Orbit is building a non-invasive neurostimulator that controls sleep, metabolism, and the autonomic nervous system

Summary

Stephen is a Thiel Fellow building Orbit, a non-invasive neurostimulator that sticks to the back of the ears and targets the vestibular system — the inner-ear organ that controls the perception of motion. By stimulating it precisely, the device can induce motion sensations without surgery, generate hallucinations of spinning or flying, and, more practically, make the brain think it's receiving the same signals it gets from rocking or rhythmic movement.

How it works

The vestibular system is evolutionarily ancient, which means it's entangled with deep brain regions well beyond its original function — a connectivity quirk that explains why rocking a child induces sleep. Orbit exploits that entanglement. Stephen describes the device as making "high-level API calls" to the autonomic nervous system: adjusting heart rate, breathing rate, and other functions without penetrating the skin. The electrode is a hydrogel patch worn in a headband configuration — no surgery, no implant, peelable on demand.

Sleep results

Stephen says he's recording 2 to 2.5 hours of deep (M3) sleep per night using the device, which he describes as an aberrant reading on his Whoop he has never achieved naturally. The effect sizes are large enough that he argues placebo is not a meaningful confound — when the device is on, the physiological response is immediate and unmistakable: heart rate drops, blood pressure shifts, the world appears to spin.

Metabolism and the Ozempic comparison

Orbit is running early experiments to test whether vestibular stimulation can replicate metabolic effects similar to GLP-1 drugs like Ozempic, without the side effects. Stephen is explicit that they're not there yet.

Development pace

Iteration cycles run two to three days when new hardware or software is needed, and as fast as two to three hours otherwise. The team confirmed a long-standing neuroscience theory the day before the conversation that had gone unproven for a decade — a byproduct of running controlled stimulation trials. Effect sizes across multiple neurostimulation categories are reportedly growing 1.5x to 2x per month, though Stephen declines to name the specific indications.

Regulatory path

The FDA pathway remains undecided. The indication they pursue will drive the regulatory strategy, and the team is still determining which therapeutic application is advancing fast enough to justify a filing.

The distribution argument

The non-invasive form factor is deliberate strategy, not just a technical choice. Neuralink currently has roughly 20 to 30 surgeons capable of performing its implant procedure. A peel-and-stick electrode can scale; a surgical robot cannot. Stephen's framing is that the binding constraint in neurotech isn't hardware — it's understanding how the brain works, and you need large-scale data collection to get there. Surgery is the bottleneck that prevents that data from accumulating.

Orbit's near-term thesis is that non-invasive access to the autonomic nervous system, deployed at consumer scale, generates the data needed to map and eventually replace functions — starting with sleep — that currently require hours of passive recovery.