In short
Midjourney released a behind-the-scenes video of its ultrasound-based medical scanner, but it still offered little evidence the prototype can deliver the imaging performance the company suggests. The startup is positioning the device as a wellness product first, while experts remain unconvinced by the lack of public validation.
- Midjourney’s new video shows the scanner’s hardware, but not proof it works as promised.
- The company is framing the product as a wellness and body-composition device to avoid the stricter diagnostic path.
- Experts continue to question whether the system can overcome ultrasound’s known technical limits.
- Midjourney says it will publish more updates, but key performance and regulatory questions remain unanswered.
Midjourney has published a new behind-the-scenes look at its experimental medical scanner, but the nearly 20-minute video does little to answer the biggest question surrounding the project: whether the company can truly deliver the kind of fast, detailed, low-cost imaging it is promising. The footage offers a better look at the machine’s construction and the small team behind it, yet it still leaves a wide gap between Midjourney’s ambitious claims and the evidence it has shown so far.
The AI company, widely known for its image-generation tools, is trying to push into an entirely different field: medical imaging. Its prototype is an ultrasound-based scanner designed around a dunk-tank-like setup that the company says could one day be installed in spas and other consumer-facing locations. Midjourney’s pitch is straightforward and unusually bold for a startup best known for generating fantasy art: use existing ultrasound technology in a new form factor, make it inexpensive enough to be widely available, and produce body scans detailed enough to be medically meaningful over time.
But the new video, released by engineer and YouTuber Marcin Plaza, provides more of a workshop tour than a scientific demonstration. It shows the hardware, the assembly process, and the scrappy engineering approach behind the prototype. What it does not show is the kind of rigorous validation that would persuade medical imaging specialists that the system can overcome ultrasound’s long-established limitations.
What Midjourney showed in the new video
The most striking thing about the video is how physical and improvised the system appears. Plaza, who also works for Midjourney, describes the scanner in blunt, hands-on terms. In his telling, it is built from a large number of ultrasound probes, custom electronics, and commodity computing gear arranged in a structure that resembles a high-tech soaking tank more than a clinical machine.
According to Plaza’s description, the prototype is essentially a collection of probes fitted into a system with an elevator-like mechanism, supported by standard computers and Raspberry Pi boards. The presentation makes clear that this is still an engineering experiment, not a polished commercial product. The company appears to be emphasizing iteration and speed rather than a finished design.
The video also makes an important strategic point: Midjourney is not presenting the device as a diagnostic medical scanner, at least not yet. Instead, it is positioning the machine first as a wellness and body-composition product. That distinction matters because diagnostic equipment faces a far more demanding regulatory path, including clinical validation and clearance from the U.S. Food and Drug Administration.
Why the project has drawn skepticism
From the outset, Midjourney’s entry into medical imaging has raised eyebrows in both AI and healthcare circles. Ultrasound is an old and well-understood imaging technique, but it comes with major technical trade-offs. Its performance can be affected by body type, tissue composition, and the difficulty of producing clean, detailed images quickly across a large area of the body.
Experts who spoke to The Verge when the project was first unveiled said the company had not yet shown enough to support the scale of its claims. The central concern was not whether ultrasound works — it obviously does — but whether Midjourney could repackage it into something dramatically better, faster, and more broadly useful without solving long-standing physical constraints.
Those concerns remain after the new video. The footage reveals more about how the scanner is built, but not much more about how well it performs. There is still little public evidence on resolution, scan consistency, scan time, repeatability, patient comfort, or the ability to distinguish clinically useful details from noise.
The limits of ultrasound are still the real test
Ultrasound imaging has been used in medicine for decades because it is portable, relatively inexpensive, and avoids radiation. But it is not a magic window into the body. Image quality can vary significantly depending on the anatomy being scanned, the skill of the operator, and the quality of the equipment.
That means any company claiming to dramatically improve ultrasound has to show measurable gains. It is not enough to build a more elaborate machine or to arrange many probes around a person. The hard part is proving that the system can produce reliable, interpretable images in settings that matter clinically.
So far, Midjourney has not made that case in public. Instead, the company has focused on concept, hardware development, and a broad vision for frequent scans that track health over time.
Midjourney’s wellness-first strategy
One of the clearest themes in the new video is Midjourney’s decision to frame the scanner as a wellness product first. That approach gives the company an easier path to launch, at least in theory, because wellness tools face fewer regulatory hurdles than devices marketed for diagnosis.
Tom Calloway, who leads the medical effort at Midjourney, says focusing on body composition would allow the company to move quickly once testing is complete. In his words, that strategy would let the team “speedrun” the process and open the product sooner rather than later.
The language is revealing. It suggests that Midjourney knows a full medical-device pathway would be slower, costlier, and more burdensome. It also implies that the company is trying to carve out an intermediate category: a product close enough to medicine to sound transformative, but framed loosely enough to avoid the strictest requirements.
Calloway argued that the body-composition use case should allow the company to move quickly and begin operating once testing is finished, rather than waiting for a lengthy diagnostic approval process.
Still, the video repeatedly returns to medical implications. The company asks what doctors might do if they could access scans on a frequent basis and compare them over time. That framing keeps the project tied to clinical ambition, even as Midjourney insists its first product will not be sold as a medical diagnostic tool.
Body composition or medical imaging?
There is a meaningful difference between a wellness scanner that estimates body composition and a device that can detect disease. The former may only need to provide broad guidance about measurements or changes over time. The latter has to support clinical decisions, which raises the bar dramatically.
