Science Corp’s retinal implant is restoring reading to people with advanced macular degeneration. The last company to try this left patients with dead hardware and unanswered phone calls.
Imagine you have spent years unable to read the menu in front of you, and then a small chip under your retina lets you do it again. The clinical results for Science Corporation’s PRIMA implant read like that. After training, more than 80 percent of recipients could read, do crosswords and use their phones, according to MIT Technology Review’s account of the device [1]. This is a real result for a disease that has no comparable treatment, and I do not want to bury it under skepticism.
I do want to ask the question the trial cannot answer. What happens in year eight, when the company’s priorities change?
What the evidence shows
PRIMA is a 2 by 2 millimeter wireless chip placed under the atrophic macula, paired with glasses that project near-infrared light onto it [2]. The implant has 378 pixels. Each carries an electrode and two photovoltaic cells that turn light into current and power the chip, so it needs no battery [1]. The target is geographic atrophy, an advanced form of age-related macular degeneration that affects more than five million people worldwide [1].
The pivotal study, published in the New England Journal of Medicine by Holz and colleagues, enrolled 38 people and 32 completed it. At 12 months, 26 of the 32 completers, about 81 percent, showed clinically meaningful improvement. The authors estimate 80 percent of all 38 would have met the endpoint, which is an estimate rather than an observed result. I could not open the paper itself, so I am relying on a trade summary of it, and that summary does not explain why six participants did not finish [4]. Science Corp’s own announcement puts the mean gain at 25.5 letters on a standard chart and says 84 percent could read letters, numbers and words [2]. That is the company describing its own trial, so I treat the numbers as a claim. The journal summary supports the headline responder rate, but not every figure in the announcement.
The three-year follow-up is more sobering and more useful. Among the 38 implanted participants there were 29 serious adverse events by 36 months. About 83 percent happened in the first six months and were mostly procedure-related, such as pressure spikes, bleeding under the retina and retinal tears. Only five occurred between months six and 36, with no late device extrusion or migration reported. Mean acuity gain reached nearly 30 letters by 36 months, and more than half the participants were using the system for harder tasks such as reading menus and checking train departures [3]. Ehsan Rahimy, MD, who reports these results, says he wants to see real-world European outcomes and five-year durability data [3].
Regulators have started to act on this. The device was approved in Europe in July 2026 under a CE mark, with a first commercial implant expected soon in Germany, while the US review continues and the device holds Breakthrough Device and Humanitarian Use Device designations [1][2].
The implant is the easy part to keep working
PRIMA is well designed against the failure that killed earlier devices. Because the chip is powered by light and has no battery or wires, it removes some of the ways an implant gives out. That matters.
But nobody sees with a chip alone. The system depends on camera glasses capturing roughly five megapixels at 30 frames per second, a cable to a handheld processor that converts video to infrared signals, a laser projector, and buttons for contrast, brightness and zoom [1]. Science Corp describes a zoom-in feature [2], while a trade summary of the journal paper notes that the current version delivers black-and-white vision and that grayscale software meant to enable facial recognition is still in development [4]. In other words, the part of this system a patient actually experiences is hardware and software that sits outside the body, is updated over time and is owned by a company.
Thirty years of watching systems outlive the teams that built them has made me wary of that arrangement. Software sitting between a person and their sight needs a named owner for patches, an update path that does not require the manufacturer to stay solvent, and a way for the patient to know which version they are running. Those are ordinary engineering questions. They become ethical ones when the user cannot simply unplug the thing.
None of the sources I read describes what happens to that external stack if the company changes course. MIT Technology Review’s piece does not cover costs, support or maintenance [1]. Science Corp’s announcement does not describe support services, and reimbursement applications are still being filed [2]. I am not claiming there is no plan. I am saying that a reader of the public record cannot tell what the plan is, and that gap is where patients have been hurt before.
We already ran this experiment
In 2022, IEEE Spectrum documented what happened to the Argus II, a retinal implant that reached the market. Second Sight, a Los Angeles company, gained European approval in 2011 and US approval in 2013, at roughly $150,000 per device in the US and still losing money. In 2019 it announced it was phasing out the retinal implant. In March 2020 it laid off most of its staff. More than 350 people worldwide had been implanted [5].
The human consequences are the part I cannot forget. Ross Doerr developed severe vertigo in 2020 and could not get an MRI, because Second Sight did not answer his calls about the implant. A CT scan found nothing, and he still does not know whether he has a tumor. Jeroen Perk’s processing unit shattered in November 2020, and he got a refurbished system by February 2021 only because patients and a doctor supplied spare parts. Terry Byland learned secondhand that his company was failing [5]. These were not people who regretted their implants. They were people whose devices had outlived the institution responsible for them.
