An AI Image Company Just Built a Body Scanner
On June 17, 2026, Midjourney announced its first piece of hardware, and it is a medical device. The company unveiled Midjourney Medical, a new division built around a full-body ultrasound scanner it calls Ultrasonic CT, and said a whole-body scan will take as little as 60 seconds, using only sound and water, with no radiation and no magnets [1] [2]. Chief executive David Holz set the ambition as high as it goes. The plan is a fleet of 50,000 scanners by 2031 running a billion full-body scans a month, starting with a single location in San Francisco that opens at the end of 2027 [1] [2].
That is a long way from generating pictures.
However, it is worth noting that MidJourney was not just an image generation company, but arguably the best AI image generation software available. They have done over $500 million dollars in revenue with no VC funding.
Midjourney built the scanner with Butterfly Network, a public ultrasound-chip company, and Butterfly disclosed the deal in a securities filing back in November 2025, worth up to $74 million over five years [6]. The reveal is new. The work is more than a year old.
This piece sorts the genuinely new from the oversold, and then gets to the part that matters for you: what a 60-second body scan can actually tell you, and who reads the result.
What Midjourney Announced
Midjourney's own framing is the cleanest summary of the pitch. The company said it "dreamed of something as powerful as MRI, and as casual as a trip to the spa," and presented the scanner as the path to it [2]. Magnetic resonance imaging (MRI) is the high-detail hospital scan the comparison leans on.
The specific claims are these:
A whole-body scan in about 60 seconds, against the 15 to 90 minutes an MRI can take [1] [2].
No ionizing radiation and no strong magnets. The scanner uses sound waves moving through water [1].
Resolution down to about half a millimeter, which Midjourney says produces images that "look a lot like today's MRIs but at nearly a hundred times the speed" [2] [5].
50,000 scanners worldwide by 2031, with capacity for a billion scans a month, which the company frames as "regular, monthly scans" for a billion people [1] [2].
Holz did not hedge. "No such device has ever been built until now," he said, and called it the first new whole-body imaging method in 50 years [3] [5]. He was candid about one thing, which is striking for an artificial intelligence (AI) company: the scanner "does not use AI yet," in his words, just "really cool hardware and software" [3] [5]. Midjourney's announcement also floated the largest claim of the day, that "with enough early imaging in the future, the world could avoid 30% of all deaths and 50% of all healthcare costs" [2]. That is an aspiration, not a finding, and the company presented it as one.
What Midjourney Actually Is
Midjourney builds AI models that generate images and video from written prompts. Customers pay a subscription, roughly $10 to $120 per month, to make pictures and short clips [3]. The company was founded in 2022 by David Holz, and it describes itself as a community-funded research lab of about 60 people with no outside investors, paying for its work out of subscription revenue [2] [3]. That independence is unusual for a company making hardware claims at this scale.
The medical scanner is one of eight projects Midjourney says it is pursuing, four hardware and four software, and the company plans to ship at least two of the hardware efforts before long [3]. Medical imaging is the one it chose to show first.
Midjourney also carries real legal weight on the image side. Warner Bros. Discovery and Walt Disney have both sued the company, claiming its models reproduced their copyrighted characters [3]. A company financing a medical-hardware moonshot out of subscription cash, while fighting two major studios in court, is taking on a great deal at once.
How the Scanner Is Supposed to Work
Midjourney describes the experience in deliberately cinematic terms. You step into "a shallow pool of golden light," a platform lowers you into the water at about five centimeters per second, and your body passes through a ring of sensors that each work "like a dolphin, using its echolocation" [2].
The engineering underneath is real and large. The ring holds roughly half a million tiny elements, each able to act as both a speaker and a microphone, firing ultrasonic waves and recording the echoes from every angle [2]. Built across 40 imaging modules from Butterfly Network, the system captures data at around 17 gigabytes per second and produces on the order of 800 terabytes of raw data per scan, which a cluster of 21 on-site servers reconstructs into images using about two petaflops of computing power [5] [6]. Midjourney's own comparison: one second of scan data equals about 500 hours of high-definition video [2].