Midjourney appears to be using that distinction to its advantage. By starting with body composition, the company may be hoping to reach market earlier, gather usage data, and build a base of customers before attempting anything closer to full medical imaging.
That strategy is not unusual in consumer health technology. Many products begin with wellness claims and later broaden their ambitions. What makes Midjourney’s case unusual is the scale of the leap: a company famous for visual AI is now trying to enter one of the most regulated and technically demanding corners of healthcare.
The company’s public messaging remains deliberately loose
Despite the new video, Midjourney has done relatively little to dispel confusion around the scanner. Calloway’s comments suggest the company does not see a need to over-explain its plans. He said there was not much to clarify and indicated that the team will publish regular blog posts showing progress.
That approach may reassure supporters who want to see a startup moving quickly and sharing work in public. But it also leaves open many practical questions. What exact imaging outcomes is Midjourney targeting? How does it compare with standard ultrasound systems? What is the intended price point? How will it handle false positives, poor scans, or inconsistent results?
Those details matter because a product in healthcare can fail not only by being technically weak but also by being difficult to trust, difficult to interpret, or difficult to regulate. The more Midjourney leans on sweeping language about medical transformation, the more it will be judged on evidence rather than aesthetics.
Midjourney’s leaders say the company’s independence gives it room to pursue the scanner on its own terms, without outside investors forcing the project into a traditional startup timeline.
That freedom may be an advantage. It may also be a reason for caution. Without external funding pressure, Midjourney can experiment longer — but it can also afford to present vision before validation.
Why this project matters beyond one company
Midjourney’s move is notable because it reflects a broader pattern in AI: companies that built reputations in one domain are increasingly trying to expand into adjacent or entirely unrelated industries. The logic is clear. If a company has a strong brand, engineering talent, and a taste for ambitious storytelling, it may be able to sell investors, users, and the public on a leap into a new market.
Healthcare is especially attractive because it promises huge impact, strong recurring demand, and potentially large revenues. But it is also one of the hardest sectors to enter responsibly. Regulators expect evidence. Clinicians expect performance. Patients expect safety. In this environment, product demos are not enough.
That is why the Midjourney scanner is such an interesting test case. It sits at the intersection of AI hype, consumer wellness, and medical regulation. If it succeeds, it could become a model for how AI companies move into hardware-backed health products. If it fails to prove itself, it may become another example of a flashy concept that outran the science.
Key questions still unanswered
The latest video adds atmosphere, not certainty. The prototype looks real enough, and the team appears committed enough, but serious questions remain about whether the project can deliver useful results outside a controlled demo.
- Can the scanner produce consistent, high-quality images across different body types?
- How fast can it scan, and how much detail can it preserve?
- What measurable advantage does it offer over existing ultrasound systems?
- Can the company validate the system without turning it into a regulated medical device?
- Will consumers understand the difference between wellness scanning and diagnosis?
These are not minor details. They are the difference between a curious prototype and a product that could reasonably claim to change healthcare.
How the project is likely to be judged next
For now, Midjourney’s scanner should be viewed as an experiment with an unusually large ambition attached to it. The company has shown more of the machine, but not more proof that the underlying concept is ready for the market it imagines.
The next stage will probably be defined less by marketing and more by evidence. If Midjourney continues to publish technical updates, the public will likely look for comparisons, benchmarks, and demonstrations that go beyond hardware visuals. Independent experts will want to know whether the scanner can actually do something distinctive, not just whether it looks futuristic.
If the company can answer those questions, the project may gain credibility. If it cannot, then the scanner will remain what it looks like today: an ambitious idea wrapped in an appealing machine, still waiting for proof.
| Aspect | What Midjourney has shown | What remains unclear |
|---|---|---|
| Product type | Ultrasound-based body scanner | Exact performance goals and clinical usefulness |
| Current positioning | Wellness and body composition product | Whether it can later support diagnostic use |
| Hardware design | Many probes, custom assembly, commodity computing | How much this improves imaging quality |
| Regulatory path | Aims to avoid early diagnostic claims | How it will navigate FDA-related scrutiny if ambitions expand |
| Public evidence | Behind-the-scenes video and team commentary | Benchmarks, clinical data, and independent validation |
The bigger story: AI companies moving into physical products
Midjourney’s scanner also reflects a larger shift in the AI industry. After years of attention focused on software, chatbots, and image generation, some of the most prominent AI companies are now looking for ways to embed their technology in physical systems. That includes consumer devices, robots, specialized hardware, and health tools.
The appeal is obvious. Physical products can be more defensible than software alone, and they can create new categories rather than competing directly in crowded app markets. But hardware raises the stakes. It requires manufacturing, maintenance, safety testing, service infrastructure, and in healthcare, a far more exacting standard of proof.
Midjourney’s scanner is therefore more than a side project. It is a signal that the company wants to be seen as more than an image generator. Whether it can succeed in that ambition will depend on whether it can turn an eye-catching machine into a dependable one.
What to watch next
If Midjourney continues to develop the scanner publicly, the most important updates will not be aesthetic. The meaningful milestones will be technical and regulatory. Watch for:
- Independent demonstrations of scan quality
- Comparisons with conventional ultrasound systems
- Clearer explanations of body-composition metrics
- Details on safety, privacy, and data handling
- Evidence that the product can scale beyond a prototype
Until then, the company has revealed a machine, a team, and a vision — but not yet a convincing case. The video may have shown more of the scanner’s guts, but it still leaves the central mystery intact: whether Midjourney is building the future of medical imaging or simply a very expensive proof of concept.
