My first visit to Neuralink came in 2018. Three years earlier, I’d published a bestselling biography of Elon Musk called Elon Musk: Tesla, SpaceX, and the Quest for a Fantastic Future, and we hadn’t talked since. Musk didn’t like some of the things I wrote in the book, and our relationship went from cordial to hostile to nonexistent. That was OK with me. I’d studied Musk and his universe for a very long time, and I could use the break.
But Neuralink, a company Musk founded in 2016 that develops brain-computer interfaces (BCIs), brought me back into his orbit. I’ve spent my career covering businesses trying to build technology at the edge of what’s possible. With its blend of neuroscience, biology, artificial intelligence, and machinery, Neuralink holds everything that fascinates me. So I reached out to Musk and asked to be let inside, and to my surprise, he said I could show up whenever I liked.
I’ve since been to Neuralink dozens of times, occasionally talking to Musk himself, but mostly to the engineers, neuroscientists, and executives building the company. I’ve traveled around the country to meet the recipients of the first wave of brain implants. It’s been eight years of reporting combined with hundreds of hours spent filming a documentary about the company.
My bet back in 2018 was the right one. Neuralink is fascinating and has only become more so as AI technology has accelerated alongside it. I cannot claim to know how fast things will progress with either AI or brain implants, but I do know that we’re on the cusp of something quite new and odd in terms of what it means to be human. We are still in the very early days of an altered world.
ASHLEE VANCE

A 2016 accident left Noland Arbaugh as a quadriplegic, paralyzed from the neck down. In 2024, he became the first patient to receive an implant from Neuralink, Elon Musk’s company that develops brain-computer interfaces.
Come November, the nights start getting cool in Yuma, Arizona, so Noland Arbaugh and his family will sometimes head to their back patio, toss some wood in the firebox, and hang out in the crisp desert air. It’s usually the three of them: Noland; his mother, Mia; and his stepdad, David, along with a menagerie of animals, both loose and caged, that includes dogs, chickens, quail, ducks, and goats. The most attention- seeking of these companions is Hope, a turkey who thinks she’s a dog. Hope cruises around the dirt backyard, then comes over to the family, pressing up against them to elicit pats and kind words.
The scene seems like it could take place anywhere in America, but that’s not the truth. That’s because this prefab house has Arbaugh in it, and he is the closest thing we have to a cyborg. He is the harbinger of a future way of life that can feel glorious, unnerving, or both, depending on how you look at it.
On January 28, 2024, Arbaugh became the first person to receive a brain implant made by Neuralink, the company founded by Elon Musk. At nearby Barrow Neurological Institute in Phoenix, a surgeon removed a circular piece of his skull so that a robot made by Neuralink could sew a series of very fine electrode-laced wire threads into his motor cortex. The electrodes nestle right up against his neurons, which allow them to read the electrical signals produced by his brain activity.

Arbaugh at home in Arizona with his mother, Mia, and his stepfather, David.
In 2016 just after he had graduated from college, Arbaugh had been horsing around with some friends at a lake when he was hit in the head and almost drowned. He was left as a quadriplegic, paralyzed from the neck down and able to communicate only with the aid of a computer, relying on a cumbersome mouth stick to tap a screen and move a cursor. He spent the next eight years in a wheelchair, spinning through cycles of complacency at best and depression at worst about his condition and the burden it placed on his family. They were not well off, and the care and time required to tend to him only made getting by that much harder. “I just draw in resources,” he says, “and give nothing out.”

When a friend told Arbaugh about Neuralink’s plans to start a clinical trial of its brain implant, he was a little hesitant at first. “He called me one day, half drunk basically, in the middle of the day, like noon, and he was like, ‘Hey, Neuralink just opened up human trials. The FDA approved it. You should sign up for this,’” he says. “I was like, ‘What the heck is Neuralink?’”
It wasn’t long before the promise of being able to do more things won him over. A quarter-size device with wires dangling off one end would be surgically inserted into his head, giving him the ability to send his intentions to a computer through a wireless connection. The prospect of having his skull cut open did not particularly faze him. “I was worried that if something went wrong, that it would hurt my mom a lot,” he says. “That was it.”
Without the implant, Arbaugh could not accomplish tasks most of us take for granted, like using his hands to push a cursor around a web browser, sending texts, and shopping. He couldn’t play video games, and, man, does he love to play video games. But with the device, Arbaugh might be able to do all these things again, as if by magic. He’d be able to think where he wanted the cursor to go, and the cursor would move. He could look at a chessboard on his laptop screen, plot his next attack, and use his mind to execute it.
The implant also held the promise of giving Arbaugh renewed purpose. He relished the idea of potentially driving science and technology forward, especially if being experimented on could one day help other people in difficult physical circumstances. As much as anything, though, he loved the idea of becoming a cyborg. He’d be getting Neuralink’s first implant, which, by definition, would mean he’d be getting the worst one the company would ever make. The technology would have major limitations compared to where Musk hoped to take it.
Still, Arbaugh would represent something profound: the start of a new era for humanity. While other people communicate with AI via chat windows and agents, Arbaugh would have it fused right into his brain — part of what’s described in Silicon Valley circles as the Merge, the moment when flesh and machine, human intelligence and artificial intelligence, become one.
Arbaugh and I began talking not long after his procedure, and I went to visit him and his family for the first time in April 2024. He was still buzzing with excitement about the effects of the technology on his life. “For a whole day, I was just giggling to myself,” he says. “I was like, ‘This is so freaking crazy.’”
At this point, I’ve known him longer than any other journalist. I obviously had lots of questions for him, and he had some for me. He wanted to learn how to use his new powers to carve out a role in the outside world. Post-implant, Arbaugh found himself attached to the mania that accompanies Musk in the media. A guy who lived off a dirt road surrounded by farm animals is now a surrogate spokesman for the richest person alive. It’s a most unusual relationship.
