More than 700,000 people have a cochlear implant [1]. A microphone picks up sound, a processor converts it into electrical pulses, and an electrode inside the cochlea delivers them to the auditory nerve. The brain learns to interpret them as sound. We have been connecting electronics to the nervous system for decades, and it works.
What is new is the scope. Today there are chips that give reading back to people who are almost blind, arms that transmit touch, tattoos that record brain waves, and paralyzed patients who “speak” at about 60 words per minute with an implant. This post goes over what exists, what is in clinical trials, and what is still hypothesis.
Visual Implants
Earlier retinal implants gave little more than flashes of light. PRIMA is different: a chip measuring 2 × 2 mm and 30 microns thick, with 378 electrodes, placed under the retina and working like a tiny solar panel. Augmented reality glasses project infrared light onto it, and the chip stimulates the retinal cells that are still alive [2].
The European trial included 38 patients with geographic atrophy (the advanced form of macular degeneration, which destroys central vision). Their average age was 78.9 years. At 12 months, 26 of the 32 evaluated (81%) had gained at least two lines on the visual acuity chart, and many were able to read again [2]. It is partial central vision, not normal sight, and it requires rehabilitation. The manufacturer, Science Corporation, has applied for authorization in Europe [2].
Arms: touch that reaches the brain
A robotic arm that moves with the mind is old news. The problem was that the user felt nothing and had to look at the hand to know whether it was gripping properly.
The e-OPRA system, developed in Sweden, combines osseointegration with electrodes implanted in nerves and muscles. Three patients used it for years in their daily lives, and they use it to modulate the force with which they hold objects [3]. It is an arm anchored to the bone, controlled by real muscle signals and with a sense of touch.
In the United States, a team at the Cleveland Clinic combined intuitive motor control, a sense of touch and a sense of hand movement in a single prosthesis. The two participants performed the tasks with a precision similar to that of people without an amputation [4]. A study funded by the NIH with eight people is now testing the combination of osseointegration, muscle reinnervation and implanted sensors [5].
Hearing, vests and new senses
Hearing is covered by the cochlear implant, which replaces the cells of the inner ear by directly stimulating the auditory nerve. Only 2.5% of people with profound deafness worldwide have one, although more than a million devices have been fitted [1]. It does not restore hearing, but gives a representation of sounds. It is the last resort when a hearing aid does not work. In Spain it is covered by the public health system.
For those who do not get an implant, David Eagleman (Baylor College of Medicine) developed a vest with vibration motors that translates sound into tactile patterns, and a wrist version called Buzz. This is sensory substitution: the skin receives what the ear does not pick up, and the brain learns to decode it with training. Eagleman and his colleague Perrotta summarize in a review the idea of going beyond substituting a sense and adding a new one, such as magnetoreception [6]. The same applies to data that no human perceives, such as the state of the market or the weather.
Neil Harbisson, who was born unable to perceive colors, has worn an osseointegrated antenna since 2004 that converts color into sound vibration, and his British passport recognizes him as a cyborg [7]. It is a unique case, with no formal study behind it, but it shows that the brain adapts to a new channel. There are also commercial body compasses: the North Sense device attaches to a small implant and vibrates when you face magnetic north [8].
Skin: chips and tattoos
NFC chips. About the size of a grain of rice, they are injected between the thumb and forefinger. Biohax, the Swedish company that has done the most to promote this practice, reported more than 4,000 installations five years after its launch [9]. They are used to open doors, clock in at work or serve as a train ticket. They have no battery or GPS and can only be read from a few centimeters away. The price is around $150 [10].
Electronic tattoos. These are ultra-thin sensors that stick to the skin or are painted onto it. In 2024, a group at the University of Texas at Austin, with Nanshu Lu and José Millán, used conductive ink to print EEG electrodes directly onto hairy scalp. In five participants they kept stable contact for up to 24 hours [11]. In 2026, a Penn State team published in PNAS a conductive ink that is painted on like makeup and records ECG (electrocardiogram), EEG (electroencephalogram) and EMG (electromyogram: muscles); in the trial, the electrodes recorded ECG during twelve hours of everyday activity [12].
This is not on the market, but it points to an EEG that can be worn for weeks.
Neural implants
Speaking without a voice
The most solid result of recent neurotechnology is speech. In 2023, a Stanford team implanted electrodes in the motor cortex of a patient with ALS and decoded attempted speech at 62 words per minute, 3.4 times faster than the previous record. Natural conversation runs at around 160 [13].
In 2024, a UC Davis group achieved 97.5% sustained accuracy over 8.4 months in a patient with ALS, with conversations at about 32 words per minute and more than 248 hours of cumulative use [14]. In 2025, the Stanford and BrainGate team decoded “inner speech”, the kind that is imagined without trying to pronounce it, with 74% accuracy over a vocabulary of 125,000 words. The signals are weaker than those of attempted speech, but they exist [15].
