Butterfly Network, Larry Robbins and a EUR 11.0m sell cluster

Butterfly Network’s stock has not been trading in a vacuum. The ultrasound and point-of-care imaging trade is still being pulled by the same forces that have helped larger med-tech names, namely portable diagnostics, AI-assisted workflow and the push to move imaging closer to the bedside. GE HealthCare and Koninklijke Philips have the scale, the installed base and the balance sheets. Butterfly Network Butterfly Network, Inc. has the smaller, more fragile version of that pitch, and the market keeps asking whether the growth is real enough to justify the volatility.

Against that backdrop, Larry Robbins’ Larry Robbins funds sold again. On July 31 they sold 1,720,129 shares at a weighted average price of $7.3925, then on August 3 they sold 156,163 shares at $8.1048. The filing value on the August 3 sale was EUR 11,040,077, euro-normalised at ingest. That is not a rounding error. It is a meaningful exit from a name that had just posted a better quarter.

The ultrasound trade is still about scale, not slogans

The sector backdrop matters because Butterfly does not get judged like a sleepy device supplier. Ultrasound is still a growth market, helped by non-invasive diagnostics and the shift toward portable systems. The broader ultrasound devices market is projected to grow at an 8.23% compound annual rate from 2026 to 2035, reaching about $29.47 billion, according to Precedence Research. That is the kind of backdrop that keeps capital interested in the category even when individual names wobble.

But the category is not one trade. GE HealthCare and Philips sit at the other end of the spectrum from Butterfly. They sell breadth, service and enterprise relationships. They also talk about AI integration and remote workflow because that is where the market is going. Butterfly sells a handheld system and a software story, which is a cleaner growth narrative and a harder operating model. You get more torque when the numbers work. You also get more punishment when they do not.

Broader market conditions have not exactly made life easier for smaller healthcare growth names. The Federal Reserve held the federal funds rate steady at 3.50% to 3.75% on July 29, and some analysts are now pricing in the possibility of rate increases later in 2026 as growth and inflation stay resilient. Higher-for-longer rates tend to keep pressure on names that still need the market to believe in future margin expansion. Butterfly is one of those names. It does not trade like a utility. It trades like a story that still has to earn its multiple.

The stock has also had to live with a mixed healthcare tape. Technology and AI have carried a lot of the market’s enthusiasm, while healthcare has been more uneven, with rotation and yields doing their usual damage to sentiment. That matters because Butterfly is not being valued only on revenue growth. It is being valued on whether the company can turn a good product into a durable business. Those are different tests.

What the peers are telling you about this corner of med-tech

GE HealthCare and Philips are useful comparables precisely because they are not close operational twins. They are bigger, slower and better capitalised. They also have the kind of product breadth that lets them absorb a bad quarter in one line and make it up elsewhere. Butterfly does not have that luxury. When a smaller imaging name posts a strong quarter, the market wants to know whether it is a one-off licensing bump or the start of something more durable.

Butterfly’s recent top-line growth has been helped by licensing partnerships, according to the grounded research. That is a real business lever, but it is not the same thing as broad-based demand across a large installed base. Licensing can lift revenue quickly. It can also flatten out once the easy comparisons are gone. That is why the stock can rally on a good print and still leave room for doubt. The market knows the difference between a cleaner quarter and a cleaner model.

Analyst targets reflect that split. Recent reports have put the average 12-month price target somewhere between $5.56 and $9.25. That is a wide band for a company of this size, and it tells you the Street is still arguing with itself about how much of the growth is durable. The stock can move sharply inside that range without resolving the argument. It often does.

Insider activity becomes more interesting in that kind of setup because it gives you a view into how the people with the longest exposure are behaving around the quarter. Not a verdict. A view. And here the view is not subtle. The selling did not come from a single holder taking a token profit. It came as a cluster, and it came after the company had already shown better revenue momentum.

The filing is not the story, the pattern is

InsiderTrades data marks this as a cluster. There are three distinct insiders in the recent declarations, with 12 recent declarations in the pattern the dossier flags. The most visible name is Larry Robbins, a director, and the sales tied to his funds are the largest single piece of the activity. The dossier also shows sales by founder Jonathan Rothberg and by executives including CEO Joseph DeVivo. That matters because a lone sale can be noise. A cluster is harder to ignore, especially when it spans more than one role.

The score attached to the filing is 53. That is not a magic number and it is not the point of the piece, but it does tell you the filing sits in a middle zone rather than the extreme end of the screen. The drivers are straightforward enough: an operating director filed, the activity came as part of a cluster, and the size was about 0.67% of the company’s market value. The euro-normalised filing value near EUR 11.0 million is the kind of size that deserves a second look, not a reflexive conclusion.