Noland Arbaugh
Having watched Arbaugh’s progress over the past couple of years, I feel confident saying that most of his hopes and dreams have come to pass. Though there have been some travails with the implant, it has transformed his life and given him the purpose he sought. Arbaugh is a cyborg and proud of it. He thinks the idea of merging with AI “would be awesome” and is ready for Neuralink to do even more with his body and mind. Per the terms of the clinical trial, his implant will either be removed or deactivated around five years after its insertion, but he says he’ll “upgrade as many times as they let me.”
It is only after I spend an extended period of time with Arbaugh, however, that the truly remarkable thing about the technology inside him becomes apparent. I notice him looking me in the eye and chatting at the same moment he executes a chess move on his laptop. The first time I witnessed him commanding two situations at once with total ease and nonchalance, it felt uncanny, like a superpower. Before long, it was no big deal.
It has taken a decade of engineering to reach a point when yesterday’s science-fiction fantasy somehow seems basic and familiar. As with all things Musk, Neuralink has sparked controversy and discussions of dystopian undercurrents. The reality of the situation, though, is that something like the Merge is, in fact, happening; and it’s going to shock us right up until the point that it feels totally normal.
In the late 1700s, Italian scientist Luigi Galvani was mucking about with the legs of dead frogs (as one does) and noticed that they would sometimes move when touched by a piece of metal. Intrigued by the phenomenon, Galvani began refining his experiments so that he could connect wires directly into a skinned frog’s body and make its legs twitch on command.
Galvani would eventually propose the idea that electricity, rather than “life spirits” or some other unknown force, controlled the excitement and movement of nerves and muscles. While initially met with skepticism, his work helped kick off a new branch of science dedicated to studying the behavior of electricity in bodies.
It wasn’t until the 1920s that researchers began probing the relationship between electricity and the human brain in earnest. In 1924, German psychiatrist Hans Berger placed electrodes on the scalp of a young boy who had a piece of his skull cut out during a tumor-removal procedure. The soft spot on the boy’s head gave the electrodes a clearer path to the brain, allowing Berger to detect some faint signals. The experiment showed that the brain hummed away with electrical impulses, giving us a crude means of measuring electrical activity in the brain.
Neuralink and its competitors — of which there are many — seek not just to monitor neural activity but to interact with the brain. They produce what neuroscientists call a brain-computer interface, a computing device that is usually placed into the brain tissue or right on top of the brain.
The critical objective is to get electrodes as close to neurons as possible, because the closer they are, the clearer the understanding around when and how strongly neurons fire. This activity signals, for example, when Arbaugh is thinking about moving his arms up and down or left and right, even though he can’t actually move them. Researchers want the most accurate representation of how Arbaugh’s neurons fire so that they can best capture what he’s trying to do.
Putting foreign objects in or on the brain, however, comes with consequences. Surgery can damage the tissue, nick blood vessels, or crush neurons. Since the body tends to react poorly to the presence of unfamiliar objects, it will form scar tissue around the electrodes over time. This can muddy the signals the electrodes pick up, which may render the BCI ineffective. The central questions haunting the field: How can you pick up the clearest electrical signals while doing the least damage, and is it possible to create a device that can sit inside someone’s brain for many years?


Scans from October 2024, taken during a routine follow-up, show Arbaugh’s brain approximately nine months after his surgery.
In the quest for answers, scientists have been placing BCIs in primates since the late 1960s and in humans since the late 1990s. For a long time, the most substantial research with the invasive technology revolved around a device known as the Utah Array, which was first implanted into a human — a paralyzed man named Matt Nagle — in 2004. About the size of a pencil eraser, the Utah Array comprises 100 or so rigid silicon spikes (each about a millimeter in height) affixed to a base. Researchers push it into the brain using a firm pneumatic pulse reminiscent of a quick hammer strike.
There’s a degree, then, to which all the attention around Arbaugh is manufactured, since Nagle, for example, could also control a computer’s actions with his thoughts. Beyond that, he could change TV channels and open and close a robotic hand. In the years since, dozens more have performed similar feats with Utah Arrays.
The problem with the device is everything else that comes with it. Patients need to wear clunky computing hardware affixed to their heads that sends their neural data to large banks of signal processing equipment via cables. Experiments with the device often take place in hospitals or research labs for short periods of time, with a team on hand to make sure the hardware works. With its computing cubes and wires attached to a patient’s head, the whole setup throws off A Clockwork Orange vibes, placing the goals of science above those of the person going about their day-to-day life.
This is what sets Arbaugh apart from the cyborgs who came before him. He feels at home living with what amounts to a consumer-electronics device attached to his skull. He can turn the neural computer on with a phone app and use it for many hours at a time by himself without wires connecting his head to large machines. The BCI sits flush with his skull, and his hair has grown over it. Unless he told you he had a brain implant, you would not know it was there.
In October 2016, neuroscientist Dongjin “DJ” Seo received a cryptic email from Musk’s executive assistant saying that Musk wanted to meet with him at 45500 Fremont Boulevard in Fremont, California, the next day. Seo had been doing pioneering neuroscience research as part of his Ph.D. in electrical engineering and computer science at the University of California, Berkeley, when he heard rumors about Musk’s interest in the BCI field. He figured that Musk wanted to talk brains, but did not know exactly why he’d been summoned.
Seo looked up the address and discovered that it belonged to Tesla’s massive electric car factory. He became nervous, feeling like he needed to cram for a test. The evening before the meeting, he read a biography on Musk (OK, it was mine) cover to cover and firmed up how he wanted to describe his neuroscience research. “I had a crash course on Elon,” he says. This led him to the conclusion that Musk “doesn’t tolerate fools, and he really wants to make sure that people know their stuff,” which only amped up the pressure.
Silicon Valley’s interest in brain-computer interfaces had been rising right alongside its interest in artificial intelligence. In 2015, the same year he cofounded OpenAI, Musk met with hundreds of neuroscientists in a bid to ferret out the most promising researchers in the field. Before he became a longevity guru, Bryan Johnson also dug into BCIs; according to him, he and Musk briefly considered joining forces.
Johnson hoped that BCIs might help us finally unlock the inner workings of the brain, leading, in turn, to better mental-health treatments. Musk was not opposed to any of that; but he was mostly fixated on the idea of BCI technology as humanity’s only hope of keeping pace with AI. Better to join forces with the exponentially powerful, alien intelligence than become its supplicant.
