Science Corporation implants its PRIMA chip in a Stargardt patient

A chip measuring two millimetres by two millimetres, placed at the back of the eye, and two rehabilitation sessions. In the first, the patient made out a letter; in the second, they read a word. That is the first result Science Corporation shared at TechCrunch Disrupt after implanting its PRIMA device in a person with Stargardt disease, an inherited retinal condition that currently has no cure.

Max Hodak, the company's CEO and previously president and co-founder of Neuralink, explained it on Tuesday on stage at the event in San Francisco. The headline is medical, but what Hodak argued for is a broader thesis about what counts as a brain computer interface.

What happened

The surgery took place in May at Sydney Eye Hospital (Australia), as part of an ongoing trial. PRIMA works as a substitute for the photoreceptors the disease has destroyed: the patient wears camera-equipped glasses that capture the image and send it to the implant, which acts as an artificial photoreceptor.

Stargardt disease affects around 150,000 people worldwide, according to TechCrunch. It causes a central blind spot that grows over the years and mainly damages the outer retina, while the inner retina can be preserved. That difference is what makes the approach viable: if the neurons that carry the signal towards the optic nerve still work, giving them an artificial input is enough.

Hodak was clear about the limits of the current device. The vision it provides is black and white, with a very narrow field. "It's like looking through a straw," he said, while adding that the image is sharp and high contrast. In practice, the patient sees a couple of letters at a time.

Why it matters

PRIMA is not a lab prototype. The device is already commercially available in Europe for patients with geographic atrophy, the advanced form of age-related macular degeneration. What is new is testing whether the same hardware works for a different disease, one that is genetic and usually appears in young people.

Hodak used the case to frame the company's strategy: "The retina is part of the central nervous system. We consider cochlear implants, retinal prostheses, motor prostheses, these are all different types of BCIs." It is a way of redefining the field with practical consequences. Instead of promising that one day we will control a computer with our thoughts, Science Corporation leans on medical applications with a specific problem, a specific patient and a measurable metric, such as reading a letter.

The roadmap he gave at Disrupt has two horizons. A next generation of the implant should reach human trials within the next year, and within five years Hodak believes they can get close to native visual acuity. The company also mentioned other lines of work: a neurodegenerative disease it has not yet disclosed, stroke recovery and brain tumour monitoring.

Who it is relevant for

Patients and ophthalmologists: it is the first sign that a prosthesis already sold in Europe for one indication can extend to another with no available treatment. A single case is not clinical evidence, but it does justify following the trial closely.

Anyone following the BCI sector: the comparison with Neuralink is unavoidable. Hodak founded Science in 2021 after leaving that company, and his bet is to start with the eye, an entry point into the nervous system that is more accessible than the motor cortex.

Teams working on AI for healthcare: the value of these systems depends as much on the hardware as on processing the signal coming from the camera. The more ambitious the acuity goal, the more weight falls on the software that decides what information fits through such a narrow channel.

What we don't know yet

The TechCrunch piece does not say how many Stargardt patients will take part in the trial or when aggregated data will be available. Nor is there a timeline for this indication to reach the market. Reading a word in the second session is a striking result, but the relevant question is how much the patient's day-to-day independence improves after months of use.

At ClaudeWave we follow projects like this because they mark the ground where technology stops being a text interface and starts working with signals from the body itself. We find Science's approach, with stated limits and measurable goals, more solid than generic promises about merging mind and machine; now it is time to wait for the data.