What Does a Spa Body Scanner Have in Common With a Horoscope App?

Midjourney is building a full-body ultrasound scanner for spas. It has also bought Co-Star, the astrology app that tells people what their day means based on where the planets were when they were born. One product is attempting to extract real information from the human body using physics, sensors and medical imaging. The other packages a belief system with no established scientific basis into one of the most successful consumer apps of its kind.

What does a spa body scanner have in common with a horoscope app? In the age of AI, they can be projects inside the same company.

The scanner is not a vague concept attached to a mood board. Midjourney says it has built a first-generation prototype and is now developing what it calls Ultrasonic CT. A person stands on a platform, descends into warm water and passes through a ring of ultrasound sensors. The system records how sound waves move through the body and attempts to reconstruct them into a three-dimensional map.

Midjourney does not want the first scanner to feel like a hospital machine. It plans to place ten of them inside a San Francisco spa with saunas, hot tubs and cold plunges. The company imagines people getting scanned repeatedly, not necessarily because something hurts, but because body imaging could become part of an ordinary wellness routine. Its current regulatory scope is body composition rather than diagnosis, which is an important distinction, but the long-term ambition is clearly medical.

Then there is Co-Star. Midjourney announced the acquisition in July 2026 and said the app would remain under founder Banu Guler's control while she and her team helped build a broader product and design organization around Midjourney's other projects. Co-Star takes birth date, time and location, applies the logic of astrology and turns it into notifications about personality, relationships and decisions.

The scanner and horoscope app do not merely come from different industries. They operate under opposite standards of truth. The scanner eventually has to show that it measures something real, does so reliably and gives people information that survives comparison with established medical imaging. Co-Star does not need to prove that the planets shape personality. It needs the interpretation to feel specific enough that people keep opening the app.

One product has to be correct. The other has to resonate.

A few years ago, this combination would have been difficult to imagine inside one company. A medical imaging startup would have been built around researchers, hardware engineers, regulatory specialists, clinical studies and a narrow technical claim. A horoscope app would have been built around consumer product design, cultural intuition, writing, notifications and social behaviour. Even if both became successful, nobody would have searched for a strategic reason to merge them.

Companies were expected to have a category. They were software companies, medical device companies, media companies or consumer apps. They could expand, but the expansion was supposed to make sense. Adobe could move from Photoshop to Illustrator because both products belonged to creative software. A medical imaging company could build another scanner because the science, customers and regulation were adjacent.

Midjourney appears less interested in adjacency. It began with image generation, added video, built a medical hardware division, bought an astrology app and says several more ambitious projects will be announced. David Holz described this as the slow blossoming of a very unusual research lab. That description may be more useful than trying to force everything into a normal product roadmap.

The image generator is the cash engine. It gave Midjourney money, talent, models, infrastructure and the freedom to explore. AI then lowers the cost of moving from curiosity to prototype. A team can read a field faster, write software faster, generate interface directions, simulate experiences, produce visual concepts and test more possibilities before committing to a full organization.

AI does not remove the need for ultrasound researchers, medical validation or regulatory work. Midjourney's medical careers page is full of roles for electrical engineering, signal processing, systems integration, operations and quality. The company still has to confront water, bodies, noise, artifacts, manufacturing and the FDA. AI makes it easier to enter the field. It does not make the field stop being difficult.

That difference matters. The most interesting change is not that one AI model can magically do medicine and astrology. It is that an AI-native company can carry several very different teams, each working on a problem that once would have required a separate startup. The shared asset is no longer only a technology or customer base. It is the ability to learn, prototype, recruit and coordinate across fields quickly enough that the company does not need to choose one identity.

This creates something between a startup, a research lab and a small conglomerate. The company can hold projects that do not share customers, business models or standards of evidence. They only need to share the same source of funding and the same group of people deciding what is interesting enough to pursue.

There is something exciting about that. Most startups are forced to repeat one idea until the market, investors or founders become tired of it. Midjourney can use the success of image generation to build projects that would struggle to survive the normal fundraising process. A full-body ultrasound scanner inside a spa is exactly the kind of idea that sounds too strange, expensive and cross-disciplinary for a conventional startup pitch.

There is also something unsettling about it. Midjourney became successful by generating images that look convincing. Co-Star became successful by generating interpretations that feel convincing. Medical imaging is a field where convincing is not enough.

A horoscope can be wrong without producing a visible failure. The user may forget it, reinterpret it or find another sentence that feels accurate. A body scan can create false reassurance, unnecessary fear or misplaced trust if the image looks authoritative but the underlying measurement is not reliable. The same company can therefore build one product where emotional plausibility is the experience and another where emotional plausibility must never substitute for evidence.

This is not only a Midjourney story. AI is making it possible for more companies to become collections of bets rather than expressions of one stable industry. A profitable product can fund several internal startups. Small teams can travel further into unfamiliar fields before they need large budgets. Founders can organize companies around curiosity, taste and appetite rather than a single market category.

The old question was: what business is this company in? The answer was supposed to help predict what it would build next.

That question becomes less useful when the same company can use science to scan your body and astrology to explain your future. Midjourney's spa scanner and horoscope app are not evidence that the two products secretly belong to one category. They are evidence that AI companies may no longer need a category at all.

Connected reading

• Why Is AI Making Me Think About Business Strategy?
• The AI Was Easy. Understanding the Company Was Hard.
• The Prototype Has Become Too Cheap to Trust
• The Algorithm Thinks Your Past Is Your Personality
• The System Around the Medical Model

AI authorship note

This essay began with Kris Haamer noticing that Midjourney could seriously pursue a science-based body scanner and acquire a horoscope app at the same time. Kris rejected attempts to invent a false philosophical connection between the products and pushed the argument toward the changing shape of companies in the age of AI, especially the contrast between a product that must be scientifically correct and one that only needs to feel personally accurate. AI supported research, drafting, editing and implementation. Kris shaped the angle, title, voice, paragraph structure and final publication decision.