RPT-BREAKINGVIEWS-‘Solar in space’ crowns white elephant energy boom

By Robert Cyran

NEW YORK, April 16 (Reuters Breakingviews) - A white elephant is a beautiful gift that proves financially ruinous for whoever owns it. For utilities, the idea of sending satellites into space to generate solar power might qualify. Yet a new generation of startups and investors is trying to revive it, alongside other quixotic green energy pushes, reprising decades-old ideas to solve the power demands of artificial intelligence. Then as now, the numbers don’t work out.

Earth-bound solar power is fast to install and, at as low as $29 per megawatt hour before subsidies, cheaper than fossil fuels, according to Lazard. Over half of new U.S. generation capacity planned by utilities for 2025 will harness the sun, the Energy Information Administration reckons.

Still, with new data centers serving AI accelerating demand, sources that provide uninterrupted power are increasingly appealing. Suspend disbelief for a moment and panels in space, staring directly at the sun and unperturbed by clouds, fit the bill.

Startup Aetherflux said on April 2 that it had raised $50 million in new funding. Counting big-name backers like Andreessen Horowitz, it plans to launch a demonstration satellite next year. It's part of an $89 billion rush of venture capital investments in clean energy between 2020 and 2024, according to Pitchbook, up from a trickle beforehand.

Others are keen too. Rocket firm China Academy of Launch Vehicle Technology unveiled a similar plan in October, likened to a Three Gorges Dam Project in space. In other far-fetched fields, nearly 100 companies are chasing the dream of miniaturized, mass-produced nuclear reactors.

Science-fiction writer Isaac Asimov proposed something like Aetherflux’s pitch 84 years ago. More concretely, the 1970s oil crisis drove research into these very same ideas. A 1979 study found that enough satellites to generate 300 gigawatts of solar power could be built – for about $4 trillion in today’s dollars, or 14% of U.S. GDP.

Fundamentally, these schemes add cost to well-understood power sources. A 2024 NASA report is unflinching: putting a single 2-gigawatt solar station in geosynchronous orbit would take 1.5 million panels and weigh about 6 million kilograms, nearly half the weight of every human-made object currently whizzing around the planet. Space-based panels would degrade twice as quickly as on Earth while being extremely difficult to repair. Beaming energy via laser or microwave is inefficient.

The rush of VC capital into the sector may have already peaked, and projects will fail. More generally, they fall victim to the appeal of complex, theoretically elegant solutions that often grow in price as existing solutions get cheaper. The thing about reaching back into the past is that it can never catch up with the present.

Follow @rob_cyran on X

CONTEXT NEWS

Aetherflux, a startup for building satellite-based solar power in low earth orbit and transmitting it back to earth using infrared lasers, said on April 2 it had raised $50 million. The firm was founded by Baiju Bhatt, a billionaire co-founder of Robinhood Markets. Investors in the company include Index Ventures, Interlagos, Breakthrough Energy Ventures, Andreessen Horowitz, NEA, Vlad Tenev, Dan Gallagher, Laurent and François-Paul Journe and Jared Leto.

In October, Long Lehao, chief designer for major rocket firm China Academy of Launch Vehicle Technology, introduced a plan for solar stations in orbit, comparing it to a “Three Gorges Dam project above the Earth” according to the South China Morning Post.