VERTEBRATE.ai

Two years after a surgeon in Phoenix first threaded Neuralink's electrode-laced threads into a human brain, the company says 21 people with paralysis are now using its Telepathy implant to move cursors and robotic arms by thought alone. The company marked the anniversary this week with a flurry of announcements: a $650 million Series E funding round, the launch of its first clinical trial outside the United States — GB-PRIME, running at two NHS hospital systems in Britain — and the hiring of David McMullen, a neuroscientist and former FDA official, to lead medical affairs. Neuralink also open-sourced datarepo, the internal software platform it built to manage the flood of neural recordings, brain scans, and manufacturing logs that come with implanting computers in human skulls.

The most striking detail, almost tossed off in passing, came from Noland Arbaugh, the first person to receive the implant back in January 2024: he described flashes of "anticipatory control," moments where the cursor seems to move before he's consciously willed it to. It's a small, strange sentence, but it gestures at something larger than cursor control — the implant isn't just reading commands, it may be catching intention on its way out the door. For a field that has spent a decade proving it can decode what the brain already decided, that's a meaningful shift in texture, if not yet in substance.

The business mechanics matter too. Five years ago Neuralink was a research curiosity with five patients and a lot of Elon Musk showmanship. Now it has a regulatory hire pulled straight from the agency that will decide its fate, a data infrastructure mature enough to give away, and a British trial running inside a national health system rather than a boutique American hospital. That's the difference between a science project and a device company preparing for scale. The $650 million round, following similarly large raises, suggests investors are betting less on a single breakthrough than on Neuralink becoming the default rail for brain-computer interfaces the way Stripe became the default rail for payments — infrastructure, not novelty.

For the robotics industry, the implications sit at the unglamorous but crucial intersection of actuation and intent. Robotic limbs have gotten dexterous; the bottleneck has always been the interface between a damaged nervous system and a machine built to obey it. If Neuralink's pool of users keeps growing past the two dozen mark, and if the UK trial produces data under a different regulatory and clinical culture than the FDA's, it widens the evidence base for an entire category — exoskeletons, prosthetic hands, assistive wheelchairs — that has mostly been proven in small, isolated academic studies.

The deeper stakes are human, not technical. For a paralyzed person, a thought that moves an arm closer by a fraction of a second, or a cursor that seems to know what's coming, is not a research milestone. It's the first draft of independence. Neuralink is still a company selling a vision as much as a device, and its claims deserve the usual reporter's skepticism. But the gap between "a person moved a robotic arm with their mind" and "a person picked up a coffee cup without thinking about it" is the gap this industry has been trying to close for twenty years, and this week, in increments, it got a little smaller.