Seo fell into a camp of scientists looking for novel ways to build brain-computer interfaces and push the field beyond the Utah Array — and perhaps beyond academia altogether. In 2013, he co-authored a seminal paper titled “Physical Principles for Scalable Neural Recording,” which highlighted the limits of the BCI technology of the time and called for new paths forward. His collaborators on the paper were Adam Marblestone, who later worked on Google DeepMind; Dario Amodei, the future cofounder and CEO of Anthropic; and Mikhail Shapiro, who went on to cofound BCI startup Merge Labs with OpenAI’s Sam Altman.
This cadre of neuroscientists shared the belief that electronics and software had progressed to a point where new types of BCI devices were feasible. At Berkeley, Seo’s Ph.D. work centered on technology described as “neural dust” — tiny sensors that could, in theory, sit next to single neurons and detect when they fired. Others proposed slotting electrodes into the brain’s blood vessels in a fashion similar to heart stents. There’s even far more sci-fi-esque systems like “neural lace” (a term coined by novelist Iain M. Banks), which covers the brain in a network of sensors.
As he waited at the Tesla factory, all these ideas swirled around in Seo’s head. Musk was, as ever, running late. Hours passed. When an assistant finally led Seo into the meeting room, she warned him that there was only time now for a quick hello. Musk, as ever, went off-script. He introduced himself, looked away awkwardly, then revealed the existence of Neuralink to Seo for the first time.

Dongjin “DJ” Seo, cofounder and president of Neuralink.
The two men ended up talking for 90 minutes, the conversation spanning Seo’s research and the state of BCIs, mixing technical questions with philosophical ones. Though AI was still a hazy, obscure concept to almost everybody who wasn’t actively working on it, Musk was already convinced that it would soon advance rapidly and pose a threat to humanity. He pushed Seo to consider what a commercial BCI company might look like if it had, say, $100 million at its disposal. Could researchers really build a new interface between humans and computers?
Seo thought such a thing was possible. “I wasn’t sure what the right format would be. But in talking to Elon and also being in these academic areas of doing this research, I knew that the fastest path and the likelihood of something like this existing was going to be outside of academia,” he says. And that’s when Seo got his first job.
Turning up for work at Neuralink in January 2017, Seo joined seven other people, including Musk, as the official cofounders of the company. Musk acted as CEO while Max Hodak, a promising young scientist and engineer, took on the role of president. The other cofounders brought a mix of neuroscience, robotics, and semiconductor expertise to the table, as well as unique points of view on how to approach the BCI problem.
They set up shop in the three-story Pioneer Building in San Francisco’s Mission District. The wooden structure, which dates to 1902, was once a factory that made trunks for cars (back when they were add-ons rather than built-in). Over the years, the building has been home to several technology companies; in 2017, it became the headquarters of both OpenAI (which Musk also cofounded) and Neuralink.
The arrangement wasn’t without its peculiarities. Even though Neuralink had a smaller team, the company needed much of the space to build out early hardware prototypes and run experiments, including a machine shop, biology lab, and surgical suite. OpenAI, which then had around 30 employees, did most of its work on laptops. It also included a small robotics lab where researchers focused on getting a machine to solve a Rubik’s Cube.
Despite the fact that both were Musk joints, the two companies maintained a surprising degree of secrecy between them. An OpenAI employee walking past a barrel marked “hazardous waste” could not help but wonder what his neighbors were up to. There was, of course, plenty of secrecy vis-à-vis the rest of the world, too. Virtually nobody outside the Pioneer Building had any idea what was going on there.

Seo with Neuralink's surgical robot.
There were, however, moments of overlap between Neuralink and OpenAI facilitated by the building’s layout. Researchers from both companies often met up in a shared lunch area on the first floor. Seo and Hodak made regular appearances, as did OpenAI cofounders Ilya Sutskever and Greg Brockman. Now and again, Musk and Altman would pop in. These lunchtime chats focused on the commonalities between Neuralink and OpenAI’s quests to understand intelligence in all its forms, and the ways in which the neural networks behind AI technology resembled the neural and synaptic networks in the human brain.
“Having lunch together was quite exciting and fun,” Seo says. “The success of each other could help propel each other’s field forward. But that was still early days, and we weren’t entirely sure if some of these marriages would happen or not.”
Neuralink’s founders were beyond ambitious in what they hoped to build: a brain implant that everyday people would choose to have inserted into their skulls. The product would almost certainly start out as an aid to paraplegics or people suffering from conditions like amyotrophic lateral sclerosis (ALS) and needed help with movement and speech. But the company also had the explicit goal of making a mind computer for the masses. Musk envisioned walk-in clinics where robots perform quick surgeries and people leave with astonishing new powers to go about their everyday lives.
This might make you wonder: What are they so sure people will actually want that? The answer is both extraordinary and simple.
Some of the Neuralink cofounders, including Hodak and Seo, grew up consuming science-fiction books and watching films like The Matrix. They tended to ignore the dystopian tone of these works, instead fixating on the good shit. Like Neo in The Matrix, they wanted to quite literally be able to download the ability to do kung fu. They fell in love with the idea of a two-way information exchange between humans and computers. A brain implant that gives people a chance to develop abilities akin to superpowers, allowing them to consume vast quantities of information and control machines, such as a phone or a robotic limb, via thought commands.
For Seo, who is now president of Neuralink and runs the company day-to-day, these desires were rooted in personal experience. When he was 4, his father, a mechanical engineer, moved the family from Baton Rouge, Louisiana, to South Korea to pursue academic work. When Seo returned to the U.S. in 2001 at the age of 13, he did not speak English. Feeling isolated and struggling to connect with his peers, he wanted nothing more than to be able to learn the language in an instant.
DJ Seo
He’d also watched both sets of his grandparents succumb to dementia or Alzheimer’s disease. “It was very tough seeing their memories, identity, and their entire mind just slowly degrading over time,” Seo says.
As he progressed through graduate school and dove deeper into brain-computer interfaces, these experiences returned to him and began to crystallize his purpose: to figure out how the brain works.
