Slowly, I become aware of a regular beeping. The muffled sound of conversation drifts over me. I feel a deep soreness, and gnawing tug of a cannula in my arm. Slowly, slowly, I fall back into consciousness, until all at once my eyes open. Of course I’m blinded by the light, but as my eyes adjust, I take in the sight of a hospital room.

The conversation takes on a suddenly-alert tone. I’ve been noticed. Half an hour later, my family are gathered around me.

“Take your time”, they say. “You don’t need to say anything just now, you’ve got a long road to recovery ahead of you.” With a wrenching effort I make the slightest of nods. Then…

“Wh… Wh… What year?”

They share concerned glances, until someone bites the bullet.

“2050”

I wince and start to process the loss of two and a half decades. The faces around me grow sombre, some avert their gaze. It takes a few minutes before I find the energy to speak again.

“Wa… Wa… Was the UK able to fully decarbonise its grid?”

The faces look up, they smile.

“Yes! We did it!”

“H… How?”

A pause.

“Do you want to guess?” someone says, before cutting themselves off. They look guilty as they realise what they’ve said.

“L… Laptop?” I respond, and the guilty face disappears.

After ten or so minutes someone is able to produce a laptop, it’s propped onto my frail torso, and my hands are gently lifted onto the keyboard. With a herculean effort, I type out this blog post.

We probably stuck with gas. The UK transitioned to gas because it became cheap, and required the least amount of building stuff near people; and it continued to use gas for the same reasons.

Hinkley Point C and Sizewell C were eventually completed, but they are the only nuclear power stations now running. The rest went offline in the late 2020s and early 2030s, in the mid 2030s there were a few years where the UK had no nuclear power at all! As renewables grew, no politician could again bring themselves to incur the costs—financial or political—of commissioning new reactors. SMRs never made it over the line either. The nuclear regulatory environment already made it near impossible to get proven reactor designs built, so developing completely novel designs never worked out. Rolls-Royce eventually decalared defeat, and cancelling Wylfa over costs.

The economics of solar, wind, and batteries on the other hand, were far more favourable. And though progress was halting as more and less permissive governments rotated in and out of power, renewables’ fraction of the grid did grow. By a lot.

The grid expanded too slowly to accommodate this growth, and curtailment grew. Lithium-ion battery storage was able to mitigate this to an extent, but it couldn’t economically provide storage for longer than 24 hours. Compressed-air energy storage plants, and their more versatile cousin, liquid-air energy storage plants came online here and there, but at nothing like the rate needed to solve the problem. Even if the grid added enough transmission lines to solve curtailment, the problem of Dunkelflaute remained. The UK would always have periods of weeks on end when there was neither the sun nor the wind to meet demand, and these periods would drain the country’s energy storage to zero. At this point, the gas turbines would have to spin up, and—to recoup the costs of standing idle most of the year—they charged exorbitant fees to do so.

This problem had become very apparent by the early 2030s, and the government of the time believed the only solution was long term energy storage. To that end, they began work on the UK’s first salt dome hydrogen storage facility, which would use excess grid capacity to produce hydrogen via electrolysis and store it in a solution-mined cavern underground. Then when renewables and short term storage couldn’t meet demand, that hydrogen would be burnt in on-site diesel engines, generating power for the grid. The facility would provide less power than a single conventional gas power plant, but it would be entirely renewable-powered, and was to pave the way for a fossil-fuel-free grid.

But as the consultation process on the hydrogen storage plant began, something happened in the United States that would ensure the facility was obsolete before it opened. For the first time, experimental synthetic natural gas (SNG) plants in the sunniest parts of the country began selling gas that was comparable in cost to natural gas. These plants were off-grid facilities in the desert, that used a giant solar farm and clever machinery to capture CO2 from the air, split hyrdogen from water molecules, and run Sabatier reactors that combined the two into methane gas.

When the sun shone, the plant ran, and produced a stream of methane gas that could be piped away and sold as a drop-in (or mix-in) replacement for natural gas. When it was dark, the plant would go dormant. The plants were designed to ruthlessly minimise capital costs, and be as cheap to run as possible. As a result they were incredibly inefficient. The energy you got once you finally burnt the gas was a fraction what had been harvested from the solar farm. But as long as the cost of solar PV panels continued to fall, it didn’t matter. SNG had become competitive with natural gas, and it was about to compete it out of existence.

The revolution began in America, but spread to China, whose manufacturing prowess drove down the cost of SNG plants even further. China began exporting SNG equipment to the rest of the world at a pace that soon outstripped America. Now any country with a desert was sitting on a potential gas field, and by the late 2030s mass-deployment to the sunniest latitudes had pushed down gas prices across the world. As the technology got cheaper, it became viable in places further and further from the equator, this not only lowered gas prices, it decreased their volatility. No longer was the cost of energy given to fluctuating wildly over geopolitical events. Supply was dispersed widely enough that regional incidents no longer caused massive shortages. And many of the countries that had previously been most dependent on gas imports could now produce their own.

This marked the end of the UK’s progress towards energy independence. While most of its power now came from renewables, the 20 or so per cent of energy that was still provided by gas each year was generated with imported SNG. That gas—mostly transported via pipes from the Sahara and Southern Europe—is now too cheap, too clean, and too reliable to be displaced by any other technology any time soon.

Fusion won’t become a reality until technology advances to the point that it can be developed by hobbyists—or until it’s needed for space travel.

It turned out the UK still needed long term energy storage, but the storage was all it needed. A lot of salt domes were solution mined, but they were filled with imported SNG, not hydrogen generated on site. While we could have developed our own SNG industry, this would have required an even bigger build out of solar and wind power, and with energy prices at historical lows there was no longer any political will to do so.

Obviously this prediction will be wrong in some way, maybe most or all ways. But it’s an attempt to sketch out what the future might look like, to do so quickly without sitting down to a do really in depth forecasting, and within the bounds of one very important contingency: that AI progress more or less stops where it is right now. This story takes place in a world where, for whatever reason, No One Built It and Everyone Didn’t Die. It’s also a world where controlled AI hasn’t advanced to the level where it can really accelerate the roll out of new physical technologies, and things still happen on roughly the timescales that we’re used to.