Neuralink in 2026: Three Humans Have Brain Chips. Here Is What the Clinical Trials Are Actually Showing — and What Comes Next.

In January 2024, Neuralink implanted its N1 chip — a device the size of a large coin, containing 1,024 electrodes — into the brain of Noland Arbaugh, a 29-year-old man paralyzed from the shoulders down in a diving accident. Two months later, Arbaugh demonstrated his ability to control a computer cursor with thought alone, playing online chess and communicating in real time — faster and more accurately than with the assistive technology he had previously relied on. By March 2026, Neuralink has three human participants in its ongoing PRIME trial, with the company reporting improvements in recording quality, bandwidth, and user performance across its second and third implants. And the technology it represents — brain-computer interfaces that restore communication and motor function to people with severe paralysis — is beginning to look less like science fiction and more like medicine.

What the Clinical Trials Are Actually Showing

• Patient 1 (Noland Arbaugh): the first implant had a setback — approximately 85% of the electrode threads retracted from the brain tissue in the weeks after implantation, reducing recording quality significantly. Neuralink's software team compensated by developing better algorithms for interpreting the remaining electrode data. Arbaugh continued to use the system for computer control and communication. The retraction issue prompted surgical technique and electrode design changes in subsequent implants.
• Patients 2 and 3: Neuralink has not disclosed the same level of detail about its second and third participants but has reported improved electrode retention following design changes, higher bandwidth neural recording, and demonstrated capabilities including controlling robotic arms and navigating virtual environments with thought.
• What BCI can currently do: high-bandwidth thought-to-computer control (cursor movement, typing, gaming), communication for people who cannot speak, control of external devices (robotic arms, computer interfaces). Current systems require a surgical procedure, a period of learning and calibration, and regular software updates.
• What BCI cannot currently do: restore natural movement to paralyzed limbs (that requires stimulation, not just recording, and is a harder engineering problem), provide sensory feedback from the environment to the brain, or operate reliably 24/7 without occasional recalibration. The current generation is a powerful assistive device, not a full restoration of function.

The Competition: Who Else Is Building Brain Chips

Neuralink gets the headlines, but it is not the only company — or even the most commercially advanced company — in brain-computer interfaces.

• Synchron: the company that beat Neuralink to human trials. Synchron's Stentrode device is implanted through a blood vessel (no open brain surgery), has been commercially available in Australia, and received FDA Breakthrough Device designation in the US. Less capable than Neuralink's intracortical approach but dramatically safer surgery-wise. Synchron has more human implantees than Neuralink.
• Blackrock Neurotech: the longest clinical history in the space, with over 40 human implants using its Utah Array technology over 20 years. Less venture-funded glamour than Neuralink but more clinical experience with chronic implants.
• Precision Neuroscience: founded by former Neuralink engineers. Its Layer 7 Cortical Interface uses a minimally invasive 'brain sticker' approach rather than penetrating the brain tissue. Early human use data is promising for intraoperative (during surgery) applications.
• The academic timeline: most neuroethicists and clinical researchers expect FDA approval for BCI devices in limited paralysis applications within 5 years. Broader applications — cognitive enhancement, memory augmentation, communication for locked-in patients at home — are likely 10-20 years away from widespread clinical availability.

The Questions Nobody Is Asking Loudly Enough

• Data ownership: neural recording data is the most intimate personal data imaginable — it is literally your thoughts. Current Neuralink terms of service give the company significant access to neural data collected through their devices. Who owns the data generated by your brain chip, and what can be done with it, is the most underexamined policy question in the entire BCI space.
• The enhancement vs therapy divide: current BCIs are being developed as medical devices for people with severe disabilities — a use case with clear ethical justification and regulatory oversight. The path from 'restoring communication for locked-in patients' to 'enhancing memory for healthy executives' exists, and the regulatory and social frameworks for managing that transition are not yet developed.
• Long-term safety: no BCI device has a 20-year safety record in humans. Electrode drift, tissue response over decades, software dependency, and the implications of device failure are real unknowns that short-term clinical trials cannot fully characterize.
• The equity question: if BCIs become cognitively enhancing devices in addition to medical ones, unequal access creates a scenario where enhanced and unenhanced humans are in direct economic competition. This is not an imminent concern for current technology but is the policy question that governments need to begin addressing now.