Midjourney Wants to Scan Your Entire Body in 60 Seconds. What a Time to Be Alive

This is Midjourney Medical, and its first product is the Midjourney Scanner. It is the kind of swing-for-the-fences project that, a decade into the crypto-and-frontier-tech era, you might think we would be jaded about. We are not. This one is genuinely thrilling.

What it actually does

Strip away the spa lighting and the concept renders and the engineering claim is specific. You stand on a platform that descends into a tank of water at about five centimeters per second. As you sink, you pass through a ring made of roughly half a million tiny elements, each the size of a grain of sand, each able to act as both a speaker and a microphone. They fire ultrasonic waves through you from every angle and listen to how those waves deform as they cross boundaries between water, skin, fat, muscle, and bone.

The result, Midjourney says, is a volumetric image that looks a lot like an MRI, reconstructed at close to a hundred times the speed, Source: MidJourney

The sensors generate terabytes of data per second — the company’s own analogy is that one second of scan data would take 500 hours to watch as HD video. Turning that torrent of noisy acoustic reflections into clean anatomical images is, by Midjourney’s admission, the hard part it has not fully solved yet. That candor is, if anything, reassuring. This is a research lab telling you where the frontier actually is, not a pitch deck pretending the work is done.

The technique itself is not fringe. Ultrasound tomography is a real and rapidly improving field, and academic groups have recently published on reconstructing full in-vivo cross-sections of the human abdomen and thighs using sound alone. What Midjourney is proposing is to take that science and weaponize it with scale, custom silicon, and the kind of AI-driven reconstruction pipeline the company is unusually well-positioned to build.

The unlikely hardware partner

Midjourney is not doing the physics from scratch. The scanner’s current prototype runs on 40 Ultrasound-on-Chip modules per system, licensed from Butterfly Network (NYSE: BFLY), the semiconductor-ultrasound company whose handheld probes already exist in the real world. Butterfly disclosed the co-development agreement in a November 2025 SEC filing that points to up to $74 million in expected payments over a five-year term, and future scanner generations are expected to pack in many more modules.

Butterfly chief executive Joseph DeVivo, who has spent decades in medical devices, framed the appeal in preventive terms, arguing the US healthcare system is still built to treat illness rather than prevent it. He called the device a continuous window into your health, designed for weekly use. Coming from a 35-year industry veteran rather than from Midjourney’s own marketing, that endorsement carries weight — and it tells you the partner supplying the chips believes the use case is real, not just cinematic.

The project is led by Ahmad Abbas, Midjourney’s head of consumer hardware, who joined in late 2023 after working on Apple’s Vision Pro. The hardware pedigree matters here, because the gap between a striking demo and an FDA-cleared diagnostic device is precisely where ambitious health-tech projects go to die.

The spa is the strategy

Here is the part that sounds like a joke and is actually the most interesting business decision in the whole announcement. Midjourney is not building clinics. It is building spas.

The first Midjourney Spa is slated to open in San Francisco around the end of 2027, with hot tubs, saunas, cold plunges, and rooms where pools of golden light quietly scan you. The pitch is that the scan becomes a side effect of somewhere you already want to be. You go for the warm water and the company; you leave with a longitudinal dataset about your own body.

The single biggest barrier to preventive imaging has never been the physics — it has been that nobody volunteers for an MRI on a Tuesday for fun. By wrapping the machine in an experience people actively enjoy and can use repeatedly, Midjourney is attacking the behavioral problem, not just the technical one. Reframe the scariest part of healthcare as a treat and you change who shows up.

The numbers that make you sit up

Midjourney’s stated ambition is a fleet of more than 50,000 scanners worldwide by 2031, with capacity for a billion scans a month — enough, the company claims, to give regular monthly scans to a billion people. Founder David Holz has suggested that fewer than a dozen of these machines running at full tilt could perform more full-body scans than every MRI on Earth combined.

Then there is the headline claim: that with enough early imaging, the world could eventually avoid 30% of all deaths and 50% of all healthcare costs. That is a staggering, almost certainly over-optimistic figure, and it is worth holding it at arm’s length. But the underlying logic — that catching disease early and cheaply is enormously more valuable than treating it late and expensively — is one almost every clinician would sign off on.

The cost argument is the most plausible part. Because the hardware is comparatively cheap and the scan is fast, Holz has floated per-scan costs hundreds or thousands of times below an MRI. If even a fraction of that holds, the economics of routine imaging change completely.

Why this matters beyond Midjourney

What makes this story resonate for anyone watching frontier technology is the pattern it confirms. The most capable AI labs are no longer content to ship models inside a chat window. They are reaching into physical infrastructure, semiconductors, and the foundations of the human experience itself.

We saw the geopolitical edge of that shift this month, when a single export control directive pulled two of the most powerful Anthropic AI models offline worldwide overnight — an episode that has supercharged the decentralized-AI thesis and reminded everyone how concentrated this power has become. Midjourney’s announcement is the optimistic mirror image of that same trend: the same wave of intelligence and compute, pointed not at control but at something close to a public good.

It is also worth noting how Midjourney intends to pay for this. The company famously has no outside investors and describes itself as a community-backed research lab funded by everyday users. A profitable consumer AI product underwriting a moonshot medical-imaging program is its own quietly radical funding model — one that sidesteps the venture-capital treadmill entirely.

The honest caveats

None of this is built yet. The hardest computational problem remains unsolved, the FDA pathway has barely begun — Midjourney is starting with non-diagnostic body-composition maps and submitting results for expanded capabilities over time — and the privacy implications of a company harvesting terabytes of intimate biological data per person are not trivial. Skeptics on Hacker News and elsewhere have rightly pointed out that “envisions” is doing a lot of work in the announcement.

But every transformative technology started as a confident sketch of something that did not exist. The internet, the smartphone, the AI models we now talk to daily — all of them looked faintly ridiculous at announcement.

A company that taught machines to dream in images now wants to teach them to see inside us, and to make that as casual as a soak in warm water. Whether or not Midjourney hits its dates, the mere fact that this is the project a frontier AI lab chose to attempt next is reason for genuine optimism. What a time to be alive.