Second Sight and the Argus II: what happens to an implant when its maker moves on

Barbara Campbell was changing trains in a New York subway station when her artificial vision switched off. "I was about to go down the stairs and all of a sudden I heard a little beep beep beep sound," she later told IEEE Spectrum. It was 2013, and for four years an implant on her retina had given her patches of light and dark. Second Sight, the company that made the device, attempted repairs, but it never worked again. The dead device is still in her eye.

Her story opens a Tech Anatomy ERROR 404 episode about the Argus II, the first retinal implant approved in the United States. The larger part of that story began six years later, when the company stopped making the implant and more than 350 people around the world were left carrying it.

How the Argus II worked

Second Sight was founded in 1998 near Los Angeles with Alfred Mann, who had built companies around pacemakers and cochlear implants. The logic came from the cochlear implant: if electrical pulses delivered to a nerve can restore a form of hearing, pulses delivered to the retina might restore a form of sight. Researchers had shown from the late 1980s that electrically stimulating a blind person's retina produces a small spot of light.

The device was designed for end-stage retinitis pigmentosa, an inherited disease that destroys the photoreceptors, the cells that capture light, while often sparing the nerve cells behind them. A camera on a pair of glasses fed a processor worn on the belt, which sent a simplified pattern wirelessly to an array of 60 electrodes on the retina. The surviving cells passed the signal to the brain, which read it as a constellation of flashes. The episode's analogy is a stadium scoreboard of 60 bulbs next to a television with millions of pixels: enough for a block arrow, not for a face.

The five-year results of the pivotal trial, published in Ophthalmology in 2016, measure what that meant. Twenty-four of the 30 participants still had a working system at five years. At that point, 81 percent of those tested performed better with the system on than off at locating a bright square on a screen, 50 percent at judging the direction of motion, and 38 percent on a test of visual acuity, at the very bottom of the measurable scale. They also did better at finding a door across a room. Nobody read a book with the Argus II, but for people with no light perception, finding a doorway changed daily life.

Approval through a narrow door

The Argus II received its CE mark in Europe in 2011 and FDA approval in February 2013 through the Humanitarian Device Exemption. At the time, this pathway was limited to devices for conditions affecting fewer than 4,000 people a year in the United States, and it asks for a reasonable assurance of safety and probable benefit rather than the full effectiveness standard. The FDA's decision rested on data from 30 patients.

When the company changed course

On 18 July 2019, Second Sight wrote to patients that it was phasing out the Argus II to concentrate on Orion, an implant that stimulates the visual cortex directly and bypasses the eye. In March 2020, citing the effect of the pandemic on financing, it laid off most of its remaining staff and announced its intention to wind down operations. In August 2022 it merged with Nano Precision Medical, a developer of drug-delivery implants whose shareholders took about 77 percent of the combined company, renamed Vivani Medical. The neurostimulation assets, Orion included, were later placed in a subsidiary, Cortigent, which now holds the Argus II's US approval.

IEEE Spectrum's 2022 investigation described the consequences. MRI providers are told to contact the manufacturer before scanning an Argus II patient; Ross Doerr, whose doctor wanted to rule out a brainstem tumour, could not get that answer and had a CT scan instead. "I still don't know if I have a brainstem tumor or not," he said, and summed up the experience as "fantastic technology and a lousy company". Terry Byland, the only person to have received implants in both eyes, said: "If something does go wrong with it, well, I'm screwed." In the Netherlands, Jeroen Perk's processor failed in late 2020: "I had no vision, no Argus and no support from Second Sight."

Two readings of the same collapse

The episode rejects the comparison with Theranos: the Argus II worked within the limits its trials described. The first reading is therefore economic. Retinitis pigmentosa affects around 1 in 4,000 people, only the end-stage subset qualified, and the device sold for about $150,000 before surgery. Robert Greenberg, a co-founder of Second Sight, told IEEE Spectrum that "the size of the market turned out to be smaller than we had thought" and that "with all the overhead of sales and regulatory people, it wasn't profitable."

A cost-effectiveness study published in BMC Ophthalmology in 2014 estimated around €14,600 per quality-adjusted life year, within the range European payers generally accept. That study was not independent: one author was a Second Sight employee and the lead author's consultancy was paid to conduct it. Its model also assumed that patients would receive "regular product support in terms of software updates, special training and rehabilitation". The second reading starts there. The value of an implant depends on a service that lasts as long as the patient, and that obligation, what the episode calls the "orphan hardware" problem, was never priced or guaranteed by anyone. As the episode puts it: "When you put technology inside a human body, you are not selling a product. You are entering a relationship that outlives your business plan."

What the record does not settle

The Argus II story is one company in one small indication. It does not show how often implant makers abandon patients, and it does not say who should have carried the support obligation: the company, its investors, the regulator or the health systems that reimbursed the device. The patient accounts come largely from a single investigation.

What to watch

The next generation is already testing the same model. PRIMA, a wireless chip placed under the retina and powered by near-infrared light from glasses, was developed by the French company Pixium Vision, whose assets Science Corporation acquired in 2024 through a French commercial court. In the PRIMAvera trial, published in the New England Journal of Medicine, 26 of 32 people with geographic atrophy due to age-related macular degeneration who completed a year of follow-up had a clinically meaningful gain in visual acuity, in a single-group study without a control arm. PRIMA received a CE mark under the EU Medical Device Regulation, announced in July 2026, and in the United States it holds Humanitarian Use Device designations, the step before the same exemption pathway the Argus II used.

Geographic atrophy concerns far more people than end-stage retinitis pigmentosa. The indicators worth following are whether reimbursement follows in Europe, whether the FDA grants an exemption, and whether anyone, the company, a payer or a regulator, writes down what happens to patients if the product is ever discontinued.

Listen to the episode

ERROR 404 - The Untold Story of Bionic Eyes

Listen · 15 min