Three Modalities, One Resolution Problem

Retinal prosthesis technology in 2026 divides into three principal sub-domains — electronic electrostimulation, photovoltaic subretinal implants, and cell-based biological replacement — each attempting to restore functional vision in patients whose photoreceptors have degenerated due to retinitis pigmentosa (RP) or age-related macular degeneration (AMD). The central challenge uniting all three is the same: even the most advanced systems in this patent dataset achieve only approximately 20/1400 vision, well below the 20/200 legal blindness threshold, because of the fundamental mismatch between electrode size and individual neuron targets.

Electronic prostheses use surgically implanted microelectrode arrays — placed epiretinally or subretinally — to deliver patterned electrical stimulation to surviving inner retinal neurons, primarily bipolar and ganglion cells. Image capture is performed by a wearable camera, with signal processing encoding a stimulus pattern transmitted wirelessly to the implant. Current electrodes range from 50 to 450 µm in diameter, far larger than individual cell targets, which limits spatial resolution. Cornell University’s EP patent explicitly identifies this electrode-to-cell size mismatch as the key constraint preventing clinically meaningful acuity.

Photovoltaic prostheses eliminate transcutaneous wires by using subretinal photodiode arrays powered by projected near-infrared or visible light from a video-processing headset. Each pixel converts incident light to charge, stimulating overlying bipolar cells. The PRIMA implant from Pixium Vision, referenced in the Stanford University CN patent, has achieved prosthetic letter acuity of 20/460 to 20/565 in AMD patients — a meaningful improvement over electronic systems but still far short of the 20/200 threshold. Stanford’s analysis indicates that 50 µm pixels would be required to reach 20/200.

Cell-based biological approaches represent a fundamentally different strategy: rather than bypassing degenerated photoreceptors with electrical stimulation, they aim to reconstitute the retinal pigment epithelium (RPE) layer that provides metabolic support to surviving photoreceptors. According to NIH-affiliated filings in this dataset, this approach uses hESC-derived or iPSC-derived RPE cells, with multiple assignees — Astellas Institute for Regenerative Medicine, Lineage Cell Therapeutics, and the U.S. Government — holding active filings in this sub-domain.

From Electrode Arrays to Scaffold Implants: The Innovation Timeline

The retinal prosthesis patent record in this dataset spans from 1999 priority dates to 2026 pending applications, tracing a clear arc from broad foundational device claims toward increasingly specific manufacturing parameters and AI-enabled monitoring infrastructure.

The earliest filings in this dataset originate with Second Sight Medical Products, whose retinal color prosthesis patent (AU, 2005, priority dating to 1999) describes electrode-based electrical retinal stimulation with color channel differentiation and telemetry. The parallel photovoltaic lineage begins with Alan Y. Chow’s subretinal MMRI-4 device patent (DE, 2005), describing PiN/NiP configurations that transduce light into retinal-stimulating currents. STMicroelectronics filed foundational retinal prosthesis patents in Italy as early as 2012.

The mid-stage development period (2012–2020) introduced the encoder-based approach: Cornell University’s EP patent (2022, reflecting a longer prosecution timeline) describes systems that mimic natural ganglion cell firing at single-cell or small-cluster resolution. According to EPO records, this patent explicitly claims that each transducer targets a single cell or small number of cells, with conversion occurring on the same time scale as the normal retina — a substantially more ambitious specification than bulk electrode stimulation.

The cell therapy cluster expands substantially in this period: Lineage Cell Therapeutics filed methods for measuring therapeutic effects of retinal disease therapies (WO, 2018) and Astellas Institute for Regenerative Medicine published early hESC-RPE transplantation results (TW, 2021). The most recent filings (2023–2026) show clear movement toward photovoltaic prostheses with optical field confinement, AI-guided retinal monitoring, advanced cell-scaffold implant manufacturing, and optogenetic/gene-therapy convergence with prosthetic restoration.

“Even advanced retinal prosthesis systems achieve only approximately 20/1400 vision — well below legal blindness thresholds of 20/200 — due to electrode size relative to individual neurons.”

Assignee Geography and IP Concentration

Retinal prosthesis patent activity in this dataset spans eleven jurisdictions — EP, US, CN, AU, IT, KR, JP, TW, WO, IN, and BR — with distinct geographic patterns by sub-domain. Electronic and photovoltaic device hardware patents are concentrated in a small number of US academic and startup assignees, while cell-based approaches show the broadest multi-jurisdictional prosecution, and diagnostic/AI enablement shows the most distributed assignee base.

The United States and the European Patent Office host the core electronic and photovoltaic device patents. China is prominent in cell-therapy and photovoltaic prosthesis filings — Stanford, NIH, Cornell, and Lineage all hold CN filings. Japan appears repeatedly in diagnostic and AI monitoring filings, with Roche, Notal Vision, Iridex, EPFL, and Magic Leap all holding JP grants or pending applications. Within this dataset, active photovoltaic prosthesis device filings are concentrated in CN (Stanford) and legacy DE filings (Chow), leaving US, EP, JP, and KR as potential prosecution opportunities for next-generation photovoltaic architectures.

Cell-based approaches show broader multi-jurisdictional prosecution — spanning TW, AU, US, WO, CN, and JP — suggesting greater commercial deployment ambitions. According to data tracked by WIPO, PCT filings in regenerative medicine have grown substantially over the past decade, consistent with the multi-jurisdictional prosecution strategy observed in the Lineage Cell Therapeutics portfolio (AU, US, WO, BR, JP) within this dataset.