The chip partner is the most concrete thread in the story. Butterfly Network (NYSE: BFLY) makes "ultrasound-on-chip" technology, which shrinks an ultrasound system onto a semiconductor, and was founded in 2011 by Jonathan Rothberg [6]. The Midjourney scanner uses "40 Butterfly Ultrasound-on-Chip imaging modules per system," licensed through Butterfly's chip-licensing business [6]. Butterfly chief executive Joseph DeVivo went on record calling the scanner "designed for weekly use" and "a continuous window into your health," and said Butterfly is "proud to support Midjourney's mission to democratize access to personal imaging data" [6]. Investors noticed. Butterfly shares rose about 17% the morning the news spread, according to one market report [7].
Two facts keep the hype honest. Butterfly's own filing lists the risks plainly, including whether it can make the chips "at the scale, quality, and cost required" and whether Midjourney can win regulatory clearance "on the anticipated timeline or at all" [6]. And the current machine is a first-generation prototype: it takes about 20 minutes per scan today, a team of roughly nine people has scanned about a dozen people, and the 60-second target waits on a second-generation design due in late 2026 and a third generation on custom silicon after that [5].
Is This Actually New?
Holz's "no such device has ever been built" claim deserves a careful check, because it is half right.
Ultrasound tomography, the method of ringing a body part with transducers in water and reconstructing an image from sound passing through it, is not new. It is a research field that goes back about 50 years, to work by James Greenleaf and others in the 1970s [11]. More to the point, the specific design Midjourney describes is already approved and on the market for breast imaging:
Delphinus SoftVue uses a 360-degree ring transducer with the breast suspended in water, no compression and no radiation. The Food and Drug Administration (FDA) approved it in 2021, after a first clearance in 2014 [10].
QT Imaging (Nasdaq: QTI) scans a breast suspended in a warm water bath, ringed by a 2,048-element array doing transmission and reflection ultrasound with software-reconstructed 3D images. The FDA cleared it in 2017 [8] [9].
Researchers have already pushed the same approach past the breast. A Caltech group led by Lihong Wang published in-vivo ultrasound tomography of the human abdomen and legs, using a circular transducer array in reflection and transmission modes [11].
So the honest framing is straightforward. The underlying physics is established, and water-immersion ring tomography has been FDA-cleared for breast imaging for roughly a decade. What would be genuinely new, if Midjourney ships it as described, is the engineering package: whole-body coverage, a 60-second scan, and a consumer-scale price. The technique is not the novel part. The scale and speed are.
What Ultrasound Is Good At
Ultrasound is old, cheap, and one of the safest tools in medicine. It uses sound instead of radiation, it works in real time, and the FDA notes it "has an excellent safety record" with no ionizing-radiation risk [20]. It is also genuinely useful. When a patient has a palpable lump, ultrasound is often the first tool a doctor reaches for, because it can tell a fluid-filled cyst from a solid mass and flag a lymph node that looks abnormal.
Sound does its best work on soft tissue, fluid-filled structures, and anything close enough to the surface for the signal to get in and useful information to come back out. That is why ultrasound is a standard tool for the thyroid, a breast lump, the testicles, lymph nodes, the gallbladder, the kidneys, a developing fetus, blood flow, and much of the abdomen.
A whole-body version could be strong at exactly those things, at a scale medicine has never had. Body composition is the clearest example. Today the reference method for measuring fat, muscle, and visceral fat is a dual-energy X-ray absorptiometry (DEXA) scan [21], and a safe, repeatable ultrasound map of body composition, muscle mass, early liver changes, the kidneys, and some blood vessels would be a real addition. One variable shapes every scan, though: fat between the sensor and the target scatters sound, so leaner bodies tend to image more cleanly than others.
Different Scans Do Different Things
Ultrasound is safe, cheap, real-time, and strong on soft tissue, fluid, and surface-accessible structures. It is weak on the lungs and bone.
CT is fast and excellent at the lungs. The US Preventive Services Task Force (USPSTF) recommends annual low-dose CT, not ultrasound or MRI, to screen for lung cancer in older adults with a heavy smoking history [17].
