Neuralink's First 100 Patients: Paralyzed Users Are Now Browsing the Web With Their Thoughts

The Human Trial Data Is In When Elon Musk unveiled Neuralink in 2016, the concept seemed pulled from science fiction: implantable brain-computer interfaces that would let humans communicate directly with machines. Skeptics dismissed it as Muskian hyperbole—another ambitious timeline that would slip years, if not decades. In 2026, the data is in. Neuralink has now implanted its N1 device in over 100 patients, and the results are reshaping our understanding of what's possible in neurotechnology. From Paralyzed to Participating The first human trial, which began in 2024 with patients suffering from quadriplegia, focused on a straightforward goal: can a brain implant allow a paralyzed person to control a computer cursor with their thoughts? The answer is a definitive yes—and then some. Noland Arbaugh, the first recipient, made headlines in 2024 when videos showed him playing chess, browsing the internet, and even playing Civilization VI using only his thoughts. The N1 device, about the size of a coin, sits flush with the skull and threads ultra-thin electrodes into the motor cortex. These electrodes detect neural activity associated with intended movement and wirelessly transmit those signals to a nearby computer, which translates them into cursor movements. By early 2026, with 100 patients enrolled across multiple sites, the data shows remarkable consistency. All patients have achieved basic cursor control. Most can type at speeds exceeding 10 words per minute—far slower than natural typing but revolutionary for individuals who previously couldn't communicate digitally at all. The fastest users exceed 20 words per minute, approaching the speed of some individuals using eye-tracking systems. Beyond Cursor Control The second phase of the trial, currently underway, moves beyond basic computer interaction to more ambitious applications. Robotic limb control represents the next frontier. Several patients are now equipped with robotic arms that they can control through the implant. While current control is coarse—think "grasp object" rather than "manipulate fingers with precision"—the improvement over current assistive technologies is substantial. Patients can feed themselves, drink from a cup, and perform simple object manipulation independently. Speech decoding is perhaps even more transformative. Neuralink is experimenting with implants placed near speech-related brain regions, attempting to decode intended speech directly from neural activity. Early results show the system can identify which phoneme a patient is attempting to produce, with accuracy around 70% for a limited vocabulary. It's not conversation yet, but it's a proof point that direct neural speech synthesis is possible. The Technical Challenges The trial hasn't been without challenges. Neuralink has faced the same fundamental problem as every brain-computer interface before it: the brain is a dynamic, living organ, and electrodes implanted in it trigger biological responses. In several patients, signal quality degraded over time as the body's immune system formed scar tissue around the electrodes. Neuralink's solution involves both surgical technique—precision insertion that minimizes trauma—and adaptive algorithms that continuously recalibrate to changing signal conditions. Device longevity remains uncertain. The N1 is designed to last years, but the trial hasn't run long enough to determine actual lifespan. Battery life is approximately 24 hours, requiring nightly inductive charging through a specialized pillow or headband. Surgical risk has been minimal but not zero. Two patients experienced minor infections treated with antibiotics. One patient had a small hematoma that resolved without intervention. No serious adverse events have occurred in the first 100 patients. The Unexpected Discoveries Perhaps the most fascinating outcomes from the trial have been unexpected. Several patients report what they describe as "expanded awareness"—a sense that they can perceive their own neural activity in ways they never could before. One patient described being able to "feel when I'm about to have a seizure," despite having no prior history of epilepsy. Researchers speculate that the implant provides biofeedback that allows patients to recognize neural patterns associated with various states. Another unexpected finding: some patients have become significantly faster at learning new motor tasks. By observing their own neural activity during practice, they can identify inefficiencies in their movement planning and consciously adjust. It's like having a window into your own brain's learning process. The Blindsight Connection Neuralink's second major product, called Blindsight, is now entering human trials. The device aims to restore vision to blind individuals by directly stimulating the visual cortex, bypassing damaged eyes or optic nerves. Early results from animal studies show the technology can produce percepts of light—phosphenes—that animals can be trained to recognize as visual information. Whether this can scale to something resembling natural vision remains unknown. The visual cortex is extraordinarily complex, and recreating the rich information stream of sight through electrical stimulation is among the hardest problems in neuroscience. But if Blindsight succeeds even partially—providing enough visual information for navigation and object recognition—it would transform the lives of millions with blindness. The Ethical Framework As Neuralink moves from experimental to therapeutic, ethical questions multiply. Informed consent with severely disabled patients raises complex issues. How do you ensure that individuals with limited communication options aren't consenting under duress? Neuralink has established independent patient advocates who work separately from the research team to ensure participants understand risks and alternatives. Data privacy becomes critical when your thoughts are being recorded. Neuralink's system captures neural activity continuously. Who owns that data? Can it be accessed by insurers, employers, or law enforcement? Neuralink has implemented technical safeguards—all data is encrypted, and patients control access—but legal frameworks lag behind technical capabilities. Enhancement versus therapy will become increasingly contentious. If brain implants can restore lost function, should they also be available to enhance normal function? A future where some people have neural implants and others don't raises profound equity questions. The Road Ahead With 100 patients implanted and data accumulating, Neuralink is preparing for the next phase: commercialization. The company expects to seek FDA approval for the N1 as a medical device for severe paralysis by late 2027. If approved, it would be the first commercially available intracortical brain-computer interface. The broader vision remains characteristically ambitious: Neuralink envisions a future where brain implants are as common as smartphones, enabling direct neural communication, instant access to information, and symbiosis with AI. That future, if it arrives at all, is decades away. But the first 100 patients have proven that the journey has begun.