The Cornell University portfolio (EP, CN) is notable for its explicit extension of the retinal encoding architecture beyond ophthalmology: the EP patent states that the system “may be used with robotic or other mechanical devices, where processing of visual information is required,” positioning the neural encoding IP as applicable to machine vision and neuromorphic computing — a potential expansion of the addressable market for licensing.

Five Emerging Directions Redefining the Field

The most recent filings in this dataset (2023–2026) reveal five distinct technology directions that are reshaping the competitive landscape beyond the established electrode-array paradigm.

1. Optical Field Confinement in Photovoltaic Implants

The Stanford University CN filing (2023) introduces a pre-charging architecture that directly addresses the resolution-limiting contrast degradation that has constrained photovoltaic prosthesis acuity. Pixels destined to go dark in the next video frame are pre-conditioned to act as transient local return electrodes, reducing inter-electrode crosstalk and improving contrast. The filing targets 20/100 or better acuity via a 25 µm pixel pitch — a substantial improvement over the 20/460–20/565 range achieved by current PRIMA implants.

2. Scaffold-Engineered RPE Patches with Defined Microarchitecture

The U.S. Government NIH filing (CN, 2024) defines highly specific PLGA scaffold parameters: 20–30 µm thickness, approximately 1:1 DL-lactide/glycolide ratio, less than 1 µm average pore size, and 150–650 nm fiber diameter for polarized RPE cell delivery. This represents a shift from cell suspension injection toward structured tissue-replacement implants with defined mechanical and transport properties — and signals that IP around scaffold geometry and cell composition is tightening.

3. AI-Enabled Remote Retinal Monitoring for Prosthetic Follow-Up

Notal Vision Ltd. holds two active US patents (2023 and 2025) on daily home-based OCT monitoring systems that generate treatment recommendations from retinal fluid progression rates. The clinical workflow — short-interval OCT with algorithmic scheduling of interventions — is directly applicable to monitoring of both cell-based implants and subretinal device performance post-operatively. Novartis (US, 2026), Hoffman-La Roche (US, 2022), and EPFL (JP, 2025) hold additional filings on AI-based retinal disease monitoring from OCT data, creating what the dataset’s strategic analysis describes as an “IP thicket” around post-implant monitoring.

4. Retinal Stimulation Combined with Spectroscopic Data Capture

Diamentis Inc. (BR, 2024) discloses a bidirectional system that both exposes the retina to controlled light flashes and records the resulting spectral response. This stimulation-and-sensing architecture may enable real-time calibration of prosthetic stimulation parameters and electrophysiological confirmation of retinal response — a capability with direct application to closed-loop prosthetic systems.

5. Gene Therapy Convergence with Prosthetic Restoration

Filings from Octant Inc. (WO, 2025) and Nightstarx Limited (SG/IL, 2021) address gene-therapy-mediated restoration of photoreceptor function in RP. These approaches are positioned not as competing technologies but as potential complements to prosthetic interventions — gene therapies that preserve residual photoreceptors while prostheses serve patients in whom photoreceptors are fully lost. As noted by FDA guidance on combination product regulation, multi-modal therapeutic strategies involving both biological and device components require careful regulatory pathway planning.

Strategic Implications for IP Teams and R&D Leaders

The retinal prosthesis patent landscape in 2026 presents five distinct strategic signals for IP professionals and R&D decision-makers evaluating this space.

Resolution gap remains the central technical barrier. Even the most advanced systems in this dataset achieve only 20/460 vision (photovoltaic) or 20/1400 (electronic). R&D teams should prioritise pixel miniaturisation below 25 µm and inter-electrode crosstalk suppression — as addressed by Stanford’s optical pre-charging architecture — as primary development pathways to commercially viable acuity thresholds. The PatSnap IP intelligence platform enables teams to monitor competitor filing activity in this rapidly evolving technical area.

Cell-based RPE approaches are entering a manufacturing-definition phase. Multiple assignees — Astellas, Lineage, and NIH — are now filing on specific cell purity, scaffold geometry, and transplantation measurement parameters rather than broad method claims. This signals IP maturation and tightening freedom-to-operate constraints around scaffold design and cell composition. Teams entering this space should conduct thorough FTO analysis before committing to specific PLGA formulations or cell preparation protocols.

AI-guided remote OCT monitoring is becoming an IP-dense enabling layer. Notal Vision, Roche, Novartis, and Genentech all hold filings on AI-based retinal disease monitoring from OCT data. Any prosthetic device requiring post-implant monitoring will need to either license, design around, or partner within this emerging IP thicket. The PatSnap patent analytics suite enables systematic mapping of this monitoring IP landscape.

Convergence of gene therapy, cell therapy, and electronic prosthetics creates a multimodal treatment paradigm. IP strategists entering this space should model freedom-to-operate across all three modalities, as clinical practice is likely to combine residual vision preservation (gene/cell therapy) with prosthetic augmentation. Blocking patents in one modality may constrain commercial strategies in adjacent ones — a risk that requires cross-domain portfolio mapping.

Geographic white spaces remain for photovoltaic device hardware. Within this dataset, active photovoltaic prosthesis device filings are concentrated in CN and legacy DE. US, EP, JP, and KR represent potential prosecution opportunities for next-generation photovoltaic architectures, particularly those incorporating the optical field confinement approach introduced by Stanford’s 2023 CN filing.