. Deployment increases manufacturing scale. Scale reduces costs. Lower costs open new markets. Those markets support further deployment.

Today's affordability map is therefore not fixed. As time moves on and electrotech costs continue to fall on learning curves, regions that are already cheap become cheaper, and regions just outside competitiveness cross the line. The global area where solar plus batteries can provide high-uptime electricity at low cost expands over time.

By 2030, our analysis suggests that more than three quarters of the world's population can get 80% uptime solar-plus-storage power for less than $80/MWh, and 90% of the world for less than $100/MWh.

This matters for investment decisions being made now. Power plants are long-lived assets: a coal or gas plant planned today will be expected to operate for decades. But over that same period, solar-plus-storage is likely to continue moving down the cost curve. That means the comparison is not between a fossil plant and today’s solar-plus-storage costs; it is between a long-lived fossil asset and an alternative that keeps getting cheaper throughout its lifetime.

What is competitive today will become irresistible in the 2030s.

Conclusion

For most of solar’s history, the question was cost. Could it ever be cheap enough to matter? That question has largely been answered.

The harder question is reliability. Can solar, with batteries, keep the lights on for enough hours of the year, in enough places, at a price that matters?

This analysis suggests it can. Sunlight is abundant across most of the world. Overbuilding turns cheap solar into surplus energy; batteries move that surplus into the hours when it is needed. Together, they can already deliver high-uptime power across much of the world, especially where demand is growing fastest and grids are weakest.

The last few percent of reliability do not overturn the case. They turn it into a system design problem: a small amount of backup, grid connection, demand flexibility, or complementary generation can cover the remaining hours.

And the economics are still improving. Solar and batteries are manufactured technologies, still moving down learning curves. A fossil plant planned today will run for decades; over that same period, the solar-plus-storage alternative is likely to keep getting cheaper.

The question is no longer whether the sun can power the world. It is how quickly we build the systems that let it.