The Singularity's Chief Prophet Signs On: Ray Kurzweil Joins Nanoparticle BCI Startup Subsense

There is a certain symmetry to it. Ray Kurzweil spent five decades predicting that humans and machines would merge — that the boundary between biological and digital intelligence would dissolve somewhere around 2045. This week, the 78-year-old inventor and Google researcher took a concrete step toward that forecast from the inside: he has joined Subsense, a Palo Alto neurotechnology startup building a brain-computer interface delivered through the nose.

Kurzweil’s new title is product and vision advisor. According to the announcement, first reported by Fortune on September 3, he will advise on Subsense’s long-term technology roadmap and, in his words, on the coming convergence of artificial intelligence and neural interfaces. For a man who has spent his career at the keyboard — OCR, speech synthesis, text-to-speech, and two decades at Google — the move puts him adjacent to actual hardware that touches the brain.

A BCI you snort

What distinguishes Subsense from the crowd of brain-computer interface companies is its delivery mechanism. There is no drill, no craniotomy, no threaded electrode array implanted by a surgical robot. Instead, the company’s system is built on functional nanoparticles that are administered intranasally — in plain terms, snorted — and a wearable wireless device that talks to them from outside the skull.

The company’s technology page describes a modular architecture with two classes of particles doing complementary jobs:

• Plasmonic nanoparticles handle the “read” side. These particles scatter near-infrared light differentially depending on the local electric field, which means neural activity modulates an optical signal that external hardware can pick up. The result, the company says, is sensing with improved spatial and temporal resolution compared with scalp-mounted electroencephalography.
• Magnetoelectric nanoparticles handle the “write” side. These convert externally applied magnetic fields into localized electric effects inside the brain, making it possible to stimulate specific regions without implanting electrodes at all.

Together, the two particle classes plus the wearable form what Subsense calls a bidirectional bridge between neural activity and computation — a non-surgical, minimally invasive alternative to the implanted arrays that define the Neuralink approach.

The company’s published preclinical milestones show the program is still early but moving: first in vivo brain stimulation with magnetoelectric nanoparticles, plasmonic-surface in vitro electrophysiology, expanded behavioral in vivo studies, and calcium fluorescence imaging of primary neurons incubated with the particles. Subsense is quick to note, as regulators require, that its products are not authorized, cleared, or approved by the FDA for marketing in the United States, and that performance claims have not yet been evaluated by the agency.

Modest funding, outsized signal

Subsense is not a headline-sized check by 2026 standards. The company emerged from stealth in February 2025 with a $17 million seed round led by Golden Falcon Capital, then expanded its total financing to $27 million by December 2025 with an additional $10 million from the same investor. It was co-founded by neurotech entrepreneur Tetiana Aleksandrova and serial entrepreneur Artem Sokolov, and has collaboration ties to University of California researchers.

But the Kurzweil appointment is a signal that arguably outweighs the balance sheet. Startup advisors of his stature are rare, and his willingness to attach his name to a nanoparticle BCI — a category many neuroscientists still treat with skepticism — instantly legitimizes the non-surgical approach in circles that implants have struggled to reach. It also gives Subsense something money can’t easily buy: a direct line to the singularity narrative, and to the community of technologists who grew up reading The Singularity Is Near.

Why the non-surgical lane matters

The brain-computer interface field has spent the past decade converging on implanted electrodes. Implanted arrays offer high-fidelity, cell-level recording, and clinical programs from Neuralink, Blackrock Neurotech, and Precision Neuroscience have produced genuinely moving results — restored communication for paralyzed patients, cursor control, speech decoding. But implants face a trio of structural headwinds: surgery carries real risk, device longevity under immune response remains unresolved, and the patient population willing to undergo elective brain surgery will always be a fraction of the population that could benefit from neural interfaces.

Non-surgical approaches invert that trade. They sacrifice signal fidelity for radically lower adoption friction. If a BCI can be administered like a nasal spray and paired with a headband, the addressable use cases shift from medicine to wellness, to human-computer interaction, eventually to the consumer augmentation scenarios Kurzweil has long described. Subsense is betting that nanoparticle-mediated sensing and stimulation can hit a fidelity bar high enough to matter without ever breaking the skin — a bet the FDA has not yet weighed in on.

Kurzweil’s own writing frames why this matters now. His 2005 forecast of the singularity hinged not on any single technology but on the compounding of several — computing, genetics, nanotechnology, robotics — with brain-machine fusion as the endgame. In recent interviews he has argued that large AI models are tracking his predicted trajectory closely, and that the remaining bottleneck for merging with our tools is the bandwidth of the human-machine channel. A nanoparticle interface that reads and writes neural activity through the nasal cavity is, whatever its eventual clinical fate, one of the more literal attempts to widen that channel.

The skeptic’s checklist

The obvious caveats deserve stating plainly. Intranasal nanoparticle delivery to the brain is a real research direction — the nose-to-brain pathway bypasses the blood-brain barrier — but most peer-reviewed work remains in small-animal studies. Scatter-based optical sensing through tissue faces depth and noise constraints that implanted electrodes simply do not. And the distance between “first in vivo stimulation” and a cleared, consumer-facing product is measured in years and regulatory cycles, not press releases. The FDA disclaimer on Subsense’s own site is the honest marker of where the technology stands today.

There is also the question of what an advisor actually contributes. Kurzweil is not joining as a neuroscientist, and his mandate — long-range roadmap, AI-neural convergence — is the kind of advisory role that can be more brand than substance. The startup ecosystem is full of famous names on advisory pages.

But symbolism has a way of becoming strategy. When the most famous living prophet of human-machine merger chooses to advise the company whose product is, quite literally, nanoparticles you snort to talk to your brain, the industry reads it as a directional signal. The non-surgical BCI lane just acquired its most credible storyteller — and Subsense just acquired the closest thing Silicon Valley has to a household name in futurism.

Whether that pairing produces a product, or simply a very well-narrated research program, depends on science that hasn’t happened yet. For now, the singularity’s chief advocate has stopped predicting the merge from the sidelines. He has taken a seat at the table where it might be built.