Midjourney just announced a full-body ultrasound machine.

That’s not a sentence I had on my bingo card for 2026.

This is a company best known for generating images from text prompts. No obvious medical device heritage. No established healthcare distribution channels. No decades of clinical research sitting behind it.

On the surface, it feels like a bizarre pivot.

But I think it points to something bigger happening in the AI market.

Over the past few years, we’ve watched companies pour enormous amounts of capital into foundation models. The assumption was that the model itself would be the moat.

I’m becoming increasingly convinced that isn’t true.

Model performance continues to improve, but the gaps between leading players are narrowing. Today’s breakthrough becomes tomorrow’s commodity surprisingly quickly. Open-source models keep getting better (well, 'good enough'). New entrants keep appearing.

If that’s the future, then owning a model may not be enough.

You need to own something else.

For some companies, that means owning a vertical. For others, it means owning hardware. Ideally both.

That’s what makes the Midjourney announcement interesting.

Medical imaging is fundamentally a visual data problem. Midjourney has spent years building expertise around visual interpretation and generation. Whether that expertise translates into clinical outcomes remains to be seen, but the strategic logic isn’t as crazy as it first sounds.

More importantly, they’re not just talking about an AI model.

They’re talking about a scanner.

Most AI companies would try to license software into existing devices. Midjourney appears to be trying to own the entire stack - the hardware, the AI, and potentially the diagnostic workflow that sits on top.

That’s a much harder game to play.

It’s also potentially a far more defensible one.

Will this particular product succeed? I have no idea. Regulatory approval, clinical validation, physician trust and commercial adoption are all massive hurdles.

But I suspect we’ll see a lot more moves like this.

As the market settles and a small number of companies emerge as leaders in general-purpose AI, everyone else will be forced to answer the same question:

Where is the moat?

For many, the answer won’t be a better model.

It will be a specific industry, a specific workflow, or a piece of hardware that competitors can’t easily replicate.

Midjourney may simply be one of the first companies willing to make that leap publicly.

#AI #ArtificialIntelligence #Strategy #HealthTech #Innovation