The Med Tech Company Behind the Midjourney Scanner
Donna Cusano asked the right question about Butterfly Network. Here’s what I found when I pulled on the loose thread.
You know when a friend points out a loose thread on a piece of clothing? Yeah that!
A good friend and insightful MedTech industry analyst left a comment on my original article published last week about the Midjourney Whole-body Scanner announcement. I’d like to quote it here in full, because it gnawed away at me:
“Another problem is Midjourney’s partner, Butterfly Network. This is a company that developed an interesting and validated technology for handheld ultrasound 15 years ago that never took off as it should.
It peaked at its IPO via a SPAC in ‘21 and downhill ever since through multiple rounds of management. Can BFLY scale? They are touting this heavily and it’s bumped the stock. Yet can BFLY do anything other than chips for a prototype?”
Donna Cusano
That’s not a hostile question. It’s a precise one. From an industry observer that has seen a lot of claims and the performance that followed. And it deserved an answer.
So I went pulling on that ‘loose thread’.
First, what Butterfly actually is
Butterfly Network is a medical device company. That framing matters because most coverage of the Midjourney announcement has described Butterfly as primarily a chip manufacturer. A component supplier whose CMUT technology happens to sit inside a scanner prototype. But that’s not the accurate read.
Butterfly’s innovation is that it reimagined how a medical ultrasound device is built. Where conventional ultrasound probes rely on piezoelectric crystals, Butterfly integrated Capacitive Micromachined Ultrasonic Transducers (CMUTs) directly onto semiconductors.
Their applied result of the technology is a software-defined ultrasound probe device: dense arrays of elements, no mechanical motion required, a frequency range of 1 to 12 MHz, and a platform that can be updated in software rather than rebuilt in hardware. The semiconductor architecture is the method of innovation. The medical device is the product.
That was the promise of the original Butterfly iQ, launched in 2017. One probe, whole-body imaging, lower cost, broader access. The kind of product that makes clinical and biomedical sense the moment you hear it described.
However, pitch deck projections rarely survive first contact with the market
Medical devices don’t diffuse like consumer apps.
Clinicians need training before they trust a new modality.
Images need to reach diagnostic confidence.
Workflows need to fit departmental rhythms and billing realities.
Health systems need purchasing justification and the ability to manage a vast, sprawling estate of different devices, classes, manufacturers, and maintenance schedules.
Radiology, emergency medicine, critical care, primary care, and obstetrics each have different incentives, and none will specifically agree on who owns point-of-care ultrasound. The device was elegant. The adoption slog was not.
The company that the announcement is trying to rescue
The public company version of Butterfly Network has been a story of maturation under pressure.
It went public via a SPAC in 2021, at the peak of a market that was briefly in love with healthcare technology moonshots. Those who have read my analysis of the pattern that connects Olive AI, IBM Watson Health, Carbon Health, and Babylon Health will recognize several of the markers: ambition outpacing adoption, borrowed legitimacy from a compelling technology narrative, and a commercialization grind that the original story never accounted for
Butterfly hasn’t collapsed. But it’s had to become something different from what it announced itself as. The handheld probe business proved labor-intensive to scale. Revenue grew, but losses persisted. Management has turned over. The company narrowed its focus and began building toward what it now calls Butterfly Embedded — a strategy in which third parties license Butterfly’s imaging technology and integrate it into their own products, rather than Butterfly having to win every clinical workflow on its own.
That pivot matters. It’s a more honest fit for what Butterfly actually has: a powerful enabling imaging platform built on a unique silicon footprint, not yet a complete commercial system at scale.
The IBM Watson Health arc is also worth holding in mind here. IBM Watson had genuinely interesting technology at its core. The gap wasn’t the invention. It was the distance between what the invention could do in a demonstration and what it could deliver at scale, in real clinical environments, against real institutional resistance, over time.
Watson’s $4 billion investment ended in a $1 billion exit, not because the underlying natural language processing was fraudulent, but because the hard problems (workflow integration, physician trust, outcome evidence, competitive pressure from better-capitalized players) were never really solved.
Butterfly has not quite reached Watson-scale failure. It has something more durable in its device architecture and regulatory track record. But the warning in the IBM Watson Health story is about what happens when a healthcare technology announcement substitutes narrative momentum for the unglamorous work of commercial execution.
Butterfly has been living with that tension for the past five years. The Midjourney partnership is the most dramatic attempt to resolve it — by repositioning Butterfly not as a struggling handheld-probe company, but as the enabling platform for an entirely new imaging category.
