AI and Astrology: When Algorithms Read the Stars (and Companies Read the Future)

I don't check my horoscope. That, I think, is the right vantage point from which to look at what has quietly become a multi-billion-dollar industry — because the interesting question about astrology today isn't whether it's true. It's why it sells so well in a thoroughly technological age, and what its success reveals about the technology we actually use every day.

The numbers are real even if the stars aren't. The global market for astrology apps is projected to reach somewhere between $9 billion and nearly $12 billion by 2030, growing at close to 20% a year — enough to make a venture capitalist's eyes shine brighter than any constellation. A 2019 MTV survey of people aged 14 to 29 found three-quarters said they trust that astrology works.

Precision vs. validity: the Co-Star paradox

The clearest specimen is Co-Star, the app that made horoscopes cool again for millennials and Gen Z, and its entire pitch is built on technical sophistication. It doesn't want just your sun sign; it demands your exact birth minute and precise coordinates, then uses real ephemeris data from NASA's Jet Propulsion Laboratory to compute the actual position of every planet at the moment you were born. It feeds that into an algorithm that generates hyper-personalized, natural-language notifications — not "good day for Pisces," but "with Saturn transiting your house of communication, you may feel constrained in expressing your ideas today."

Here's the paradox worth sitting with: the astronomy is genuinely accurate. Co-Star pulls real, precise planetary data from one of the best space agencies on Earth — only to feed it into a framework that has never once survived a controlled scientific test. The inputs are NASA-grade; the model they feed is worth nothing. It's a monument to the difference between the precision of data and the validity of a theory.

Even Silicon Valley reaches for the stars

This instinct to wrap technology in cosmic archetypes isn't limited to startup apps. When Google launched its flagship AI under the name Gemini, on 8 February 2024, it chose — deliberately or not — the name of a zodiac sign whose stereotyped traits (quick-thinking, communicative, mercurial, a little glib) read suspiciously like a spec sheet for a chatbot.

The official corporate story is mundane: Gemini means "twins," marking the merger of two of Google's AI teams, with a nod to NASA's two-seater Project Gemini. No astrologer was consulted. But that's exactly the point — even a trillion-dollar engineering company reached for a cosmic archetype because it resonates, because the celestial frame sells. The same instinct Co-Star monetizes one push notification at a time.

And the universe set up a free joke: Gemini is ruled by Mercury — the trickster, the planet of communication, pranks, and deception. Google named its talking machine after the sign of the smooth talker — and sounding convincing whether or not it's right is the defining trick of the whole genre, this model no more than the rest.

The real engine: the Forer effect

Strip away the NASA data and the corporate branding and you find the same psychological engine powering both Co-Star and generative AI: the Forer effect — our readiness to accept vague, general statements as precise personal insight, as long as we believe they were written for us.

Co-Star runs it through push notifications calibrated to sound profound: "Your vulnerability is your strength today." "You can't control others, only your reaction to them." True of you, and of everyone, on any day — yet delivered so that it feels like the universe is speaking to your specific struggle. The AI chatbot runs the identical play in a different costume, answering in a fluent, attentive, apparently empathetic register that produces the illusion of a mind that truly gets you, while serving responses that, with minor substitutions, would land for almost anyone. Both pull the same lever: the human hunger to be seen, to find a thread of meaning in the noise. Stars or neural network, we're always shopping for something that will look us in the eye and say "I get you."

And here's the part that should unsettle anyone building a business on large language models: the better the model gets, the better it performs this trick. More fluency, more apparent warmth, more "it really understands me" — which is precisely the Forer effect running at higher resolution. Capability and the illusion of being understood rise together, and they are not the same thing.

The true commodity

While the apps compute planetary positions to four decimal places, the underlying claim has been tested and has failed. Decades of studies — from physicist Shawn Carlson's 1985 double-blind test in Nature to the long-running "time twins" research following people born moments apart — found that astrologers cannot match birth charts to personalities any better than chance. The verdict has been in for decades, and it changed nothing, because empirical truth was never the product.

So whether it's Co-Star turning NASA data into a daily affirmation or an LLM answering you in tones of bottomless understanding, the commodity is identical: the feeling of being understood. Not understanding — the feeling of it. Astrology has sold that feeling for four thousand years; large language models just sell it faster, cheaper, and around the clock. The smart move isn't to sneer at the people buying it — the hunger is universal, and I feel it too. The smart move is to recognize when you're being sold the sensation of being known, and to keep a hand on your wallet, whether the pitch comes from the cosmos or the autocomplete.

Sources

• Astrology app market forecasts of roughly $9–12 billion by 2030 at a ~20% CAGR (industry estimates, 2025).
• MTV Insights survey (2019, n = 1,000, ages 14–29): 75% said they trust that astrology works.
• Co-Star's use of NASA Jet Propulsion Laboratory ephemeris data (company materials).
• Shawn Carlson, "A double-blind test of astrology," Nature (1985).
• Google, origin of the Gemini model name (merger of AI teams; NASA's Project Gemini; Jeff Dean).