MRI gives rich soft-tissue contrast and is the standard for the brain. It is unaffected by bone, but its images are degraded by motion, including breathing and bowel movement, which makes some moving organs hard to capture [22].
Colonoscopy does something no scan can. It sees inside the colon and removes precancerous polyps in the same visit. Removing those polyps reduced colorectal cancer deaths by an estimated 53% in the long-running National Polyp Study, so it prevents cancer rather than only spotting it [18].
Elective full-body MRI companies publish long lists of what their scans can find, and the lists are real. What the lists leave out is what the scan is bad at. A scan that misses early lung and colon cancer, two of the deadliest, is not a clean bill of health. The risk in a glossy "one scan to know everything" message is that someone walks out reassured when all they actually learned was a few things.
Adding whole-body ultrasound does not replace CT, MRI, or colonoscopy. Each still fills a gap the others cannot. The reasonable hope is that it adds a new gap-filler.
A Spa, Not a Clinic
The launch location is not a clinic. It is a "Midjourney Spa," about 25,000 square feet across four floors at 300 Grant Avenue near Union Square in San Francisco, holding nine to ten scanners alongside hot tubs, saunas, and cold plunges, with the lease signed and designs underway [2] [5]. The company says it should be open around the clock, and that "the scans are a side-effect" of a spa visit rather than the reason for it [2]. Expansion to more cities is slated for 2028 [2].
Selling a body scan as a spa visit places Midjourney inside a market that already exists. Elective full-body MRI screening for healthy people has grown for years:
Prenuvo charges $2,499 for a whole-body MRI, runs about 21 clinics, was backed by Cindy Crawford and Rande Gerber, and went viral after Kim Kardashian called it a "life-saving machine" on Instagram, an endorsement radiologists criticized [12] [13].
Ezra, acquired by Function Health in 2025, launched a $499 full-body MRI at the time of the deal and now lists scans starting around $999 [14].
There is a constructive way to read this. The first customers will be people paying out of pocket for data about their own bodies. If those scans are collected at scale, researchers gain a large new pool of imaging, and the field can learn what cheap, repeatable, whole-body ultrasound is actually good for, including uses no one has guessed yet.
The harder question is whether scanning healthy people helps them, and here the evidence is not kind to the pitch. The American College of Radiology (ACR) states plainly that it "does not believe there is sufficient evidence to justify recommending total body screening" for people without symptoms or risk factors, and warns that such scans mostly turn up "non-specific findings that will not ultimately improve patients' health but will result in unnecessary follow-up testing and procedures" [15]. A peer-reviewed review of whole-body MRI across 5,373 asymptomatic people found that about 32% had a critical or indeterminate incidental finding, with a pooled false-positive rate near 16%, and only about 13% of flagged findings turned out to be real disease [16]. No major guideline body recommends whole-body screening for average-risk adults. The recommended approach is the opposite: targeted screening by organ and risk, such as low-dose CT for high-risk smokers and colonoscopy for colorectal cancer [17] [18].
Comparing You to You
One advantage of a cheap, safe, repeatable scan gets lost in the argument over whether it beats MRI, and Butterfly's chief executive named it directly when he called the scanner "designed for weekly use" and "a continuous window into your health" [6]. Holz framed the same idea personally, saying he is "not the most measured man on Earth yet," but that he might want that kind of daily read on his own body [4].
Most of medicine has to compare how you look today against how people usually look. A finding is "probably fine" because it resembles a population average. A scan cheap and safe enough to repeat every year changes the comparison. It can measure you against your own past. Did that spot grow, shrink, or shift since last year? Is your visceral fat trending up? Did a kidney measurement change?
Change over time is often more informative than a single snapshot, and it is the kind of data medicine rarely gets, because most scans cost too much or carry too much radiation to repeat casually. If whole-body ultrasound delivers safe, repeatable imaging at scale, that year-over-year view could become a real part of how prevention works. Whether people actually scan themselves year after year is a separate bet, and plenty of useful health habits go unused. The capability would still be new, and it could matter.