Spectrum’s authors called the field’s mood a gold rush and warned that companies can fail and leave patients with unsupported devices [5]. Four years later, the question is whether anyone in the field learned from it.
What regulators approve and what they do not
Here is the structural problem. As I understand them, approval pathways examine whether a device is safe and effective in trial conditions. The trial for PRIMA enrolled 38 people at 17 sites in five countries [2]. A trial cannot tell you about the year-eight support burden, the spare-parts supply or the corporate lifecycle, because those only become visible after the commercial phase begins. The Argus II had approvals on two continents and its patients were still stranded.
Cost is the other unanswered detail. MIT Technology Review does not report a price [1], and Science Corp says only that reimbursement applications are underway [2]. For scale, the Argus II carried a US price tag of around $150,000 [5]. A sight-restoring device that only some health systems can fund will reproduce the access gap it was meant to close.
The rehabilitation burden is another dependency. Rahimy says ideal candidates need cognitive health, a strong support system and motivation to complete intensive postoperative rehabilitation, and he expects practices to add occupational therapists and low-vision specialists, with regional centers of excellence emerging [3]. Access to this technology will therefore track access to a service ecosystem, not just access to a surgeon, which raises an equity question that the device’s results alone do not answer.
The fair counterargument
There is a strong case against my worry. PRIMA targets a far larger population than the 350-plus people who received Argus II, since more than five million people have geographic atrophy [1][5]. The passive design is simpler. Reimbursement applications are underway in Europe, and serious adverse events were concentrated in the first six months [2][3]. Demanding heavy continuity obligations could raise prices or discourage small teams from attempting hard medical problems. A treatment for a blinding condition that delays for governance paperwork has a human cost too.
I take that seriously. Nothing in the record suggests Science Corp intends to walk away, and I am not accusing it of anything. Continuity rules are not a verdict on a particular company. They are what makes a hopeful early result safe to trust over a decade, for the patient, for the surgeon recommending it and for the next company trying something hard. The counterargument still argues for proportionate rules, not none.
What I would ask for
Four things, none of them exotic. First, a published continuity plan as a condition of market access, covering who services the external hardware if the manufacturer exits. Second, escrowed interface specifications and software so that a successor, a hospital or a patient community can keep devices running, which is effectively what Perk’s circle improvised by hand [5]. Third, a patient-held record of the implant’s configuration and safety conditions, including scanner compatibility, so that nobody is guessing about an MRI. Fourth, a funded wind-down commitment that does not depend on the company surviving.
The principle behind all four is control. The person wearing the system should not depend on a single company’s continued attention to keep their sight.
The test that matters
PRIMA may prove to be one of the first sensory prostheses that ordinary patients can actually receive. If it succeeds, millions of people will rely on a stack of hardware and software for a basic human sense. The measure of success will not be the first reading test. It will be whether a patient with a failing processor in 2034 has a number to call and a part to install. The Argus II patients are the evidence that this cannot be left to goodwill.
About the Author
Len Noe is a Solutions Architect at BeyondTrust, Transhuman, Podcaster, International Cyber Security Speaker, Author, Technical Evangelist, and Biohacker with 11 implanted microchips. A former blackhat with more than 30 years in technology, he has presented in over 70 countries and is featured in the documentary I Am Machine, which premiered at DEF CON 2025.
References
- MIT Technology Review. “This implant can restore sight for patients with macular degeneration.” MIT Technology Review, October 16, 2026. technologyreview.com
- Science Corporation. “Science Corp. Announces European Commercial Launch of PRIMA…” Science Corporation press page, July 22, 2026 (company self-report). science.xyz
- Ehsan Rahimy, MD. “PRIMA at 3 years: Early surgical risk, durable vision gains in GA.” Ophthalmology Times, 2026. ophthalmologytimes.com
- Review of Optometry. “Retinal implant restores vision in geographic atrophy” (trade summary of Holz FG, Le Mer Y, Muqit MMK, et al., “Subretinal photovoltaic implant to restore vision in geographic atrophy due to AMD,” New England Journal of Medicine, epub October 20, 2025). reviewofoptometry.com
- Eliza Strickland and Mark Ellis Harris. “Their Bionic Eyes Are Now Obsolete and Unsupported.” IEEE Spectrum, February 15, 2022 (updated February 17, 2022). spectrum.ieee.org