For most of the history of medical imaging, the scan cost more than the doctor. That is inverting. A full-body scan is now, in some cases, cheaper than the specialist appointment that would have ordered it, and the price is still falling. That is a genuine shift, and like most genuine shifts it arrives with a story attached that is partly true and partly a pitch.

A body scan for the price of a dinner. What changed?

Track one device and the collapse is clear. Ezra’s full-body MRI ran about 60 minutes at roughly $1,500. After Function Health acquired Ezra in May 2025, the same scan was cut to about 22 minutes at $499. That is time down by nearly two thirds and price down by two thirds, on one imaging line, inside about a year. Neko Health sells a different scan, roughly 60 minutes, for about $300 to $400, and reports more than 350,000 people waiting for one. Against the incumbent reference point, Prenuvo at about $2,500 for a full-body MRI, the new floor is a fraction of the old one.

What changed is that scanning is being re-engineered as a consumer funnel rather than a referral, with software cutting the scan time and volume cutting the price. The number to hold is the inversion: a scan that costs $300 to $500 is cheaper than many specialist visits, which reverses the usual order in which imaging is the expensive thing a cheap appointment orders. That reversal is what makes the category move, and it is real regardless of what one thinks of the medicine.

What the record shows:

  • Ezra’s full-body MRI went from about 60 minutes at roughly $1,500 to about 22 minutes at $499 after Function Health acquired Ezra in May 2025, per Function’s announcement.
  • Neko Health sells a roughly 60-minute scan for about $300 to $400 and reports more than 350,000 people on its waitlist, per the company; Prenuvo has long been cited at about $2,500 for a full-body MRI.

The most common failure: the price read as the whole story. Treating a falling scan price as if cost were the only variable that moved. The price fell; whether the scan should be done on an asymptomatic person is a separate question the price does not answer.

A billion scans a month. Built, or announced?

The newest entrant has no healthcare history at all. Midjourney, the AI image-generation company, announced a medical division in June 2026 and a whole-body ultrasound it calls Ultrasonic CT, described as using sound and water rather than radiation or magnets. Its founder’s company said it is investing over $74 million in the effort. The specifications are striking, and every one of them is a goal rather than a result.

This is the line to hold exactly. Midjourney states a goal of a scan taking no more than 60 seconds, a first location in San Francisco at the end of 2027, a fleet of more than 50,000 scanners by 2031 and a capacity of a billion scans a month. Those are targets, announced by a company that says it will spend the coming period refining its algorithms and hardware and running research trials. The device is not built. There is a concept video and a set of numbers, and the numbers describe an ambition, not a machine that has scanned anyone. An image-generation company announcing a medical device is a genuine event worth recording, and it is an announcement, which is a different kind of thing from a product.

What the record shows:

  • Midjourney announced a medical division and a whole-body ultrasound concept, Ultrasonic CT, in June 2026, stating goals of a 60-second scan, a first San Francisco location at the end of 2027, and later fleet and volume targets, per the company.
  • The device is unbuilt; the figures are stated targets, and the company describes forthcoming work on algorithms, hardware and trials.

The most common failure: the target quoted as a spec. Repeating an announced goal as though it were a measured capability. A billion scans a month is an ambition on a slide, and printing it without the word goal turns a plan into a fact that does not yet exist.

A cheaper scan finds more. Is finding more the same as helping?

The part the price and the pitch both skip is what a scan does when it works. Screening asymptomatic people with whole-body imaging turns up incidental findings at high rates, and most of them are nothing that needed finding. A systematic review of whole-body MRI for preventive screening pooled incidental findings at about 32 percent and false positives at about 16 percent. A separate meta-analysis of whole-body MRI in asymptomatic people put the cancer-detection rate at about 1.6 percent while noting that downstream biopsy rates ranged widely, evidence that most of what a scan flags leads to follow-up that ends in nothing.

That is the cost that does not appear on the price page. Each incidental finding can become another scan, a biopsy, a specialist visit and a stretch of fear, and academic medicine has been explicit that whole-body screening of the well is not the unmixed good it markets as. Cheaper scanning means more scanning, more scanning means more findings, and more findings means more of this cascade, for a detection rate that is real but small. None of this makes the scan worthless; early detection genuinely helps some people. It makes cheaper and more frequent a trade rather than a gift, and the trade belongs in the sentence next to the price.

What the record shows:

  • A systematic review of whole-body MRI for preventive screening pooled incidental findings at about 32 percent and false positives at about 16 percent, and a meta-analysis in asymptomatic people put cancer detection at about 1.6 percent with wide variation in downstream biopsy rates.
  • Cheaper, more frequent scanning increases findings and the follow-up they trigger, most of which resolves as nothing, which is a cost the price does not show.

The most common failure: the finding mistaken for a benefit. Assuming that detecting more is helping more. A scan that finds a great deal in a well person mostly finds things that would never have caused harm, and counting those as wins is how a cascade of tests gets sold as prevention.

What a record does instead

A record does not decide whether you should be scanned, and it does not clear the announcements at face value either. It does two smaller things. It keeps the price and time trajectory honest, so the collapse from $1,500 to $499 to $300 reads as what it is. And it holds the announced apart from the operating, which is why Midjourney Medical is not yet a clinic in the map.

The Atlas of the Healthspan Economy records Midjourney Medical as a watch item rather than an organization, because it is announced and not yet operating. That is the same rule that keeps a press release about a device off the map until the device exists. When the San Francisco location opens and a person can actually be scanned, it becomes an organization with a row. Until then it is a plan, recorded as a plan, next to the incumbents whose scans and prices are real and whose findings, cheaper by the year, still have to be reckoned with one anxious follow-up at a time.

The Atlas of the Healthspan Economy is a neutral record of organizations working on healthspan. It records operating providers and holds announced ventures as watch items until they open. It does not recommend for or against screening. See Ezra, Neko Health and the Healthtech & Diagnostics pillar, and read the methodology.

A note on what this piece does not claim. Midjourney’s specifications are stated goals for an unbuilt device, not achievements. Nothing here recommends or discourages a scan; the incidental-findings and overdiagnosis figures are from the cited literature and describe population averages, not any individual’s case.