What This Means for Your Doctor
Here is the part that matters if you are deciding what to do with your own health. A scan is an input. It produces an image and a set of measurements. By itself, an image does not know your history, cannot examine you, and cannot decide what to do next.
Consider the most common real outcome of a full-body scan, the one the evidence above describes: it finds something small. A 6-millimeter nodule on a kidney. A cyst on the liver. A spot on the thyroid. For a healthy adult, the large majority of these are harmless. Whether to watch it, re-test it in six months, or refer you to a specialist this month depends on your age, your family history, your symptoms, and your other lab work. That judgment is medicine, and it is what a physician who knows you provides.
This is why the concierge and direct primary care (DPC) model fits the moment. A concierge physician carries a small patient panel, often under 300 people, and a DPC doctor up to around 800, against 2,000 to 2,500 for a typical primary care doctor. The smaller panel buys time, which is what it takes to read this year's scan against last year's, explain what a finding means, and keep you from a cascade of unnecessary procedures. A spa technician handing you a report cannot do that. A doctor with a 45-minute appointment and your full history can.
The pattern is the same one across consumer health hardware. When Oura added a doctor connection to its smart ring, the device still could not manage a chronic condition or coordinate care. It produced data and routed you somewhere. A body scanner does the same thing at higher resolution. The data is only as useful as the physician who reads it and owns what happens next.
One more issue belongs in plain view: the data itself. A billion scans a month would be one of the largest medical-imaging datasets ever assembled, held by a private company. Midjourney's stated goal is to "democratize access to personal imaging data" [6], and its hiring lists standard health-data protections under the Health Insurance Portability and Accountability Act (HIPAA). As of launch, though, there is no consumer-facing statement about who owns your scans, whether you can delete them, or whether they train Midjourney's models. That is a fair question to ask before you step into the water.
The Patient Takeaway
For now, Midjourney Medical is a prototype with a 2027 opening date, not something you can use. The smart way to read the announcement is as a signal about where wellness is heading, not as a product to wait in line for.
If you are healthy and curious, watch this space, and treat any future scan the way you should treat any screening test. The scan is a starting point. Before you act on a result, take it to a physician who knows your history and has time to think about it with you. If you do not have that kind of doctor, that is the gap worth closing first, well before a body scanner opens near you. You can compare what concierge medicine actually costs and includes, or read what personalized medical care really means in 2026.
Hold onto one idea above all. Different scans do different things, and a new one does not retire the old ones. Whole-body ultrasound could add something real, especially a safe scan you can repeat year after year. A machine can take a picture inside your body in 60 seconds. Deciding which scan answers your question, what the picture means, and what to do next is still a job for a doctor who knows you.
FAQ
What is Midjourney Medical?
Midjourney Medical is a new division of the AI image company Midjourney, announced on June 17, 2026 [1] [3]. It is building a full-body ultrasound scanner the company calls Ultrasonic CT, which it says will image the body in about 60 seconds using sound and water, with no radiation or magnets [1] [2].
How does the Midjourney scanner work?
You stand on a platform that lowers you into a shallow pool of water and pass through a ring of roughly half a million tiny ultrasonic elements that send and receive sound waves from every angle [2]. A cluster of servers then reconstructs a 3D image of your internal anatomy [5]. It is built on 40 Butterfly Network ultrasound-on-chip modules per system [6].
Is the Midjourney scanner as good as an MRI or CT scan?
No, and the comparison is the wrong way to think about it. Different scans answer different questions. Ultrasound is strong on soft tissue, fluid, and surface-accessible organs, but it cannot see through air-filled lungs or bone [19]. A low-dose chest CT is what catches early lung cancer, and a colonoscopy can remove precancerous polyps [17] [18]. Whole-body ultrasound does neither, so it would add a tool rather than replace the others.
Is the Midjourney body scanner available now?
No. As of June 2026 it is a first-generation prototype that currently takes about 20 minutes per scan, and roughly a dozen people have been scanned [5]. The first location, a "Midjourney Spa" in San Francisco, is scheduled to open at the end of 2027 [1] [2].