“My introduction to the brain and neuroscience was really from an engineering perspective of: Wait, how do we not understand this system that we’ve carried around for millennia? How can we actually start to build a set of tools to understand this, and what does that process look like?” he says. “I think everything kind of came together as: Wow. This is the core mission of what I want to dedicate my life to.”
In July 2019, Neuralink held its coming-out party at the California Academy of Sciences in San Francisco. Many of the world’s top neuroscientists had been invited, and they mingled in a glass-roofed atrium in front of an enclosed rainforest full of spectacular insects and exotic flora. Leading up to the event, there had been plenty of buzz about Neuralink among researchers and the press, but few outside of the company knew exactly what Musk and his team had been up to. The carefully curated set of attendees felt lucky to be there, downing cocktails and snacks as they prepared to be freaked out.
When the event began, Musk took the stage dressed in black jeans and a black sportcoat. He walked the audience through a brief history of brain technology, touching on the cochlear implant (1957), the Utah Array (1991), and deep-brain stimulation for Parkinson’s disease (1997).
Then he came to the heart of Neuralink’s plot. The company had raised $150 million (about $100 million of which came from Musk) to push the BCI field dramatically forward. Its brain implants aren’t just to treat illnesses; they’re the next great turn in technology, intended for everyone.
“This is going to sound pretty weird, but ultimately, we will achieve symbiosis with artificial intelligence,” Musk said at the event. “This is not a mandatory thing; it is a thing you can choose to have if you want. This is something that I think will be really important on a civilization-level scale.”
Neuralink, he said, had built a surgical robot that could place electrode-dense wires into a brain and an early version of an implant that could process neural signals faster than anything that came before. Beyond that, researchers had already completed 19 surgeries on animals, including mice and primates — something Neuralink had planned to keep a secret for a while longer, since animal testing often spurs controversy and negative media coverage. Musk, though, could not stop himself from disclosing the news that Neuralink had already outfitted a primate with a functioning BCI.
“We definitely need to address the monkey in the room,” he said in his signature glib style. “This is a sensitive subject: A monkey has been able to control a computer with its brain.” As the words flew from Musk’s mouth, Seo, Hodak, and some of the other Neuralink cofounders onstage looked on in semi-shock. The audience got all the freak show it could handle.
The company was only a couple of years old but moving at breakneck speed for the biotech field. Neuralink left the Pioneer Building and took over a large industrial office-cum-laboratory in Fremont, not far from the Tesla factory where Seo had met Musk for the first time. The facility included an area for Neuralink to build its surgical robots, multiple rooms for housing animals, laboratories, and a machine shop. There was even a clean room where employees created custom computing parts. About 100 people were now working at the company.
At the July event, Neuralink declared its intention to begin human BCI trials, possibly as early as 2020. To do so, the company had to gain approval from the U.S. Food and Drug Administration. As all of Musk’s timelines do, this one felt aggressive to some and implausible to most.
The traditional way of getting a new medical device approved by the FDA is to settle on a fixed design, then test it over and over until regulators feel confident that the product will do no harm — or at least no harm above an acceptable level. The problem in this case is that, at a Musk company, nothing stays fixed for very long.
To begin human BCI trials, Neuralink had to get two products — the robot and the implant — blessed by the FDA at the same time. Both pieces of hardware were constantly being updated with new materials and technology. Much as he did with SpaceX, Musk challenged the FDA to keep up with him.
Neuralink dedicated a large team to dealing with the FDA, sometimes describing the effort as “going to war.” Government regulators had never seen a device with fine threads combined with a wireless plastic implant that was surgically inserted by a robot. As such, the agency did not really have criteria at the ready by which to judge the technology, putting the FDA under as much pressure as Neuralink. Should something go terribly wrong with one of the early devices, the public would turn on Musk and the government at the same time.
Over the course of many months, the FDA asked Neuralink to do things such as short-circuiting the battery and delivering physical blows to the implant to prove that its casing would remain sealed under violent conditions.
When it came to other tests for chemical leeching, the FDA kept moving the goalpost by pushing the detectable limit threshold ever lower. Neuralink’s initial submission to the agency included 15,000 pages of engineering documentation and test results. And each time the company made a change to its hardware or software, employees had to update the filings with updated supplemental information. It took until May 2023 for the FDA to grant Neuralink an Investigational Device Exemption (IDE), and set the company off on its path to conduct clinical trials.
Over the next five years, Neuralink followed a familiar pattern. The team held periodic events at its Fremont headquarters to reveal new developments — invite-only affairs for people affiliated with the company, including investors, prominent scientists, and friends of employees. The public could watch the proceedings via video feed, but that was not much of a consideration for Musk. He viewed these sessions mostly as a recruiting tool for promising engineers; he needs all the talent he can get.
For the most part, Neuralink kept to itself, choosing not to interact with the media. As far as I’m aware, I’m the only reporter to have visited the company during this period.
At one session, Neuralink famously showed a video of a 9-year-old macaque named Pager playing a version of the video game Pong with its mind. At other events, Musk talked about the company’s new efforts to cure paralysis and blindness. He also gave updates on the ever-increasing computational horsepower of the implant, as well as new skills its surgical robot had attained.
Though many on the internet reacted to these revelations with awe, some members of the neuroscience community were dismissive. They saw Neuralink as recycling old research breakthroughs and using the world’s infatuation with Musk as a means to overhype the company’s accomplishments.
“Neuralink hasn’t done anything that I consider innovative at all,” Miguel Nicolelis, a neuroscience professor at Duke University, told Inverse in 2021. “I just find it a little offensive that these tech guys who behave like gods come out and say, ‘Oh we are going to do much better.’” He portrayed the company as callous and inhumane: “What I see is most of these guys, these techie guys, going there and talking about technology like there is no human being behind what is going to be done.”
As Musk became ever richer and drifted deeper into politics, the media began to view Neuralink as a fruitful venue for critique and controversy. Musk has a well-earned reputation for pushing his companies and employees hard. Reporters wondered aloud whether taking this approach to biotech was a recipe for disaster. It’s fine to blow up rockets as part of an ongoing development effort; no one, however, would tolerate brain implants going haywire and hurting people.
