On Wednesday morning in San Francisco, Midjourney CEO David Holz stood on a stage and announced something nobody had on their bingo card: a full-body ultrasound scanner. Not a new image model. Not a video generator. A piece of medical hardware that lowers a person through a ring of ultrasonic sensors submerged in water, producing a 3D body map in about sixty seconds.

The company known for making generative AI into a consumer product — the one that gave us six-fingered hands and dreamlike architecture renders — now wants to scan your internal organs at a strip-mall spa.

The immediate reaction from the tech press split into two familiar camps. Camp One: this is the weirdest pivot in AI history, a head-scratcher from a company that should be shipping V7. Camp Two: the FDA will eat this thing alive, and Midjourney has no idea what it’s stepping into.

Both takes miss what’s actually interesting here. Midjourney didn’t just announce a scanner. It announced a business model — one that treats diagnostic imaging the way Netflix treats video rental.

The Hardware Is Almost Beside the Point

Midjourney says it wants to deploy 50,000 scanners within six years and perform a billion full-body scans every month. The first flagship location, branded as a “Midjourney Spa,” is slated for San Francisco by the end of 2027. The partner providing the underlying ultrasound-on-chip technology is Butterfly Network — a publicly traded company whose stock popped on the news.

Notice what’s absent from the announcement: a CPT billing code. A reimbursement pathway. Any mention of an FDA 510(k) clearance or PMA submission timeline. Holz showed a prototype, not a regulatory filing.

This is not an oversight. It’s the strategy.

The entire premise of the Midjourney Scanner is that it bypasses the traditional medical billing apparatus. No prior authorization from UnitedHealthcare. No facility fee from the hospital system that bought up every imaging center in a twenty-mile radius. You walk into a spa, you pay a flat fee — or, more likely, a monthly subscription — and you get scanned.

One person close to the project, reached via Slack on Wednesday afternoon, described the pricing model under discussion as “somewhere between a Peloton membership and a boutique gym” — recurring, predictable, and entirely out-of-pocket. No insurance claim required.

Why People Pay Cash for Things Insurance Already Covers

Skeptics will point out that Americans already have access to diagnostic imaging through their health plans. This is technically true and practically misleading. The actual experience of getting an MRI or CT scan in the U.S. involves a referral from a primary care physician who is booked three weeks out, a pre-authorization fax that vanishes into a payer’s queue, a facility with parking that costs $12, and a bill that arrives four months later for somewhere between $400 and $4,000 depending on a deductible algorithm nobody can explain.

The median wait time for a routine MRI in a major U.S. metro area is north of a week. The median consumer understanding of what they’ll actually pay is zero until the explanation of benefits arrives.

Midjourney is betting that a subset of consumers — the same people who pay cash for direct-primary-care memberships, who use at-home lab tests from companies like Function Health, who wear continuous glucose monitors without a diabetes diagnosis — will pay a predictable monthly fee for body data they can access without asking a gatekeeper.

That’s not a medical device strategy. That’s a consumer subscription strategy wearing a lab coat.

The Real Competition Isn’t Siemens — It’s the Front Desk

Every conversation about this announcement will gravitate toward the technology. Is ultrasonic CT actually comparable to MRI? Can water-coupled ultrasound really produce diagnostically useful images at speed? Those are fair questions. Midjourney claims image quality comparable to MRI; independent validation is zero.

But the technology risk is the obvious risk. The overlooked risk — and the overlooked opportunity — is distribution.

Hospital systems do not want imaging to become a consumer product. Imaging is a profit center. A 2025 analysis by the Rand Corporation found that hospital outpatient imaging prices averaged 3.2 times what independent imaging centers charged for the same procedures. The gap exists because hospitals can bill facility fees and because insurers negotiate different rates for hospital-based services.

A stand-alone scanner in a spa, priced at a consumer subscription, sits outside that entire system. It doesn’t bill insurance, so it doesn’t need to negotiate with insurers. It doesn’t need a hospital’s chargemaster. It needs a storefront, a technician, and enough subscribers to cover the lease.

If the technology works — a significant if — the moat isn’t the hardware. It’s the direct-to-consumer relationship and the recurring revenue model. That’s what Midjourney, a company built on selling subscriptions to creative software, actually knows how to do.

The FDA Question Is a Distraction

Regulatory approval will be the chorus of every skeptical take this week. And yes, the FDA will eventually need to weigh in on whether a device that produces full-body 3D scans is a diagnostic tool requiring clearance.

But Midjourney can plausibly position the scanner as a wellness device in the near term — the same regulatory category that lets companies sell $300 at-home blood test kits without premarket approval, so long as they don’t make specific diagnostic claims. “Here is a high-resolution image of your body” is not the same sentence as “here is a cancer diagnosis.” The line is blurry, legally contested, and extremely useful for a company that wants to start scanning bodies before it finishes filing paperwork.

None of this means the Midjourney Scanner will succeed. Most hardware moonshots don’t. But treating this as a hardware story misses the thesis entirely. Midjourney isn’t betting it can build a better imaging machine than GE HealthCare. It’s betting that consumers are so fed up with the existing medical billing experience that they’ll pay cash for an alternative — and that the alternative doesn’t need to be better technology, just better access.

If that bet pays off, the lesson won’t be about AI or ultrasound. It will be about how many industries are one subscription model away from being disrupted by a company that doesn’t care about their reimbursement codes.

Sources