The Sunlit Half

The observation

At every moment, one half of the Earth faces the sun. That half moves, but it never goes away.

Solar power has one famous weakness: it stops at sunset. Most of the effort in clean energy today goes into working around that, mainly by storing midday power in batteries until the evening.

But sunset is a local event. When it is dark in Melbourne, it is mid-morning in London and the middle of the night in New York. Somewhere, always, the sun is high and panels are producing more than the people near them can use.

So consider a different answer to the night. Do not store the sunlight. Move it.

The idea

What if the world’s grids were joined into one, the way the world’s computer networks were joined into the internet?

The drawing shows the simplest version I can think of. A ring main of high-voltage direct current cable circles the northern continents: across Europe and the Asian steppe, down to the Pacific coast, over the Bering Strait, through Canada, and back across the North Atlantic by way of Greenland and Iceland. Spurs run south from it to the places with the most sun: North Africa, the Gulf, India, Southeast Asia and northern Australia, and from North America down to Brazil.

Each country connects to the ring through a converter station, the way a house connects to the street. During its day, a country sells its surplus into the ring. During its night, it buys from whoever is in daylight. The terminator, the line between day and night, sweeps around the planet at 15° of longitude an hour, and the flow on the ring follows it.

The ring is the important shape. A line can be cut once and fail. A ring that is cut once becomes a line and keeps working, because every station can still be fed from the other direction.

Why it is possible now

Twenty years ago this would have been a fantasy. Three things have since moved it into the range of a very large engineering project.

Long-distance direct current. Alternating current loses too much over thousands of kilometres. High-voltage direct current does not. China’s Changji–Guquan line, energised in 2019, carries 12 GW over about 3,300 km at ±1,100 kV, with line losses of roughly 1.5% per 1,000 km. That is one segment of the ring, already built.

Long sea cables. Viking Link between Britain and Denmark runs 765 km under the North Sea and carries 1.4 GW. Proposals for 3,800 km (Morocco to Britain) and 4,300 km (Darwin to Singapore) have been engineered in detail. The ring’s longest sea leg would be about 1,000 km.

Cheap panels. Solar modules now cost a small fraction of what they did in 2010. Generation is no longer the expensive part. I think the bottleneck is distance, not generation.

A sketch of the numbers

These are rough. They are here to show the scale, not to settle it.

Quantity	Value
World electricity use	≈ 30,000 TWh a year
Average draw	≈ 3.4 TW
Solar capacity to match it, at 25% capacity factor	≈ 15 TW
Land for that, at ~40 MW per km²	≈ 375,000 km², about the size of Japan
Share of the Sahara	≈ 4%
Ring length, great-circle	≈ 27,000 km
Sea cable in the ring	≈ 2,300 km, in four crossings
Loss, evening peak, sun 90° to the west	≈ 10%
Loss, local midnight, sun 180° away	≈ 20%
Lithium battery round-trip loss, for comparison	≈ 10–15%

The last three rows are the honest heart of it. For the evening peak, when demand is highest, the sun is still up about a quarter of the way around the world, and the ring loses about as much as a battery would. For the dead of night, the nearest strong sun is on the far side of the planet and the ring loses more.

There is a useful pattern here. Local time runs later as you go east. So at 7 pm, the places still in afternoon sun are to your west, and at 7 am, the places already in late-morning sun are to your east. A city’s evening is fed from one direction around the ring, and its morning from the other.

What the ring is really good at

I started with day and night, but I now think the stronger argument is the seasons.

A battery can carry noon into evening. It cannot carry July into January. When Europe is in a dark, still December week, Australia and southern Africa are in midsummer. The spurs that run south of the equator let one hemisphere’s summer cover the other’s winter. No storage technology we have does that at a reasonable cost.

The second thing it is good at is weather. A cloudy week is a regional event. On a grid the size of the planet, someone is always under a clear sky.

So the ring does not replace storage. Storage handles the hours; the ring handles the seasons and the weather, and shares the load on the hours.

The hard parts

Distance. Losses compound. At 1.5% per 1,000 km plus about 0.7% at each converter, sending power halfway around the world costs about a fifth of it. That pushes the design toward many medium-distance trades rather than a few heroic ones.

Trust. Energy is security. Europe learned in 2022 what it means to depend on a neighbour’s gas. A ring is only as dependable as the countries along it. I think the ring shape helps, because no single country can cut a station off, but it does not solve the politics.

Sea cables are fragile. Cables in the Baltic have been damaged repeatedly since 2023, some by dragged anchors. A world grid would be a target. It needs redundant crossings, monitoring, and the ability to run as separate regional grids when a link goes down.

DC networks are young. Most direct current links today join two points. A ring with dozens of stations needs DC circuit breakers that can interrupt a fault in milliseconds. These exist (China’s four-terminal Zhangbei grid has used them since 2020) but not yet at this scale.

The Arctic. The Bering and North Atlantic crossings run through sea ice, permafrost and very few people. They would be the hardest and most expensive kilometres on the ring.

Who pays. A single overland segment like Changji–Guquan cost roughly US$6 billion. A first ring of one 12 GW circuit is likely in the order of US$100 billion. Carrying a tenth of the world’s load would take many circuits in parallel, and several trillion dollars. That is large, but it is the same order as a year or two of what the world already spends on the energy transition.

Where I would start

Not with the ring. The ring is the end state, and it only makes sense once the arcs that make it up already pay for themselves.

• North Africa to Europe. Strong winter sun, short sea crossings at Gibraltar, and large demand to the north.
• Australia to Southeast Asia. The Darwin–Singapore link has been designed in detail.
• Across Asia. China’s internal UHVDC network is already most of an arc from the Gobi to the Pacific coast.
• The North Atlantic. Iceland to Britain has been studied for its geothermal power; Greenland is the next step.

Join the arcs, and the ring appears. The Bering crossing, short but remote, would be the last link to close.

Others who have drawn this

This is not a new idea, and I am glad of it.

• Buckminster Fuller argued for a world electric grid in the 1960s and 70s, as the single most useful thing humanity could build.
• The Global Energy Network Institute, founded by Peter Meisen in 1986, has promoted Fuller’s grid since.
• Desertec, launched in 2009, planned to power Europe from Saharan sun. It mostly faded by the mid-2010s, for reasons of politics and finance more than engineering.
• GEIDCO, the Global Energy Interconnection organisation started in China in 2016, has published detailed global grid plans, including a Bering link.
• One Sun One World One Grid, proposed by India in 2018 and launched with the UK at COP26 as the Green Grids Initiative, uses almost the same phrase as this folio.

What I add is mostly the picture: the ring, and the night half sliding around it. I think the picture makes the case faster than the reports do.

The sun has never set on the world. Only on places.