Is the Midjourney scanner FDA approved?
No. Midjourney says its first step is to provide body-composition maps and to submit results to the Food and Drug Administration for expanded capabilities over time [2]. It has not published accuracy data, and it is not cleared for diagnostic medical use. Notably, similar water-immersion ultrasound tomography devices are already FDA-cleared for breast imaging [9] [10].
Should I get a full-body scan to check for cancer?
Talk to your doctor first. Major bodies including the American College of Radiology do not recommend whole-body scans for healthy, average-risk adults, because they mostly find harmless spots that lead to more testing [15] [16]. Proven screening is targeted by organ and risk, such as low-dose CT for high-risk smokers, mammography, and colonoscopy [17] [18].
Find a Concierge Doctor
A scanner can take a picture of your insides. A doctor who knows you can tell you what it means and what to do next. Use NextMD to find concierge and direct primary care practices in your city, compare pricing, and view physician credentials.
Search concierge doctors near you on NextMD →
Sources
Midjourney. (2026). Midjourney Medical. Midjourney. Read on Midjourney
Midjourney. (2026). A New Era of Midjourney (Midjourney Medical announcement). Midjourney. Read on Midjourney
Welch, C. (2026). AI startup Midjourney pivots to health with 'full-body ultrasound machine'. Bloomberg, via Business Standard. Read on Business Standard
Lawler, R. (2026). Midjourney's next act is a full-body medical scanner. The Verge. Read on The Verge
Latent Space. (2026). Midjourney Medical: scan your organs like you step on a scale. Latent Space (AINews). Read on Latent Space
Butterfly Network. (2026). Butterfly Network Provides Commentary on Midjourney Medical's Full Body Ultrasound Scanner Announcement. Business Wire. Read on Business Wire
Investing.com. (2026). Butterfly Network stock jumps on Midjourney Scanner unveiling. Investing.com. Read on Investing.com
QT Imaging. (2026). Breast Acoustic CT Scanner. QT Imaging Holdings. Read on QT Imaging
U.S. Food and Drug Administration. (2017). 510(k) Summary K162372 (QT Ultrasound Breast Scanner). FDA. Read the FDA clearance (PDF)
U.S. Food and Drug Administration. (2021). Premarket Approval P200040 (Delphinus SoftVue). FDA. Read the FDA approval record
Garrett, J., Xu, Y., Aborahama, Y., Ku, G., Maslov, K., & Wang, L. V. (2023/2025). Whole Cross-Sectional Human Ultrasound Tomography. arXiv. Read on arXiv
Prenuvo. (2026). Pricing. Prenuvo. Read on Prenuvo
Lawrence, L. (2023). Kim Kardashian's full-body MRI endorsement draws scrutiny. STAT. Read on STAT
Function Health. (2025). Function Health Acquires Ezra, Introduces $499 Full-Body MRI Scan. PR Newswire. Read on PR Newswire
American College of Radiology. (2023). ACR Statement on Screening Total Body MRI. ACR. Read on the ACR
Kwee, R. M., & Kwee, T. C. (2019). Whole-body MRI for preventive health screening: A systematic review of the literature. Journal of Magnetic Resonance Imaging, 50(5), 1489–1503. Read on PubMed
U.S. Preventive Services Task Force. (2021). Lung Cancer: Screening. USPSTF. Read on the USPSTF
National Cancer Institute. (2022). Colonoscopy Reduces Risk of Death from Colorectal Cancer. NCI. Read on the NCI
Radiopaedia. (2026). Reflection (ultrasound physics). Radiopaedia.org. Read on Radiopaedia
U.S. Food and Drug Administration. (2026). Ultrasound Imaging. FDA. Read on the FDA
Messina, C., et al. (2020). Body composition with dual energy X-ray absorptiometry: from basics to new tools. Quantitative Imaging in Medicine and Surgery, 10(8), 1687–1698. Read on PMC
Radiology (RSNA). (2023). Respiratory Motion Management in Abdominal MRI. Radiology. Read on RSNA