David Holz stood before an audience in San Francisco on June 17 and unveiled hardware his company had never built before. The founder of Midjourney, famous for conjuring surreal images from text prompts, introduced the Midjourney Scanner. A person steps into a shallow pool. A platform lowers them slowly through water. Sensors fire ultrasonic waves from every direction. Sixty seconds later, out comes a detailed 3D map of the body.
The claims came fast. Superior to MRI machines in many respects. No radiation. No massive magnets. Nearly 100 times faster. A full-body scan in about one minute instead of an hour or more. Yet the twist arrived almost immediately. “We’re not even using any AI in this yet, just really cool hardware and software,” Holz said, according to Bloomberg.
The imaging relies on ultrasound technology and signal processing. Artificial intelligence enters later. It segments and labels the resulting views. For now the output consists of detailed body composition maps. Anything closer to medical diagnosis awaits regulatory clearance the company does not yet possess.
Midjourney licensed technology from Butterfly Network in November 2025. The prototype deploys 40 of the firm’s ultrasound-on-chip modules arranged in a ring. Each chip contains thousands of tiny transducers, some no wider than 200 microns. Future versions will pack substantially more. Butterfly Network confirmed the co-development agreement and noted it could receive up to $74 million over five years, per its press release.
Joseph DeVivo, chief executive at Butterfly, expressed pride in backing Midjourney’s mission. He highlighted the absence of radiation, low cost potential, half a million sensors, and more than two petaflops of processing power. The partnership gives Midjourney access to proven semiconductor-based ultrasound components. Still, turning those pieces into a reliable full-body tomographic system marks uncharted territory for an organization built on generative models.
The company’s own announcement struck an almost philosophical tone. “Today we’re gonna announce something a little weird and a little crazy, but also spectacular and filled with hope,” it read on the Midjourney Medical blog. The post described sensors acting like dolphins with echolocation. Waves travel through water and tissue. Changes in density and stiffness reveal organs, muscle, fat and bone. The reconstruction produces views that resemble MRI slices yet arrive far quicker.
Resolution reaches fractions of a millimeter. That matches many clinical MRI systems. But experts familiar with ultrasound computed tomography note the gap between prototype demonstrations and consistent diagnostic performance across varied body types. Midjourney has released concept videos and sample reconstructions. Real-world clinical data remains limited.
Ambitions stretch well beyond the prototype. Midjourney Medical wants more than 50,000 scanners deployed worldwide by 2031. The target capacity hits a billion scans per month. Early and frequent imaging, the company argues, could avoid 30 percent of deaths and 50 percent of health care costs. Such figures represent aspirational modeling rather than proven outcomes. They echo the optimism that propelled Midjourney’s image generator to cultural prominence.
Integration into daily life takes an unexpected form. The first Midjourney Spa opens in San Francisco near Union Square around the end of 2027. Roughly 25,000 square feet. Nine or ten scanners. Hot tubs, saunas, cold plunges. The scan becomes almost incidental. Visitors come for relaxation. Data accumulates quietly in the background. Scans start as body composition maps, a category that sidesteps immediate FDA diagnostic requirements.
Over subsequent years the plan calls for algorithm refinement, hardware generations two and three, and expanded regulatory submissions. Generation three brings custom silicon and sharper performance. The first location functions as a research spa. Real usage data informs improvements. Yet the path from spa-based wellness scans to approved diagnostic tools spans years of clinical trials and review processes.
Skepticism surfaced quickly. The Next Web titled its coverage “Midjourney’s full-body scanner: big claims, no track record.” The article pointed out Midjourney’s lack of experience in physical products or medical devices. It noted ongoing intellectual property lawsuits from Disney and Warner Bros. Discovery over the image tool. No outside investors. The organization describes itself as a community-backed research lab.
Whole-body screening in asymptomatic people carries known risks. False positives. Incidental findings that trigger cascades of additional tests, some invasive. Multiply that risk by a billion scans monthly and the medical community grows uneasy. Midjourney acknowledges the regulatory hurdles. It intends to submit test results regularly to the FDA as capabilities expand.
Butterfly’s involvement lends credibility on the hardware side. Its handheld ultrasound devices already operate in clinics and remote settings. The chips have cleared regulatory pathways before. Scaling them into a ring configuration that reconstructs full 3D volumes from tomographic data demands new computational approaches. The announcement mentions terabytes of raw data per scan, compressed and processed across clusters of computers.
Recent coverage echoes both excitement and caution. A Medical Design & Outsourcing piece from yesterday details the 40-chip prototype and Butterfly’s expectation of greater module counts ahead. It frames the spa concept as an AI-powered medical imaging destination opening in 2027. Discussions on X, formerly Twitter, range from enthusiastic predictions of transformed preventive care to questions about whether the physics truly matches MRI quality across all tissue contrasts.
One thread highlighted existing research in ultrasound computed tomography at institutions like Caltech. Midjourney did not invent the underlying concept. It aims to engineer a practical, affordable, consumer-friendly version. The water bath simplifies coupling of sound waves to the body. The slow descent at five centimeters per second allows dense sampling. The approach avoids the claustrophobia and noise of MRI tunnels.
Cost remains undisclosed. Holz declined to provide pricing when asked. Affordability sits at the center of the vision. The company speaks of megabytes of health data per second per dollar. Frequent scans. Longitudinal tracking. Conversations with physicians, nutritionists, trainers and AI systems informed by personal baselines rather than population averages.
The broader shift merits attention. An AI image company pivots to medical hardware. It partners with a public ultrasound specialist. It plans physical locations that blend wellness and data collection. Success hinges on execution across domains far removed from Discord-based prompt battles. Engineering. Clinical validation. Regulatory navigation. Operations at scale. Data privacy for intimate body scans.
Midjourney insists this fits its identity as a research lab asking fundamental questions about human experience. It wants scanners as commonplace as spa visits. It invites community input on design and priorities. Regular updates with new images and concepts will follow. The first research trials aim to showcase raw capabilities within the next year.
Whether the Midjourney Scanner becomes a routine tool or joins a list of ambitious medical concepts that never fully materialized will unfold over the coming years. Regulators will scrutinize performance. Physicians will debate clinical utility. Patients will weigh convenience against risks of over-testing. The hardware exists in prototype form. The claims stand tall. Delivery, as always, determines the outcome.
And the stakes feel real. If even a fraction of the promised impact materializes, the approach to preventive health could shift dramatically. But hype has surrounded imaging breakthroughs before. Track records matter. Midjourney now carries one in a new field. Observers will watch closely as the spa doors prepare to open.
WebProNews is an iEntry Publication