SoftVue proves the geometry, but it doesn’t provide the moat.
Here’s where that SoftVue comparison becomes more analytically useful than it first appears.
SoftVue, developed by Delphinus Medical Technologies and FDA-cleared in 2021, uses a 22-centimeter ring transducer with 2,048 elements arranged in a full 360-degree array. The breast sits in a warm-water bath. The ring scans from nipple to chest wall in 2-millimeter increments, capturing reflection, speed of sound, and attenuation simultaneously, which Delphinus calls TriAD. The result is a volumetric reconstruction of the breast that finds up to 20 percent more cancers than mammography alone in dense-breast tissue.
To be clear SoftVue doesn’t use Butterfly’s imaging technology. The comparison isn’t specifically about the device design. SoftVue uses a purpose-built circumferential transducer array rather than Butterfly’s CMUT-on-wafer approach. The comparison is rather about system architecture.
SoftVue proves that surrounding tissue with many acoustic emitters and receivers, using water as a coupling medium, and reconstructing volumetric information from multiple angles isn’t science fiction. It’s clinically cleared and operating in hospital sites today.
That precedent makes the Midjourney/Butterfly scanner more plausible. It also makes the moat thinner.
The Midjourney scanner reportedly uses 40 Butterfly imaging modules per system, each containing approximately 8,960 transducer elements (roughly half a million acoustic sensors in total) and more than two petaflops of processing power. The density is remarkable, and Butterfly’s semiconductor-based device architecture is the key ingredient enabling it at that scale. But the ring-array concept itself, surrounding the tissue, collecting from all angles, reconstructing the volume, is already in the world.
If this modality category is validated, no incumbent imaging company will need to be informed about the geometry. GE HealthCare, Philips, Siemens Healthineers, Samsung Medison, and a fleet of established medical imaging manufacturers and semiconductor specialists can all read the same physics. What they lack today is Butterfly’s particular device architecture and Midjourney’s compute ambition. Neither of those advantages is permanent.
Butterfly’s own SEC filings have acknowledged this, noting that competitors could design around patents, independently develop alternatives, or compete in jurisdictions where protection is weaker.
That’s not a death sentence for Butterfly. It’s an honest description of where the defensibility actually sits — and where it doesn’t.
The sentence buried in Midjourney’s own roadmap
There’s one tiny little (explosive) detail in Midjourney’s own roadmap that deserves more attention than it has received so far.
Midjourney’s own announcement is explicit on this point:
“In 2028, we’ll start scaling to more cities, and we’ll upgrade to a 3rd generation scanner.
Gen3 is where it gets ‘serious,’ the silicon for this design will be completely custom, and image quality and scan times will be night-and-day.”
If you read that carefully, Butterfly isn’t the permanent foundation of this product. It’s the bridge technology for generations one and two. Midjourney has already signaled, in its own planning horizon, that it expects to move beyond what Butterfly can currently offer.
For Butterfly, this cuts both ways. On one side, it validates the embedded strategy and generates real revenue over a multi-year development arc. On the other hand, it means Butterfly’s role may be durable for the near term while remaining structurally replaceable in the medium term. A bridge partner can make money. An indispensable platform makes history. Right now, Butterfly’s position in the Midjourney story looks more like the former.
The uncomfortable asymmetry
This is where Donna’s awkward question rises.
Midjourney’s announcement gave Butterfly a chrome coating of futuristic inevitability. The whole body USCT scanner is a beautiful concept. The pitch is audacious. The association with an AI company that has dominated creative image generation for years is a narrative gift.
Butterfly’s stock moved on the news, and the framing (Butterfly’s imaging technology inside the future of medicine) is genuinely flattering.
But Midjourney can pivot.
If Butterfly’s devices do not scale cost-effectively, Midjourney has the capital and the technical ambition to find alternatives.
If the category validates sufficiently to attract a major imaging OEM, Midjourney may find a more commercially powerful partner.
If custom silicon becomes available by 2028 as planned, Butterfly’s role diminishes precisely at the moment the category is supposed to scale.
Midjourney is making a long bet, and Butterfly is its current best instrument for that bet not necessarily its permanent one.
Butterfly can’t replace what Midjourney gives it. The announcement repositioned Butterfly’s entire strategic story. Without it, Butterfly is a medical device company with a genuinely innovative imaging platform and a hard commercialization history, executing a pivot toward embedded licensing in a market that hasn’t yet been proven. With it, Butterfly is “the technology inside the future of whole-body imaging.” That narrative isn’t available from any other source.
