Subsense raised $27M for a non-surgical brain interface, with Kurzweil advising

If it works, it removes the constraint that has kept brain-computer interfaces in a small clinical population. Neuralink, Precision Neuroscience and Synchron all require a surgical procedure — implantation, or in Synchron's case delivery through a blood vessel. Surgery restricts the addressable population to patients whose condition justifies the risk, which is why the field's progress is measured in individual patients rather than deployments.

A non-surgical interface would not be a better version of that market. It would be a different one.

The reasons for scepticism are specific

Each element of the approach is individually difficult, and the combination has not been demonstrated in humans.

Nasal delivery to the brain is a real and studied route, and it is also inconsistent — distribution is hard to target and varies between individuals. Magnetic readout through the skull faces the physics that has always limited non-invasive neural recording: signal falls off sharply with distance and the skull scatters it, which is why non-invasive methods have poor spatial resolution while implants have good resolution and surgical risk.

Mouse results do not resolve either problem. A mouse skull is thin and a mouse brain is small, which flatters exactly the two variables that make the human case hard.

And the safety question for anything involving persistent nanoparticles in neural tissue is not a formality. Clearance, long-term accumulation and immune response would need answering before a first-in-human trial, and that is a multi-year regulatory path.

The adviser is a signal about the audience

Ray Kurzweil's involvement will draw attention and is not evidence about the technology. He is a futurist associated with long-horizon predictions about human-machine convergence, not a neuroscientist, and his name on a preclinical BCI company tells you who the story is being told to.

That matters because $27 million is a modest seed for a company with a decade of regulatory work ahead of it. This is early money buying a lottery ticket on a physically plausible but unproven mechanism.

Subsense has not published peer-reviewed results, named a clinical timeline, or disclosed what resolution or bandwidth the mouse work achieved. Until any of those exist, the honest description is a preclinical bet — which is a legitimate thing to fund, and not a thing to describe as an interface.