Midjourney Moves Into Medical Imaging With Fast Ultrasound Scanner

Midjourney, the AI company known for image generation, announced on 18 June 2026 that it is launching a medical hardware business called Midjourney Medical. The venture centers on a full-body ultrasound scanner and a consumer wellness venue called the Midjourney Spa.

The scanner is the technical centerpiece. Unlike MRI machines, which use powerful magnets, or standard CT scanners, which rely on radiation, the Midjourney Scanner uses water immersion and phased ultrasound arrays — think of multiple ultrasound beams working together rather than a single probe. A full-body scan takes 30 to 60 seconds, requires no sedation, and exposes the patient to no radiation. The Verge reports that founder David Holz plans to install 10 units at a Midjourney Spa in San Francisco's Union Square, with that location scheduled to open before the end of 2027.

Holz called the scanner, according to Latent Space, "the first new whole-body medical imaging modality in 50 years." That frames the device against the arrival of clinical MRI in the late 1970s, which is a fair historical reference even though radiologists and imaging specialists will likely question the claim, given advances in PET/CT, advanced MRI variants, and newer CT technologies that have emerged since then.

How the Technology Works

The Midjourney Spa experience is meant to deliver a full-body scan in roughly 60 seconds. Water-immersion ultrasound itself is not new — researchers have tested it for breast imaging and limb scans for decades. What Midjourney appears to be claiming is a system that can scan an entire body quickly, which would be a genuine engineering step forward if the image quality holds up.

The company has not disclosed key technical details: how many ultrasound transducers the system uses, what frequencies it operates at, how it reconstructs images, or what role AI plays in turning raw data into the final picture. For any product making claims about diagnostic imaging, those details matter greatly. Ultrasound has long-standing limitations — sound doesn't penetrate bone well, and it struggles with air-filled structures like the lungs — so covering a whole body is a genuinely difficult physics problem.

A second-generation scanner is reportedly planned by the end of 2026, which suggests the current device is far enough along in development that the team is already working on improvements.

Regulatory and Business Questions

The Midjourney Spa brand positions this as a consumer wellness product. That choice carries serious regulatory consequences. In the United States, a device that makes claims about medical diagnosis requires FDA clearance, which is a formal approval process. Wellness scans sit in a grayer regulatory zone — companies like Prenuvo have marketed whole-body MRI scans this way, typically selling directly to patients outside standard clinical pathways.

Midjourney has not said whether it will pursue formal FDA clearance, stick with wellness positioning, or attempt both. That uncertainty is the most pressing question the company needs to answer before the San Francisco location can operate, and it will influence how the medical community views the product.

The move from generative image software to physical medical hardware is a substantial shift, though Holz has indicated Midjourney has bigger ambitions beyond image generation. The computational skills that power Midjourney — running large neural networks efficiently, synthesizing images, reconstructing complex pictures from raw data — do carry over in principle to the AI-assisted image processing that modern ultrasound systems increasingly rely on. Whether Midjourney has the regulatory expertise and organizational muscle to run a medical device business is a different question.

The broader medical imaging market is dominated by companies like Siemens Healthineers, GE HealthCare, Philips, and Canon Medical. They have built strong positions through decades of clinical testing, deep relationships with hospitals, and established insurance reimbursement systems. A faster, cheaper, radiation-free full-body scan would have genuine appeal in clinical settings if the image quality is good. Radiologists will want to see peer-reviewed studies on phantom materials first, then real patient data, before they rely on these images for diagnosis. Midjourney should expect that scrutiny from the start.

The more intriguing question may be whether Midjourney can compress the typical decade-long process of validating a medical device by using AI to improve image reconstruction and by collecting large amounts of real-world data from its spa network. Outside of pandemic emergency situations, we have rarely seen that kind of accelerated timeline in medical device approval. Whether regulators will accept that approach remains an open question, and the answer could set a precedent for how other AI-driven medical hardware companies approach validation.

In this author's view, if Midjourney pulls this off, it would be a notable example of how AI expertise developed in one domain can reshape an established industry — though the company is pursuing a path with few historical precedents, and the regulatory bar will be neither quick nor forgiving.