Reuters and Wired published ongoing stories about Neuralink’s animal trials. Some of these pieces presented the case that the company had been too aggressive with its surgeries on primates, causing undue suffering. Neuralink issued detailed rebuttals that did little to salvage its reputation. It was easy to conclude that, yes, the company was chasing progress at any cost.

The finished implant sits flush in the skull, here on a practice model.
Blowback from academics is understandable — Neuralink did repeat some past work and has been known to hype its accomplishments — but critics may be too close to the subject to appreciate the galvanizing effects of Neuralink’s work. Before Musk came along, the BCI field had stalled for years with relatively little innovation around the Utah Array. What Neuralink did was shift BCI research from academia to the business world. Musk’s presence in the field created a brain-computer interface boom. Dozens of BCI startups appeared, buoyed by billions of dollars in venture capital. A race is afoot. Competing in it requires the usual Silicon Valley machinations.
More to the point, these academics seem to undervalue the engineering and technological efforts required to make the surgical robot and implant. From day one, Musk urged his staff to build both products for widespread general-purpose use. He didn’t want to make a rare and prohibitively expensive device; his goal is a consumer product. Neuralink’s engineers can recount numerous stories of Musk pushing to make the implant cheaper, smaller, and faster. Of all the people in the world, he is in the unique position of having built complex products across different industries; he also boasts the pattern-matching skills to know what is needed to produce large amounts of implants affordably.
“There’s making a wire that you stick in the brain for a day, and then [there’s] making a wire that you stick in the brain for 10 years, and it’s going to work and not shock someone and not break down or be toxic. That’s an infinitely harder problem,” says Sumner Norman, who cofounded a pair of BCI startups, including Merge Labs. What Neuralink did, he says, was shrink millions of dollars of computing equipment down into something that can fit neatly into the skull — a breakthrough that academic critiques mistook as a fait accompli.
As for the animals, well, there is nothing pleasant about animal testing. Though the purpose is noble — to avoid human suffering — the process is imperfect and sometimes very ugly.
Neuralink runs its experiments on rodents, pigs, sheep, and primates under the supervision of an experienced animal-husbandry team. I’ve visited them several times over the years, and the primates have toys and other diversions in their enclosures, including videos and music, and they can pick items like fruit smoothies from a food menu, even doing so with their implants at times. I’ve also been to many other university and corporate animal labs and will note that Neuralink’s facilities are far larger and the animals better cared for than at any other lab I’ve seen.
The contemporary version of Neuralink has come to life on the southeastern outskirts of Austin, Texas, at a place known as River Bottoms Ranch, which is much as it sounds: 100 acres of rural land that abuts the Colorado River. Pecan and mesquite trees are scattered around the edges of the grassy grounds, where longhorn cattle have been known to graze. When Neuralink first moved in and put up a barn to house its testing animals, neighbors shared rumors that mysterious Iberian pig farmers had come to town.
Neuralink’s presence and objectives became obvious rather quickly. As so often happens in Musk Land, employees reshaped the site in a matter of months. Buildings went up, fiber-optic cables were laid, and power and water were brought in on an industrial scale.
The company, which has now raised a reported $1.3 billion, is creating a grandiose BCI production and testing facility. There’s a building dedicated to manufacturing implants, barns, land for sheep and pig test subjects, operating rooms filled with surgical robots, a pathology lab for analyzing brain samples, a materials lab, and office space for hundreds of workers. There’s also a neon-lit bar with old-fashioned arcade games, lest any prospective employees worry that working there is no fun.
Just as he did with SpaceX and Tesla, Musk set up the ranch so that Neuralink can accomplish as much as possible on its own. For example, most companies in the BCI space send brain samples to outside pathology labs. Neuralink, by contrast, performs surgeries on-site, gathers samples, then walks them over to a lab designed by one of the world’s top pathologists to speed up sample prep and analysis. This setup saves the company months of time by allowing it to make tweaks to its underlying technology, gather data on the results quickly, and show the findings to the FDA. The faster Neuralink can prove that a new hardware design is safe, the faster it can move forward with better technology.

River Bottoms Ranch, Neuralink's campus outside Austin, Texas.
The site also includes an imaging center with top-of-the-line computed tomography and magnetic resonance imaging machines, the hardware that makes it possible to peer into an animal or human skull and see exactly how well an implant has gone in. When needed, they can also be used to image the implant itself to see how well it’s been put together.
As a perk, employees can sign up to use the machines when they’re free. “I’ve seen my brain,” says Scott Padilla, a project manager who oversaw the construction of the ranch. “No tumors yet found, so that’s kind of nice.”
Much of River Bottoms Ranch is dedicated to solving the central challenges that come with BCI technology: making an implant that does little damage to the brain and can last inside someone’s head for a long time.
To explain how all of this works in practice, let’s start with the surgical robot. The machine stands almost 8 feet tall and weighs 1 ton. It’s mostly all-white and looks like something Apple would design if it were in the surgery game. The bottom half of the robot’s body is made up of a cube-shape base that holds motors and electronics; it also gives the contraption heft so that it’s sturdy when placed in an operating room. The upper half looks more exotic, almost like a manufacturing robot, with moving arms that can center the robot over a patient’s head. The business end of the hulking machine, which looks like a Stormtrooper helmet, contains five camera systems and a specialized imaging device that can see not just the brain, but the vasculature beneath its surface.
All this hardware is required to perform the most delicate of tasks. The robot holds a needle made of tungsten alloy wire that’s a fraction of the thickness of a human hair. It’s designed with a tiny, laser-milled groove that slots into an equally tiny loop at the end of one of the implant’s 128 threads, each containing eight electrodes. The robot uses its cameras to look at the brain and pick the right spot and time to grab a wire and sew it into the patient. The threads are flexible, thin-film ribbons that move as the brain pulses. They’re made from sandwiched layers of conducting and insulating materials, gold and polymers, respectively, to provide both function and protection.
“I’ve seen my brain. No tumors yet found, so that’s kind of nice.”
