SonoThera

Here is a fact about gene therapy that the field would prefer you not dwell on: we are pretty good at knowing which gene to fix, and pretty bad at getting the fix to the right place. The gene is the easy part. Delivery is the wall everyone runs into. And the standard tool for getting over that wall - a hollowed-out virus loaded with genetic cargo - comes with a set of problems that are, on close inspection, kind of embarrassing for a cutting-edge technology.

Viruses are expensive to manufacture. Your immune system remembers them, which means you usually get exactly one dose before your body starts treating the therapy like an infection. And they don't go everywhere - some organs are simply hard for a virus to reach. So you have this elegant molecular medicine bolted to a delivery mechanism that is finicky, one-shot, and geographically limited.

SonoThera, a roughly 50-person company in South San Francisco, looked at that arrangement and decided the problem was the virus, not the gene. Its whole pitch is a different delivery system: focused ultrasound plus injectable microbubbles. You inject the bubbles, aim the ultrasound at the organ you care about, and the acoustic energy makes the bubbles oscillate and briefly punch transient pores in nearby cell membranes. The genetic payload slips through the pores. The pores close. The bubbles are gone. No virus involved.

The technical name for this is sonoporation, which sounds like a spa treatment and is actually a biophysical process. The elegant part - the part that makes an investor lean forward - is what it implies. If there's no virus, there may be no immune memory. If there's no immune memory, you might be able to dose the therapy again. And again. For a chronic genetic disease, "redosable" is not a feature. It's a different category of medicine.

There is a reason to take this seriously beyond the pitch deck, which is that the underlying tool is not speculative. Ultrasound microbubbles are already an FDA-approved, boring, widely-used product. Cardiologists inject them roughly 35 million times a year in the United States just to get a clearer picture of the heart. SonoThera's co-founder, Steve Feinstein, is the person who invented the first two FDA-approved versions of them. The company's insight, reduced to a sentence, is: what if the bubbles carried medicine instead of just improving the image?

This is the sort of idea that is either obvious in hindsight or quietly very hard, and SonoThera has spent since 2022 finding out which. So far the answer looks like "hard, but working." In June 2026 it closed an oversubscribed $125 million Series B, which is the market's way of saying the animal data was convincing enough that a lot of sophisticated people wanted in.

The Janssen collaboration, signed early on, is the tell. A pharmaceutical company paying a startup to help deliver its own next-generation DNA and RNA therapeutics to non-liver organs is a company outsourcing the exact problem it can't crack internally. When incumbents pay the new kid to solve their hardest thing, that's usually more meaningful than the size of any single funding round.

None of this proves the therapy works in humans - that's what the 2027 Duchenne trial is for, and the gap between convincing mouse data and a first patient is where a lot of biotech stories get honest. But the capital is patient, the investor list is strategic, and the team owns the underlying tool. It is, at minimum, the right table to be sitting at.