The manufacturers
Neuralink (USA) had about 21 participants in early 2026 [16]. Synchron introduces its device through a vein, like a stent, without opening the skull, with lower resolution but less surgical risk. Its COMMAND study reported six patients with no serious adverse events at 12 months [17]. Precision Neuroscience uses a sheet of 1,024 electrodes that rests on the surface of the brain, and has FDA authorization for uses of up to 30 days [17].
China is advancing in parallel. NeuCyber (Beijing) had implanted seven patients with its semi-invasive device by March 2026, and its director admits a lag of about three years behind Neuralink [16]. StairMed (Shanghai) did three implants in 2025 and plans about 40 in 2026 [18].
As of the sources I have been able to cross-check, no brain implant for paralysis yet has full commercial approval. Industry estimates point to 2028-2030 for limited availability [17].
Memory and intelligence
There is no data on “implants that give more intelligence” in healthy people. What exists is a 2018 pilot study (Wake Forest and USC) with eight epilepsy patients who already had electrodes in the hippocampus. The researchers recorded the activity pattern of correct answers in a memory test and “replayed” it through the same electrodes. Short-term memory improved by between 35% and 37% over baseline [19]. It is funded by DARPA and no one has replicated it at scale.
I think the first real use of this idea will be restoring memory after a stroke or in early-stage Alzheimer’s, not increasing intelligence. With 40-100 stimulated neurons, as in that pilot, the effect is a targeted aid, not an enhanced brain.
Communicating without speaking, over the internet
This has already been done in the lab. In BrainNet (University of Washington, 2019), two people (senders) decided via EEG whether a piece in a Tetris-like game should be rotated, and that decision reached the brain of a third person (receiver) over the internet through transcranial magnetic stimulation. Five groups of three people completed the task with 81.25% accuracy [20]. It is a one-bit channel (yes/no), with equipment the size of a room.
It could be that the practical path does not go through connecting two brains, but through combining what already exists: an implant that decodes attempted or imagined speech [13][14][15], a language model that turns it into text, and a network that sends it. That is already technically possible for patients with an implant. I think it will arrive sooner as a communication tool for people with ALS or stroke than as “telepathy” for everyone.
Another hypothesis: if electronic tattoos [11][12] come to provide a reliable and comfortable EEG, it could be that non-invasive BCIs gain ground for simple tasks (cursor, selection) without surgery, and that the implant remains for those who need speech-level speed.
Practical guide: what exists today and how to access it
Severe or profound hearing loss. The cochlear implant is an established technique [1]. Ask at an otolaryngology department whether you meet the criteria.
Arm amputation. Ask about centers with osseointegration and targeted muscle reinnervation. There are few, and the studies with touch are still in trial [3][4][5].
Advanced macular degeneration. PRIMA is under regulatory review in Europe and the United States [2]. In the meantime, the way forward is to consult an ophthalmologist about open trials in your area.
Paralysis or ALS and communication. BCI trials recruit in several countries. Search for them in registries such as ClinicalTrials.gov or the European clinical trials registry (CTIS), using the terms “brain-computer interface”.
No condition, just curiosity. The NFC chip is the only thing you can buy today [9][10]. Its function is to identify you to a nearby reader; it does not record or transmit your location. Haptic bracelets like Buzz are the no-surgery option for experimenting with an added sense [6].
Electronic tattoos and most brain implants. They cannot yet be obtained outside of trials [11][12][16].
References
[1] NIDCD (National Institute on Deafness and Other Communication Disorders, NIH). Cochlear Implants (fact sheet; 736,900 devices registered as of December 2019). Reliable. https://www.nidcd.nih.gov/sites/default/files/cochlearimplants.pdf
(The “more than a million” figure and the 2.5% coverage come from the statistics summary: https://voxbooster.com/blog/cochlear-implant-statistics-2026/ With reservations.)