Scott Padilla
The threads must be sufficiently miniscule to dodge blood vessels and neurons and be considered as inert objects by the body, while retaining enough structure that the robot can push them through tissue. It’s a near-impossible balance that Neuralink has spent years trying to perfect. Recently, the company disclosed a new breakthrough: Its robot-and-thread combo has gotten good enough to pass not just through brain tissue but also the dura, a protective membrane on top of the brain. This means a simpler, safer, quicker procedure.
Once all the threads are in place, the attached circular implant slots into the hole that had been made in the patient’s skull. This quarter-size device contains chips that amplify the signals gathered from the neurons, a Bluetooth radio to send out data, and a battery. Many of these components are custom-built to reduce heat and power consumption. The whole nesting doll of electronics then gets hermetically sealed with a special plastic.
There are dozens of workstations on the implant manufacturing line. Up until recently, it took workers 5.6 hours to make a single device. Engineer Ankita Urala Van Der Werff can recite the precise time off the top of her head because Neuralink has cameras installed at each workstation that monitor the building process. She studies the images and the data logs, anguishing over every minute. Once an implant is completed, it goes through hundreds of inspection checks.
“Each implant has been thoroughly examined from every single angle possible,” she says, “and the entire team is working really hard to make sure that it is safe and functional to be handed off to a patient.”
Once the device is inside a human, it recharges wirelessly via another device Neuralink produces. Build engineer Izumi Kawahara handmakes soft textile cases that go around the charging unit to reduce any discomfort a patient might feel right after surgery, when their skin is extra-sensitive. Before joining Neuralink, Kawahara worked for a Formula 1 racing team and Apple. She’s developed a reputation for solving the unique problems the company faces in creative ways. As one of her prototyping coworkers puts it, “She makes us look like cavemen.”
Until recently, Neuralink had been building the implants in an old axe-throwing bar that it converted into a factory and machine shop, producing 40 implants a day. As interest from clinical-trial applicants grows, the company is in the process of opening multiple manufacturing lines at River Bottoms Ranch to boost production to thousands per month.
“There are millions of patients that could benefit from this,” Seo says. “Their time’s running out, and we feel deeply, deeply responsible to move as quickly as we can.”
As he takes me on a tour of the property, Padilla tells me that Neuralink executives have a future vision of the ranch as “a meeting of technology and nature.” The campus will grow to include more futuristic buildings while preserving its pastoral ambience.
According to the plan, the setting is designed to help engineers think big thoughts and give implant recipients a place to relax. “When you bring a patient out to the site to see what we’re about, it’s cool to go to a gargantuan building,” Padilla says, “but I think it’s also nice to feel like, ‘Hey, I’m in a place where I can heal and become better.’ What is it Elon called it one time? ‘A hippie commune.’ That’s what we’re going for.”
Musk’s security detail turned up midafternoon, and the head of the operation ordered everyone inside Neuralink’s Fremont office to leave the building. Dozens of employees got up from their desks and walked outside. They hung out in a large group for about half an hour while a handful of former military and FBI personnel searched the office for anything suspicious, including bombs. Eventually, the “all clear” arrived, and everyone received a security wanding as they piled back into the office and returned to work.
This was the middle of June 2025. Musk was in town for one of Neuralink’s update events. Employees had spent weeks preparing for the occasion, which would consist of them detailing the latest and greatest elements of their work in front of Musk, Seo, and video cameras. Many had been pulling all-nighters in the immediate run-up as they practiced and polished their presentations. Seo and Shivon Zilis, a longtime Neuralink executive who shares four children with Musk, were grinding, too, reviewing every word and image the Neuralink engineers had prepared.
The event was meant to start around 6 p.m., but Musk was running late, of course, and didn’t arrive until about 8 p.m. He came in through a back entrance, accompanied by security guards and several of his children. He then ducked into a prototyping area, where Seo briefed him on the plan for the night. The lighting set up by the camera crew accidentally caught Musk at just the right angle to project his towering shadow onto one of the main walls. Musk, rather comically, loomed over the proceedings.
The event kicked off around 9 p.m. with a video of Neuralink patients using their devices and talking about how the technology had changed their lives. The company has only implanted about 30 people but has pending applications from 40,000 more. One man used his implant to control robot arms so he could work on engineering projects; others who had lost their voices could now speak again. The presentations that came next went on for three hours as one employee after another did their best to nerd-peacock for Seo and Musk.
It was past midnight when the event ended and Musk got up and made his way to the back exit. At some point during the evening, he received word that one of SpaceX’s giant Starship rockets had exploded during a test. He already looked tired, and this news did not brighten his mood.
Right before joining the convoy of exiting Teslas, Musk took Seo aside and told him that the company needed to do better. People were using too much jargon to describe their work, and Neuralink needed more, deeper data. After this brief, pointed critique, he drove off to the Palo Alto, California, headquarters of his startup SpaceXAI, where he held meetings throughout the night and into the early morning.

One of the barns at River Bottoms Ranch.
The experience of being chided by his boss left Seo looking deflated and exhausted, moving through the afterparty in a semi-catatonic state. As people ate and drank all around him, I could see him processing Musk’s comments and rolling them over in his mind as he forced a couple of smiles and congratulated people on their efforts.
One of Musk’s gifts is instilling a near-religious devotion in his employees, tied to the idea that his businesses have grand missions. Tesla is meant to save our planet; SpaceX will create a colony on Mars, a backup plan for humanity in case Earth can’t be saved. His legion of detractors dismiss all this as hucksterism, part of the relentless Musk Show. He dangles a fantastic future in front of people long enough to get them to do what he wants, secure more money for the ventures, and keep his empire empiring. The multitudes of people who work for him, however, tend to be believers, genuinely committed to what they see as the greater good.
Seo is certainly a believer. He and Musk are the last of the eight Neuralink cofounders who are still with the company. Hodak stepped down as president in 2021 after clashing with Musk, and the others departed for a variety of reasons. Seo has been the constant. He has the depth of technical knowledge to oversee Neuralink’s engineering efforts, the force of will to push the company through ordeals like the FDA approval process, and the calm demeanor to offset at least some of the tension that surrounds an ambitious company governed by the world’s most ambitious human.
