Brain-Computer Interfaces in 2026: Neuralink, Synchron, and the Patient Pathway
Where BCI sits in 2026 — Neuralink and Noland Arbaugh, Synchron Stentrode, Precision Neuroscience Layer 7, Onward, Paradromics, and the research-to-product path.
In January 2024, a paralysed man named Noland Arbaugh became the first human recipient of a Neuralink N1 implant. Within weeks of his recovery he was playing online chess using only his thoughts, and within months he was streaming on social media about life with a chip in his motor cortex. That moment did for brain-computer interfaces what AlphaGo did for AI in 2016 — converted a niche research field into something the broader public understood was actually happening. Two years on, in 2026, the BCI field has multiple human-implanted patients across several companies, real regulatory pathways at the FDA, and a coherent technology landscape with three or four genuine paradigms.
This post walks through where BCI is today — Neuralink and its patient cohort, Synchron’s vascular Stentrode, Precision Neuroscience’s surface Layer 7, Onward, Paradromics, the NIH BRAIN Initiative substrate, and the patient pathway from research curiosity to deliverable medical product.
Neuralink, two years past first-in-human#
Neuralink’s N1 implant is the most invasive of the major BCI approaches. The device sits in a 25-millimetre coin-shaped housing inserted directly into the skull, with 64 ultrafine threads spreading into the motor cortex carrying 1,024 electrodes. The R1 surgical robot does the actual thread placement, which is the engineering innovation that makes the procedure repeatable without a human neurosurgeon doing each insertion by hand.
The PRIME (Precise Robotically Implanted Brain-Computer Interface) trial expanded through 2024 and 2025. Beyond Noland Arbaugh, additional patients received implants — Alex, the second patient publicly identified, and several others whose identities Neuralink has not disclosed. The headline early-2024 setback — Noland’s threads partially retracting from the cortex within weeks of implantation, reducing electrode yield — has been the dominant engineering focus since, and the company has reported reduced retraction in subsequent implants through firmer anchoring and revised insertion depth.
By 2026 Neuralink has expanded into Canada and the United Kingdom with parallel CAN-PRIME and CONVOY trials, and the company has begun early work on the second-generation product targeting vision restoration via a cortical implant in the visual cortex, branded Blindsight. That product remains in animal trials and the timeline to human use is still measured in years rather than months.
Synchron and the Stentrode, the vascular approach#
Synchron’s Stentrode is the most interesting counter-example in the field because the surgical approach is fundamentally different. The device is a stent-mounted electrode array that is delivered through the jugular vein and parked in the superior sagittal sinus, a blood vessel sitting against the motor cortex. There is no craniotomy, no direct cortical insertion, no piercing of the brain tissue. The procedure is performed by an interventional neuroradiologist in a catheter lab, not by a neurosurgeon in an operating theatre.
Synchron received FDA breakthrough device designation in 2020 and IDE approval to begin US human trials in 2021. The COMMAND trial has implanted multiple patients through 2022, 2023, and 2024, and the company has reported successful chronic recording over multi-year follow-up. The trade-off is unambiguous. The Stentrode has dramatically lower electrode count — sixteen electrodes versus Neuralink’s thousand-plus — and therefore lower signal resolution, but the procedure is genuinely repeatable at the volume modern healthcare can support. The path from research to product runs through cardiology catheter labs that already exist in every major hospital.
Through 2025 Synchron expanded its partnership with Apple to enable Vision Pro and iPhone control directly from Stentrode signals, and the company moved into pivotal trials targeting US regulatory clearance for the first general-population BCI indication.
Precision Neuroscience, the surface approach#
Precision Neuroscience, founded by a former Neuralink co-founder, took yet a third route. The Layer 7 Cortical Interface is a flexible film thinner than a human hair that sits on the surface of the brain rather than penetrating it. The film is inserted through a slit in the dura, no skull bone is permanently removed, and the device can be removed at the end of the study or treatment. The electrode count is high — over a thousand on the current generation — but the recording is from the cortical surface rather than from individual neurons, which is closer to high-resolution ECoG than to single-unit recording.
Precision received FDA 510(k) clearance for the wedge-shaped Layer 7 in April 2025, the first BCI of its kind to clear regulatory in the US, and through late 2025 the company expanded into chronic studies. The clinical positioning is closer to surgical-mapping and epilepsy localisation initially, with the broader cognitive and motor-prosthesis indications following.
Onward, Paradromics, and the wider field#
Onward Medical, headquartered in the Netherlands and Switzerland, has the spinal cord half of the BCI story — the implanted ARC-IM stimulator pairs with a brain-side device to bridge spinal cord injuries, restoring movement to paralysed patients in a way that pure brain recording cannot. The 2023 Nature paper showing a patient walking through thought control was their work in collaboration with the EPFL group in Lausanne.
Paradromics, based in Austin, is taking the high-channel-count cortical approach closer to Neuralink but with a different surgical and electrode-array design. The Connexus Direct Data Interface aims for higher per-electrode signal quality than Neuralink and a more modular surgical procedure. Paradromics received FDA breakthrough device designation in 2023 and began human studies in 2024.
Blackrock Neurotech, the long-time research workhorse whose Utah Array has been in dozens of academic patients for over a decade, continues to expand its MoveAgain BCI for commercialisation. And the NIH BRAIN Initiative — the federal research substrate that funded much of the underlying neuroscience — continues to support the academic groups at Pittsburgh, BrainGate at Brown, UCSF, and Stanford that do the work no individual company can.
The patient pathway from research to product#
The honest framing of where BCI sits in 2026 is that there are now multiple chronically-implanted patients across multiple companies, the FDA has a working regulatory framework, and the first commercial indications are real but narrow. The pathway from research to broadly-available product looks like this. Phase one is paralysed patients with severe motor disability, where the alternative is no communication at all and the device pays for itself against a lifetime of care costs. Phase two is broader motor-prosthesis indications and possibly vision restoration. Phase three is cognitive enhancement and the broader speculative applications that get the press but are nowhere near a clinical trial.
The realistic timeline for phase one becoming a commercially-available medical device for tens of thousands of patients is the late 2020s. The realistic timeline for phase two is the 2030s. Phase three remains genuinely uncertain.
What this means for the broader tech ecosystem#
For software and AI teams not in the medical-device industry, the BCI story has two practical implications. First, the upstream neural-decoding stack — the models that translate spike trains into intent — has converged on transformer-style architectures and is increasingly shared between companies, which means the AI talent pool overlaps with the broader LLM ecosystem. Second, the downstream consumer pathway — controlling phones, computers, AR headsets, smart-home devices from a BCI — is being built on top of the same Apple, Google, and Microsoft accessibility APIs that already power eye-tracking and switch-control. The product surface a Stentrode patient sees is recognisably an Apple device, not bespoke neurotech software.
Where pdpspectra fits#
Our AI and LLM integration practice builds the high-throughput data and modelling infrastructure that medical-device companies and academic labs need for chronic neural recording, signal decoding, and clinical-trial data management.
Related reading: the multimodal AI 2026 post, the Apple Vision Pro evolution post, and the AI clinical trials post.
BCI is no longer science fiction; it is a medical-device industry with real patients and real regulators. Talk to our team about your neurotech project.