[2] Sahel J-A (senior author), PRIMAvera group, et al. “Vision Restoration with the PRIMA System in Geographic Atrophy Due to AMD”. New England Journal of Medicine, Oct 20, 2025. Recent. https://presse.inserm.fr/en/un-implant-sous-retinien-restaure-partiellement-la-vision-de-personnes-atteintes-de-dmla/71276/
(81% figure in: https://new.nzoptics.co.nz/live-articles/implant-restores-sight-in-amd-trial)
[3] Ortiz-Catalan M, et al. “Self-contained neuromusculoskeletal arm prostheses”. New England Journal of Medicine, 2020. With reservations. https://opedge.com/prosthetic-hand-touch-sensation-functions-in-real-world-users/
[4] Marasco PD, et al. “Neurorobotic fusion of prosthetic touch, kinesthesia, and movement in bionic upper limbs promotes intrinsic brain behaviors”. Science Robotics, Sept 2021. With reservations. https://www.newswise.com/articles/cleveland-clinic-researchers-develop-bionic-arm-that-restores-natural-behaviors-in-patients-with-upper-limb-amputations
[5] Shirley Ryan AbilityLab / Northwestern Medicine / Integrum. e-OPRA study with osseointegration, muscle reinnervation and implanted EMG (NIH funding, 8 subjects). With reservations (ongoing, no results). https://www.sralab.org/research/labs/bionic-medicine/news/87-million-nih-grant-first-its-kind-bionic-arm-osseointegration-study
[6] Eagleman DM, Perrotta MV. “The future of sensory substitution, addition, and expansion via haptic devices”. Frontiers in Human Neuroscience, 16:1055546, 2023. Theoretical framework. https://doi.org/10.3389/fnhum.2022.1055546
[7] Harbisson N. Profile in The Guardian, May 6, 2014. With reservations. https://www.theguardian.com/artanddesign/2014/may/06/neil-harbisson-worlds-first-cyborg-artist
[8] Teknisk Ukeblad. Report on North Sense (Cyborg Nest). With reservations. https://www.tu.no/artikler/dette-implantatet-til-3500-kroner-gjor-deg-til-en-kyborg/377013
[9] AFP / Yahoo News. “Swedish cyborg craze sees more than 4,000 Swedes insert chips under their skin”. With reservations. https://sg.news.yahoo.com/swedish-cyborg-craze-sees-more-073926222.html
[10] Entrepreneur Middle East (citing The Economist). Cost and operation of RFID/NFC chips in Sweden. With reservations. https://mena.entrepreneur.com/technology/get-under-the-skin-why-swedens-citizens-are-opting-for/318123
[11] University of Texas at Austin (Lu N, Millán JdR, He X, et al.). EEG electrodes printed as an electronic tattoo on the scalp, Cell Biomaterials, 2024. Recent. https://cockrell.utexas.edu/news/printed-e-tattoo-ink-credible-at-reading-brainwaves/
[12] Penn State University. Paintable conductive ink for ECG/EEG/EMG electrodes. PNAS, 2026. Recent. https://www.designboom.com/technology/fierce-shark-shy-tiger-tattoos-etattoos-health-paint-on/
[13] Willett FR, et al. “A high-performance speech neuroprosthesis”. Nature, 620, 2023. Reliable. https://www.nature.com/articles/s41586-023-06377-x
[14] Card NS, et al. “An Accurate and Rapidly Calibrating Speech Neuroprosthesis”. New England Journal of Medicine, 2024. Reliable. https://profiles.stanford.edu/francis-willett
[15] Kunz EM, et al. “Inner speech in motor cortex and implications for speech neuroprostheses”. Cell, Aug 14, 2025. Recent. https://www.the-scientist.com/brain-computer-interface-lets-users-communicate-using-thoughts-73275
[16] Reuters (Mar 20, 2026), picked up by MDDI. Data on Neuralink (≈21 participants) and NeuCyber (7 Beinao-1 implants). With reservations. https://www.mddionline.com/neurological/the-china-based-neuralink-competitor-catching-up-to-elon-musk
[17] Next Wave Insight. Summary of trials by Synchron (COMMAND) and Precision Neuroscience (Layer 7, 510(k) clearance). With reservations. https://nextwavesinsight.com/bci-neuralink-synchron-clinical-trials-2026/
[18] DealStreetAsia / Reuters. Funding and trial plan of StairMed. With reservations. https://www.dealstreetasia.com/?p=478117
[19] Hampson RE, et al. “Developing a hippocampal neural prosthetic to facilitate human memory encoding and recall”. Journal of Neural Engineering, 15(3):036014, 2018. With reservations. https://viterbischool.usc.edu/?p=15589
[20] Jiang L, Stocco A, et al. “BrainNet: A Multi-Person Brain-to-Brain Interface for Direct Collaboration Between Brains”. Scientific Reports, 9:6115, 2019. With reservations. https://pmc.ncbi.nlm.nih.gov/articles/PMC6467884
Notes on reliability:
[2]: trial without a control group and funded by the manufacturer; no independent replication yet.
[3], [4], [5]: samples of 2 to 3 people (or ongoing study); no replication at scale.
[6]: review written by the founders of the company that sells these devices.
[7], [8]: individual cases or reports, without peer review.
[9], [10]: figures provided by the company or by press citing third parties.
[11], [12]: very small samples (5 participants in [11]) and no replication.
[15]: four participants, no independent replication.
[16], [17], [18]: data from companies or statements by executives, not published with peer review; [17] is a secondary aggregator.
[19]: pilot with eight patients, same research group, DARPA funding.
[20]: five groups of three people and a single-bit task.
Some links point to press releases or news coverage instead of the original article because I did not locate a verified DOI; it is advisable to search for the article by title before publishing.