For those who detest Musk and think Neuralink is just one more of his evil plans, Seo is a confounding figure. He is polite, unassuming and — there’s no better way to put it — nice. He’s either a fantastic actor, or he truly cares about the people Neuralink is trying to help.
Over the years, Seo has found himself struggling with the constant tension between his reserved nature and all the baggage that accompanies working at Elon Musk’s Brain-Control Company. There’s the internal pressure to perform and do better, and the external pressure that comes with an awful lot of people assuming the very worst about Neuralink’s motivations. Seo has watched members of the animal husbandry team break down in tears after receiving death threats, and he’s been forced to deal with the repercussions of Musk’s political outbursts.
If you catch him in a quiet moment, Seo will confess that the drama gets to him at times. “Attention goes both ways,” he says. “There’s positive attention, but there’s also an enormous amount of attention on every misstep that we may take.” Mostly, though, he’s a workaholic — a former academic who now talks about needing to move with “a maniacal sense of urgency.” This is no doubt in part to meet Musk’s demands, but it’s also an attempt to help as many people as possible as quickly as possible.

About a month after receiving his implant, Arbaugh noticed that his newfound superpowers seemed to be fading. Previously, he’d been breaking just about every BCI record and moving his cursor around his laptop screen at incredible speed. Suddenly, though, it didn’t want to respond to his commands. “One day, it just got much worse,” he says. “And every day after that, it got worse and worse and worse.”
Arbaugh’s doctors and Neuralink’s engineers quickly figured out what happened: A large portion of the threads implanted in his brain had retracted from his neurons. Our brains shift around in our skulls more than animal brains had during the company’s years of testing. Humans are just built differently.
After Seo broke the news to him, Arbaugh had a big cry. “I was thinking that I was given this amazing technology, this ability to interact with the world in a much more impactful way,” he says. “And it was all going to be taken away from me after a month. I thought it was incredibly unfair.”
But Arbaugh has deep reserves of faith. After praying about his situation, he resolved to make the best of things. He is part of a clinical trial for this very reason — to test the limits of the technology and give Neuralink a chance to fix them.
“I was thinking that I was given this amazing technology, this ability to interact with the world in a much more impactful way. And it was all going to be taken away from me after a month. I thought it was incredibly unfair.”
Noland Arbaugh
The company’s field team swoops in to work with each patient and help them learn how to use their implant. In the days after the robot surgery, engineers arrive at a patient’s house and spend morning to night by their side. In Arbaugh’s case, they asked him to perform tasks like thinking about moving his hands, then looked at the data the implant picked up.
The engineers then put their software to work building a model of the unique way Arbaugh’s neurons fire in order to find patterns that will translate into smooth actions on the computer.
Arbaugh emerged from the incident with his optimism undimmed. “I knew that I was going to get the surgery and I knew that it was going to work out,” he says, “and I knew that it was going to be at least the start of what God wanted me to do in my life.”
Other biotech companies do some version of this, but the manpower Neuralink deploys for each patient is unusual. Engineers work long hours not just to get the implant cranking but to continually improve its performance. They write updates to the code overnight, come in the next day and test them, then repeat the process. It was through this type of effort that the Neuralink engineers were able to get Arbaugh’s implant working properly again. The man in Yuma has his superpowers back.
One of the toughest challenges the field team faced involved Brad Smith, the third patient to receive a Neuralink implant and the first with ALS. Smith started to feel the early effects of the disease eight years ago at age 37, and his situation has degraded along a predictable path. He lost control of his body and voice, and now spends his days in a wheelchair and requires a ventilator to breathe. To communicate, Smith used to rely on eye-tracking technology that enabled him to pick out letters and phrases on a computer screen mounted to his wheelchair.
While helpful, the device was slow and only worked if Smith could control the levels of light in his surroundings. Before developing ALS, he could type around 100 words per minute; with eye tracking, he was lucky to eke out five. People would talk over him as he struggled to keep up with the flow of conservation. The limitations of the technology also meant that Smith could not really leave his house. If he went outside or into a brightly lit room, his pupils constricted and the device would fail.
Neuralink installed Smith’s implant in late 2024. He received pep talks from Arbaugh and Musk and a home visit from Seo before the procedure. Then he began to work with the engineers who became part of his family in the following weeks. He was beyond excited to use his new BCI, but rewiring his brain proved challenging because he’d become so used to both the eye-tracking technology and to the limitations ALS imposed on his body.
Bit by bit, the Neuralink engineers helped Smith form a symbiosis with his implant. Instead of telling him to focus on hand movements, they asked him to think about moving his tongue up, down, left, and right. The engineers — including Sammy Nio, Nir Even-Chen, and Sehej Bindra — sat by his side, encouraging him to try new methods until they discovered which ones worked.
Soon enough, Smith got the hang of things. He started using his BCI not just to communicate faster, but to play video games with his children. More importantly, he could now go with them to the park and attend their school events. For the first time in years, he left his Arizona home and went on a road trip with his family to Utah.
Smith felt liberated. Like Arbaugh, he found renewed purpose via the implant and felt that he could help others by providing feedback to Neuralink. “Suddenly, the world started listening to our thoughts again,” he says.
For the past few years, I have been visiting the most adventurous biotech labs the planet has to offer, and I can tell you that Neuralink is not alone in promising a future that will scare some and overjoy others. Much of the public’s attention has been drawn to the promise and perils of artificial intelligence and the silicon that fuels it. AI is presented as the dominant technological force we should be concerned with. I would argue, however, that our flesh, too, may be on the verge of unprecedented change, and what’s happening in biology could come to dominate the headlines.
There is, for example, a laboratory in downtown San Francisco that belongs to a startup called Becoming. In recent months, the lab’s contraptions managed to grow a mouse embryo inside an artificial placenta. The company has an incredible amount of work ahead, but this breakthrough points toward a synthetic womb that may allow humans to give birth without carrying children to term inside their bodies.