That’s not a stable dependency on which to build a valuation.
IF I sound a bit grumpy, consider buying me a coffee :-)
Kudos earned. Inevitability, not quite yet.
None of this makes Butterfly’s technology fraudulent, or Midjourney’s plan impossible. The Demystifying Medical Digital Twin piece I published earlier this year introduced a maturity model for evaluating technology claims that I applied to Digital Twin in health IT. The lesson? Clarify the gap between what a capability appears to be and what has actually been demonstrated.
The Midjourney Scanner is sitting at the lower end of that scale right now: a generation-one prototype, approximately a dozen people scanned, a sixty-second target that currently takes twenty minutes, and a clinical evidence base that doesn’t yet exist. That’s not disqualifying. It’s the honest state of play.
Butterfly has earned respect. It hasn’t yet earned inevitability. And Midjourney’s scanner, for all the glorious ‘golden-room spa’ imagery surrounding it, still has to prove that the device at its core is the device that gets it all the way there.
Donna asked the right question.
The answer is, as yet, not yet settled.
Watch the silicon roadmap. Watch the revenue concentration. And watch what happens if this category gets big enough for the dominant global incumbents to notice and jump in.
Keep haverin’. SM.
Selected sources
Butterfly Network. “Butterfly Announces FDA Clearance of its Next-Gen Handheld Ultrasound System, Butterfly iQ3.” January 8, 2024. https://www.butterflynetwork.com/press-releases/fda-clearance-butterfly-iq3
Butterfly Network. “Ultrasound-on-Chip™ Technology.” https://www.butterflynetwork.com/technology
Napolitano, Gina. “Midjourney launches a healthcare division with ‘Ultrasonic CT,’ a whole-body ultrasound scanner built on Butterfly Network’s chip technology.” RD World Online. June 18, 2026. https://www.rdworldonline.com/ai-image-firm-midjourney-spins-up-medical-division-unveils-ultrasonic-ct/
“Midjourney Medical: Full-Body Ultrasonic Scanner, 60 Seconds, SF Spa 2027.” ExplainX.ai. June 18, 2026. https://explainx.ai/blog/midjourney-medical-full-body-scanner-spa-2026
Delphinus Medical Technologies. “Delphinus Receives FDA Approval for its SoftVue 3D Whole Breast Ultrasound Tomography System.” October 12, 2021. https://www.delphinusmt.com/news/recent-news/team-news/delphinus-receives-fda-approval-for-its-softvue-3d-whole-breast-ultrasound-tomography-system/
Butterfly Network Annual Reports / SEC 10-K risk disclosures. Available via SEC EDGAR: https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=BFLY
I further dug into the Midjourney and Butterfly deal here. Many credits to you. Butterfly announced another Embedded research partner a week later, though Aleph doesn't have a revenue path.
https://telecareaware.com/b/ArM
First of all, Stuart, thank you for taking apart my question re Butterfly Network, and putting it back together! I agree with your insight that for Midjourney, the Butterfly imaging module is a waypoint in their development, should they go all the way.
Butterfly has plenty of competition in handheld scanners and has since the beginning. Vscan by GE Healthcare and Philips Lumify from the start along with others like Exo and a few others in the POCUS dust. What was always mystifying to me was GE, Siemens, or Philips *not* buying Butterfly for the chip tech 7 or 8 years ago. Certainly before the SPAC it was doable, I believe.
Here was the other problem with all the upstart POCUS. When I was in the travel industry pre 9/11, we had a saying (meme)--if you were a company buying corporate travel management, no one ever got fired for contracting either American Express or Rosenbluth. (the former bought the latter in 2003). It made it tough for the also rans which could carve out a niche in certain types of companies, on cost, or in regions. Take the analogy to hospitals and clinics. Brand. Name. Safe. Which is why Butterfly still doesn't have real traction where the bulk of the medical device business is...hospitals. (And not veterinary, where they developed a niche.)
One last point: Butterfly is merching this like mad on social media, notably LinkedIn (paid) and X, and others are picking it up. Just an example... CK Capital @CKCapitalxx on X is hustling BFLY as an investment with big upside....
"Midjourney brings the software and the brand. $BFLY brings the actual physics of the imaging, licensed under a deal worth up to $74 million over five years.
And the scale is staggering. Midjourney is targeting 50,000 scanners and a billion scans a month by 2031. Future versions use even more modules each. Butterfly gets paid on every single unit.
Everyone is looking at Midjourney. The picks and shovels play is $BFLY"
A buildup that leads to investment or a buyout?
.