Midjourney, the AI image-generation company, announced on 18 June 2026 that it is entering the medical hardware space with Midjourney Medical — a venture built around a proprietary full-body ultrasonic CT scanner and an accompanying consumer wellness venue called the Midjourney Spa.

The scanner itself is the technical centrepiece. Unlike MRI, which relies on strong magnetic fields, or conventional CT, which uses ionising radiation, the Midjourney Scanner works via water immersion and phased ultrasound arrays. A full-body scan completes in 30 to 60 seconds, requires no sedation, and delivers no radiation dose to the patient. The Verge reports that David Holz plans to deploy 10 units at a Midjourney Spa location in San Francisco's Union Square, with that facility slated to open before the end of 2027.

Holz described the scanner — cited by Latent Space — as "the first new whole-body medical imaging modality in 50 years." That claim frames the device against the introduction of clinical MRI in the late 1970s, which is a reasonable historical anchor even if the assertion will attract scrutiny from radiologists and imaging physicists who may point to intervening advances in PET/CT, spectroscopic MRI, and photon-counting CT.

What the Technology Actually Does

The Midjourney Spa experience — branded as the "whole-body scan" — is designed to complete in approximately 60 seconds. The water-immersion approach is not novel in isolation: immersion ultrasound has been explored in research settings for breast imaging and limb scanning for decades. What Midjourney appears to be claiming is a system capable of volumetric whole-body coverage at that speed, which, if validated, would be a materially different engineering achievement from existing immersion probes.

The specifics of transducer count, operating frequency, reconstruction pipeline, and the role of AI inference in image formation have not been disclosed in the current announcement. For a product making whole-body diagnostic claims, those parameters matter considerably. Ultrasound's well-known limitations — poor penetration through bone and gas-filled structures like the lungs and bowel — make whole-body coverage a genuinely hard physics problem, not simply an engineering packaging challenge.

A Generation 2 scanner is reportedly planned by the end of 2026, suggesting the current device is already in a late enough development stage that Holz's team is iterating on it publicly.

The Business Model and Regulatory Unknowns

The Midjourney Spa framing positions this as a consumer wellness product first. That classification matters enormously from a regulatory standpoint. In the United States, a device making diagnostic medical claims requires FDA 510(k) clearance or premarket approval, depending on how the intended use is characterised. Consumer "wellness" scans occupy a greyer zone — one that companies like Prenuvo have navigated with whole-body MRI offerings, typically by marketing to self-pay customers outside the standard clinical referral pathway.

Whether Midjourney Medical intends to pursue clinical clearance, wellness positioning, or both is not stated in the current announcement. That gap is the most consequential open question the company will need to answer before the Union Square location can operate, and it will likely shape how the medical community receives the product.

The pivot from generative image software to physical medical hardware is a wide lateral move, though Holz has previously signalled broader ambitions for Midjourney beyond the image-generation product. The underlying computational expertise — large-scale neural inference, image synthesis, reconstruction pipelines — does translate in principle to the AI-assisted image formation that modern ultrasound systems increasingly depend on. Whether the organisational and regulatory muscle required for a medical device business is in place is a separate question.

Looking at what this means for the broader medical imaging market: the incumbent players — Siemens Healthineers, GE HealthCare, Philips, Canon Medical — have built defensible positions on decades of clinical validation, installed-base relationships, and reimbursement infrastructure. A faster, cheaper, radiation-free whole-body modality has genuine clinical appeal if the image quality holds up. Radiologists will want peer-reviewed phantom studies and, more critically, prospective clinical data before drawing diagnostic conclusions from the output. Holz's team should expect that bar to be set early and publicly.

In this author's view, the more interesting story here may not be the hardware itself but whether Midjourney can compress the typical decade-long medical device validation cycle using AI-accelerated image reconstruction and large-scale real-world data collection from its spa network — and whether regulators will accept that approach. The precedent for that kind of compressed timeline, outside of the pandemic emergency-use context, is thin.