The story began with a quotation. Reading from Elon Musk's remarks that week, the Moonshots panel relayed his argument that switching to sustainable energy is necessary but not sufficient for humanity's survival: "extremely severe extinction events happen every 100 million years or so, and just switching to sustainable energy will not be enough to stop them." Musk's proposed answer, as described on the show, is a fleet of solar-powered AI satellites parked between the Earth and the sun, making continuous small adjustments to the sunlight reaching the planet. His timeline, also quoted: "we have about 50 years or so to take action, which should be more than enough time for the space satellites to solve any heating problem."
Peter Diamandis said he had been pitching almost the same thing for the better part of a decade at XPRIZE's Visioneering sessions, where donors and advisers debate which prizes to design and launch. He calls the idea solar shades — "a thermostat for the Earth" — spacecraft positioned between Earth and the sun that titrate the solar flux hitting the planet, dialling the temperature up or down.
Blocking light, or aiming it
Alex, who called himself "a huge fan of geoengineering," wanted to go further than shading. He said he had been looking for a startup to fund that would work on engineering global weather, and what he wants is not only a sunshade but a way to aim sunlight: low-Earth-orbit satellites, or mirrors on the ground, that could "optimize hurricanes out of existence." If a hurricane were bearing down on a coast, he asked, wouldn't it be wonderful if AI could direct some energy to steer it away from populated shorelines?
His recipe has two halves. One is global AI weather models — software that predicts atmospheric behaviour across the whole planet. The other is what he called "points of actuation": physical means of changing something. He named satellites "in the style of reflect orbital" as one option, able to divert sunlight down onto weather patterns, and said he did not much care whether the method was focusing light or, as Diamandis prefers, blocking it. With thousands or millions of orbiting mirrors, he argued, "that's a recipe for global weather engineering. And I think we're going to get it."
The company he referenced describes something considerably narrower. Reflect Orbital proposes satellites that redirect sunlight onto requested locations after dark, with the orientation of each satellite steering the beam and additional satellites adding brightness. Its published roadmap separates early illumination demonstrations from later energy services: a 2026 target of roughly five minutes of light at about the brightness of a full moon, with solar-generation services still in testing through 2028. The applications it lists are extending generation at existing solar farms and lighting remote operations, and it says deployment would depend on local approval, sharing satellite positions in advance and avoiding sensitive locations. Weather modification is not among the uses it advertises.
"Who gets to control that?"
The objection came from Dave, and it was about people rather than physics. Warming is not uniformly unwelcome, he pointed out: many nations do not want it, but Russia might, because it opens up waterways. A thermostat for the planet has one setting, and it reaches everyone. "Who gets to control that," he asked, "because you're impacting, you're not just one country or a dozen countries. You're impacting, you know, a couple hundred countries."
Alex's answer was to make the disagreement tradable. Treaties already exist for things that cannot otherwise be exchanged, he said; weather could work the same way, with municipality A trading with municipality B for rain. "You can have a global weather market."
Dave thought that was naive, and reached for a smaller problem as evidence. Urban planning, he said, is simple compared with geoengineering — and it goes badly, producing traffic jams upon traffic jams. Most cities, he added, were never designed at all.
Alex took the historical record the other way. For the past few hundred years humanity has been massively changing carbon levels, temperatures and ocean levels, he said, just not intentionally. Now the technology to design the world on purpose is arriving, so "let's do it." He framed the whole turn as overdue, arguing that Western civilisation became allergic to large-scale physical engineering some time around the late 1960s and lost roughly half a century it could have spent developing fission reactors, continuing moon landings and trying early geoengineering techniques.
Diamandis agreed that the hardware is not the obstacle. Blocking or reflecting sunlight, he said, is "pretty damn straightforward"; the decision about who controls it and what the right temperature is, is "the process that's just frighteningly broken." Historically, he added, nothing gets a unified response until it becomes drastic.
His proposed prize is deliberately modest against that background: a demonstrator showing that such a system can be built fail-safe. "You don't want to cause an ice age by blocking too much sunlight," he said, "but you want to be able to titrate it at just the right amount." On governance he offered a bet rather than a mechanism — that people now under 25, growing up AI-native and talking to each other across languages, will not tolerate today's slow, divided world governance, and will find a new way of deciding these things. He pointed to the panel's earlier story about a probe headed for Alpha Centauri as the kind of project that unifies people across the world.
The rain that falls somewhere else
The test arrived later, in the listener questions. Jim Plamundon 637 asked whether inducing rainfall in one region could reduce rainfall in areas downwind. Alex took it and conceded the point immediately: to a first approximation water is conserved on this planet, so water that falls in one place is water that cannot fall further downwind.
Then he answered what he called the question behind the question — isn't that a problem? — with no. "I think we will trade atmospheric effects. We will trade weather. It'll be a vibrant market." Some municipalities want rain and some do not, he said, and the case that interests him is not the field next door but the exchange across a planet: a desert such as the Sahara trading for precipitation from somewhere in North America to re-green itself. That, he said, is possible "with a suitable weather control system and good enough planetary-scale AI models" — the two things his market would need, and the two things nobody on the panel said anyone has built.