By James Burkhamer

Chronicle Contributor

Future dystopian scenarios can always be worse than imagined.

At least there’s not a megalomaniac trillionaire proposing to launch a million artificial-intelligence (AI) satellites into earth’s orbit. Oh wait. Elon Musk indeed has designs for a mega-constellation of 1 million satellites called “StarMind," Space.com reports.

The idea of StarMind is not internet communications, or even the reflection sunlight back to earth – StarMind would relocate a large chunk (if not all) of AI computing power from earthbound data centers into the skies above us. As Musk quips: “it’s always sunny in space!”

This isn’t a pie-in-the-sky scenario; it’s all too real. First, SpaceX filed an application with the FCC in July 2026 to launch 100,000 Starlink satellites into orbit; each one weighing about two tons.

These are third-generation broadband satellites, which will greatly expand space-based internet capacity for entire world. SpaceX currently has only 10,000 satellites in orbit. However, this is a mere preview of what’s to come.

Starlink already has artificial-intelligence capabilities, powering everything from autonomous fleets to AI training models. Calling Elon Musk’s vision of future Starlink capabilities “ambitious” is an understatement.

He is quoted in August as saying, “All cars will have Starlink in the future," envisioning all vehicles in the world connected under his system, according to basenor.com. This would sound impossible if not for the approval from the U.S. government.

However, Starlink is only the foothills of the AI future – the Skynet mountain is still looming in front of us.

Musk’s vision of space-based artificial intelligence moved closer to reality this month, when SpaceX signed an agreement with AI giant Nvidia, gate.com reports.

Nvidia is the gold standard in artificial-intelligence hardware, with a commanding 80% market share in chips and graphics processing units (GPUs). SpaceX would prefer to use Nvidia’s underlying hardware structure, rather than spend time and money reinventing the wheel.

Nvidia’s already ahead of the curve, having launched a space computing platform in March 2026. The StarMind application is currently pending with the FCC.

The first concern with StarMind is the prospect of Artificial General Intelligence (AGI). Current AI as we know it (ChatGPT and such) is called “narrow” AI, which is task specific. Experts have been sounding alarm bells about AGI for years, but fears seem exaggerated because AI technology can only advance so far.

The bottleneck is earth-bound data centers. Many communities don’t want them, plus there’s only so much electricity and water-cooling capability.

Space-based AI solves the bottleneck problem, as there is unlimited energy and plenty of cold surface area in outer space. AGI is much more advanced than narrow AI, with computers able to mimic or even surpass the human mind.

AGI computing power could theoretically use logic and reasoning beyond human capacity, teaching itself faster than any person ever could.

There are numerous nightmare scenarios concerning AGI controlled from the skies. The

surveillance capability alone would make Flock cameras look like tinker toys.

The movies Patriot Games (1992) and Enemy of the State (1998) show satellite surveillance capabilities long before the internet of today – let alone enhanced by AI.

Other scenarios include advanced weaponry, humanoid robots, Neuralink, bioweapons, digital currency, transportation, healthcare, education, and even knowledge itself – all in the hands of invisible artificial intelligence over our heads.

If AGI isn’t concerning enough, the next step in AI evolution is downright frightening. Artificial superintelligence (ASI) is a theoretical state where machine intelligence isn’t just smarter than any one person – it would be smarter than all of human intelligence COMBINED.

This would be in outer space, with no “off switch”. ASI brings up the prospects of abstract thinking, self-awareness, consciousness, even sentience – basically the premises of The Terminator and Matrix movies from the 80s and 90s.

The other chief concern about MindStar is what happens when the one million satellites start colliding into each other?

The Outer Space Institute utilizes the Collision Realization and Significant Harm (CRASH) Clock, which measures congestion and collision risk in low Earth orbit. Currently, the CRASH Clock estimates a 2.5 day countdown to a catastrophic satellite collision, if all avoidance maneuvers were to suddenly stop – and that’s with only 20,000 satellites currently in low-earth orbit. The CRASH clock was 164 days in 2018. A sudden satellite crash of that magnitude would wreak havoc on earth, plus rendering our orbit useless for a long period of time.

A sudden stoppage in satellite maneuverability would happen if the sun were to get angry and flare up a giant solar storm. That would create a traffic jam over our heads that not even ChatGPT could figure out.

In addition, most things on earth wouldn’t work anymore, since most everything would depend on orbital artificial intelligence. There is no international governing body that concerns itself with satellite collision prevention.

Of course, why should SpaceX and the United States have a monopoly? StarMind wouldn’t be the only player in the game. Google, Blue Origin, Amazon, and others are planning to launch several thousand more AI satellites into orbit, Technology Magazine reports.

The orbital AI space race is also being run by China, Russia, India and the European Union. Even third-world nation Rwanda is pioneering a satellite data program in Africa.

Proponents of orbital artificial intelligence point out that this might solve the problem of terrestrial data centers. Local communities are hesitant to have them, plus the resources and electricity usage is enormous.

AI capabilities are constrained to the task-centric narrow bandwidth on earth – but StarMind and other satellite constellations would solve all these problems. So what’s the concern?

Plan 9 from Outer Space was a 1950s science fiction movie about aliens seeking to stop humanity from creating a doomsday weapon, that could set off a chain reaction and destroy the universe.

In an ironic twist, the doomsday scenario won’t be coming from aliens – it might be coming from people. Our insatiable thirst for more data and artificial intelligence might lead to a technology utopia, only dreamed about by sci-fi authors.

Unfortunately, the majority of science fiction is DYStopian – as something inevitably goes awry in these technological grand schemes (as proven by the resurrected dinosaurs of Jurassic Park).

The scenarios range from AGI taking control of everything from on high (whether we want it to or not) – to a cascading satellite collision that would throw us back into the Stone Age.

In the middle of the scenario spectrum, massive amounts of data would controlled over our heads, tracking all of us. But for what purpose? And by whom? There wouldn’t be any escaping the grid, not even into the dark continent of Africa. Much the like the Matrix (1999) – the “grid” would be everywhere, both inside and outside. Who’s worried?

"For we wrestle not against flesh and blood, but against principalities, against powers, against the rulers of the darkness of this world, against spiritual wickedness in high places." Ephesians 6:12 ■