What makes this more interesting is timing. The sales landed shortly after Butterfly reported record second-quarter 2026 revenue of $32.61 million, up 39% year over year, with narrowed losses. If you wanted a clean narrative, you would prefer insiders buying into that print. They did not. They sold into it. That does not tell you the business is broken. It tells you the quarter was not enough to stop a meaningful amount of distribution.

The market usually gives insiders some benefit of the doubt when they sell after a strong run or around a tax event or a prearranged plan. Here, the pattern is broader than one date and one person. That is why the cluster matters. It does not prove anything about the next quarter. It does tell you that the people filing these forms were willing to reduce exposure after a better-than-expected revenue report, and the market should not pretend that is irrelevant.

What our cohort data says about this kind of trade

The relevant historical bucket in the dossier is director-level buys at mid-cap names, which is not the same thing as this sale cluster. That distinction matters, and it is why you should not force the cohort read into a shape it does not have. The historical T+90 cohort return for that bucket is 5.58%, with a 53.1% win rate over 90 days and a 64.75% average return over 365 days. Those are historical cohort data, not a forecast, and they belong to a different role-and-size bucket than the selling we are looking at here.

That mismatch is useful in its own way. It reminds you that insider data is not a single universal language. Role matters. Size matters. Direction matters. A director buying into a mid-cap name has a different historical profile from a director-led selling cluster after a revenue beat. You can use the cohort framework to keep yourself honest, but you cannot flatten every filing into the same template and expect the result to be useful.

The strategy framework in the dossier is also worth a brief mention, because it tells you how the system is meant to be used, not how to worship it. The holding period is 90 days, the max position is 0.08, and the out-of-sample headline is 0.81, 26.4 and 51.5 on the restricted EU venue universe, with the usual caveat that those figures live in a short, single-regime window and do not survive search-aware deflation. That is a screen, not an alpha claim. You do not buy the token. You use it to keep your process from drifting into story time.

Butterfly’s quarter was better, which is why the sales matter

Butterfly’s second-quarter 2026 revenue of $32.61 million, up 39% year over year, is the kind of print that can change the tone around a small-cap med-tech name. The company also narrowed losses. That combination usually buys management some breathing room and gives the market a reason to revisit the growth case. It is exactly the sort of moment when insider selling gets more attention, because the easy excuse, that insiders are just reacting to a bad quarter, is not available.

The company’s business model still has to prove that licensing can sit alongside product demand and not just temporarily mask it. That is the real question under the quarter. If the top line is being helped by partnerships, the market wants to know whether the underlying device business is also improving. If it is, the stock can keep working. If it is not, the quarter becomes a nice headline and not much else.

InsiderTrades data gives the company a fundamental score of 27, with a rank of 23,594 out of 27,967. The quality pillar is 25 and the value pillar is 30, while growth is not provided in the dossier. I would not turn that into a thesis by itself. I would treat it as a reminder that the market is not paying for perfection here. It is paying for a path. The path still has to be walked.

That is where the insider cluster becomes more than a filing footnote. When a company is still in the proving stage, insider sales after a strong quarter can be read as simple diversification, or as a sign that the people closest to the story are not eager to add exposure at the current price. You do not need to invent motive to see the tension. The forms already show it.

The stock still has to answer the same question

Butterfly Network is a small company in a market that rewards scale. That is the whole problem and the whole opportunity. The handheld ultrasound pitch is attractive because it is portable, software-heavy and tied to a real shift in care delivery. It is also exposed to the usual small-cap hazards, including execution risk, financing sensitivity and the market’s impatience when growth does not translate into cleaner economics fast enough.

The recent insider sales do not erase the quarter. They do not cancel the sector backdrop either. What they do is sharpen the question around the stock. If the company can keep growing revenue, keep narrowing losses and show that the licensing lift is not a one-quarter trick, then the market has something to work with. If not, the selling will look less like routine portfolio management and more like a timely reduction in exposure.

For now, the most concrete thing to watch is the next filing and the next revenue print. The cluster is already on the record, the quarter is already on the record, and the stock has already been asked to absorb both. The next update will tell you whether insiders were simply taking money off the table after a better quarter, or whether they were leaning into a pattern the market has not yet priced properly.

The next filing will matter more than the commentary

The useful part of this setup is that it is still testable. Butterfly Network has already shown it can post faster revenue growth. It has also shown that insiders are willing to sell into that strength. Those two facts can coexist for a while. Eventually the market decides which one matters more.

If you are looking for the cleanest near-term marker, it is not a slogan about disruption or a generic promise about med-tech adoption. It is whether the company can keep the revenue line moving while the insider forms stop clustering on the sell side. That is the next concrete checkpoint, and it will come from the next quarter and the next round of filings, not from the commentary around them.

This is not investment advice.