Another startup, Kind Biotechnology, is editing the genes of mice embryos to stop them from developing limbs, a central nervous system, or a brain. What’s left after the embryo matures is a sack of organs. Kind’s plan is to grow organs inside the wombs of pigs and other animals in order to harvest them for use in humans. You might think this sounds ghoulish. But many people felt the same way about the first heart transplants, and now we can’t get enough organs to satisfy the demand.
Or take Bexorg, a startup based in New Haven, Connecticut, which has developed a machine that can keep a human brain — harvested from recently deceased donors — functioning outside of the body for several days. Fed with a proprietary blend of nutrients, the brains, fully disembodied and sitting in a vat, can perform many of their usual biological functions. The startup has been using the technology to test medicines for diseases we cannot currently treat well or at all, like Parkinson’s and Alzheimer’s.
All of these companies are piggybacking off tremendous advances in gene-editing technology and in the machinery and sensors needed to make their experiments work. Here again, sci-fi concepts that once seemed laughably far-fetched are tantalizingly possible.
The neuroscience field remains divided on whether Neuralink’s approach to BCIs really makes sense. The company’s surgical method stands out as the most invasive and aggressive in the field. Companies like Precision Neuroscience and Phantom Neuro have found ways to help people with surgeries that do not require entering the brain tissue at all. Their argument is that we can get a clear enough signal by being near either neurons in the brain or motor neurons in the body to match the capabilities of Neuralink’s device.
There are also questions around Neuralink’s business model. Will enough people suffering from paralysis, ALS, or other conditions choose to get implants? Will insurance companies cover the costs? Phantom Neuro founder and CEO Connor Glass argues that Neuralink’s potential market is relatively small, making it hard to justify the level of investment.

A model skull shows the exact placement of the implant in the brain.
“Brain-computer interface is almost entirely starting with a spinal-cord injury or people with locked-in syndrome,” Glass says. “There are north of 100,000 individuals with spinal-cord injuries in the U.S. If you are not Elon Musk and you’re talking to investors, they’re going to tell you that is way too small of a market. The math doesn’t math.”
The main issue is the gulf between where Neuralink is now and where it wants to go. The company’s primary goal was never to help people with injuries or disabilities, but to accelerate the merging of man and machine. It built surgical robots because there are not enough human brain surgeons to perform the millions — or billions — of procedures we’ll see in the not-so-distant future. Deep implants are necessary, the company believes, to unlock the full potential of BCIs. Rival implants might accomplish similar gains for injured people, but Neuralink does not consider that the final destination.
Still, even that goal is murky. To help people recover control of their bodies, Neuralink needs to put an implant into the motor cortex. Recovering speech requires going into a different part of the brain. What type of implants would be required to endow all of us with a broad set of new abilities and what exactly would we use them for?
It’s easy, for example, to imagine a future where people message their friends or order an Uber via their thoughts, or even look up someone’s name while shaking their hand. But we currently have little idea how a person would download information from a computer into their brain or have any meaningful back-and-forth with an artificial intelligence.
“If Neuralink ultimately achieves its goal, I think I will never die.”
DJ Seo
Silicon Valley, per usual, has little interest in what’s not possible today. Neuralink is one of dozens of companies pushing at the frontier of how the brain functions. Go to a dinner party in San Francisco and you’ll hear people talking in very serious tones about downloading their minds, with their associated personalities, knowledge and memories, into computers. Immortality, to them, is not achieved by following stuff like Bryan Johnson’s obsessive longevity regime but through the intertwining of flesh, silicon, bits, and neurons.
Seo is very clear on how he feels about this subject. “If Neuralink ultimately achieves its goal,” he says, “I think I will never die.”
In July, Arbaugh (known as Patient 1, or P1) visited the home of Brad Smith (P3) for a Neuralnaut party, the first ever in- person gathering of modern cyborgs. P2 Alex Conley, and P10 Brandon Keena were also in attendance. It took months of scheduling to pull the event off, with the patients navigating their caregivers’ schedules and various clinic visits, plus the logistics of getting folks in wheelchairs to Gilbert, Arizona.
The Smith household tends to be quite chaotic. There’s Brad, his wife, Tiffany, and their three children, including the youngest, Abigail, who has Down syndrome. This party, though, was chaos-plus. The cyborgs all came with a variety of machinery, family members, and lots of things to say about their various conditions. Introductions included tales of drunk drivers and accidental blows to the head that had left them injured — and, of course, plenty of stories about what it’s like to live with a brain implant.
As the only one among the group with ALS, Smith had the most trouble speaking and had prepared some remarks. He’s been able to feed old recordings of his voice into an AI model, so he can now speak with his pre-ALS voice instead of in a Stephen Hawking-like computerized tone.
“We’ve been in the same Slack for months creating wins, failures, and memes. But nothing replaces looking someone in the eye who has the same scar on their head and the same quiet miracle happening inside their skull,” he said. “It means I’m not alone in this.”

The Neuralnaut network has, in fact, been expanding rapidly of late. Neuralink has increased the rate of surgeries and extended its clinical trials to Canada and the United Kingdom. The OG cyborgs often hold video chats with new participants, coaching them on how to adjust to the technology.
All of them wonder what the future may bring, especially the time limit on their implants. For the moment, they’re bound to the terms of the clinical trials, which don’t include the option for upgrades. Unless Neuralink comes up with a commercial device, these early patients will have to return to their pre-cyborg lives. Arbaugh can feel the clock ticking. He would like nothing more than to keep pushing forward.
“Maybe they’ll get something through the FDA that says they can do more — they can extend the study, add a different sort of avenue for the study, whatever it is,” he told the group. “We’ll see.”
An owner of Bedford Media, the publisher of LIFE, is also the founder of Thrive Capital. Thrive Capital is an investor in Neuralink. Thrive Capital has no editorial control over content featured in LIFE. In an effort to promote transparency, any potential or actual conflicts are timely disclosed.
Production
Photographs in Austin by Eli Durst
Photographs in Yuma, Arizona, by Rosie Marks












































