For years — decades, even — news accounts and scientific journals have featured videos of human beings controlling computers with their minds. A person, often with some kind of paralysis, from either injury or disease, gets a computer chip embedded in their head — tiny, needle-like electrodes plunged into the gloppy folds and wrinkles of the brain. Those electrodes sense the activity in nearby nerve cells — communication among the tens of billions of neurons in the brain, all networked together. Given enough time and the right software, a computer can learn to equate specific patterns of neuroelectric zaps in the person’s brain with a designated action, and then perform the requested function: move a robot arm, point a computer cursor, say the word “yes.”
Scientists came up with the idea of implantable brain chips in the late 1960s; the first experiments in humans began in the early 2000s. That’s when I started reporting on them, and like any good nerd I was compelled by what scientists could learn with these “brain computer interfaces.” BCIs seemed to promise a bionic, Matrix-like future of mind-controlled giant robots and back-of-the-skull plugs for surfing cyberspace. But despite decades of research and development, no implant has moved past the lab. I figured brain chips were equal parts science and science fiction. They weren’t a workaday technology capable of treating actual illness or disability — and they certainly weren’t a business.
I’ve changed my mind. I now believe BCIs are about to cross what tech folks call the Valley of Death, the chasm between research and commercialization. That’s in part because the science has improved. But even more important, it’s because venture capital has finally come up with a potentially lucrative market for brain chips — and, in typical VC fashion, a way to unlock even more value from the user data the chips will generate. Federal regulators are also on board: The Food and Drug Administration has green-lit more human trials of BCIs, and made clear what the agency will expect of companies asking for full-on approval. Implants are no longer some futuristic tease, like jet packs or flying cars. Computer chips are coming to a brain near you.
A direct interface between our brains and our computers has the potential to dramatically reshape the digital landscape. If it works, it will signal a change as profound as the arrival of PCs and smartphones. The companies that make these things are now competing, quite seriously, to be the Apple or Microsoft of neurotechnology, the one that turns an early market foothold into global dominance. The race for implantable brain chips has been a long, deliberate marathon. Now it’s entering the final, heated stretch.
There are a bunch of technical approaches to creating a computer interface with the brain. You can use arrays of microelectrodes, or skullcaps that pick up EEGs, or devices that read from the peripheral nervous system rather than from the brain itself. But if you’ve heard about BCIs in the past few years, it was probably something about Neuralink, the company founded by Elon Musk. Three years ago, Musk boasted that his implants might someday help people mind-control a computer into communicating at 40 words a minute, control virtual avatars, and endow them with “really high-bandwidth telepathy.” It would even keep artificially intelligent computers from taking over the world by letting us “merge with AI” so we can “go along for the ride.”
Um, no. Here’s what’s actually likely to happen: At first, people with severe paralysis or missing limbs will get BCIs to help them move through the world with more autonomy. Brain chips will enable them to perform simple actions on their own and reduce the need for round-the-clock care. Then, as the technology begins to scale up and more doctors are trained in the procedure, BCIs will be easier and quicker to get. They might help with more ordinary, everyday things, like managing anxiety. Playing video games with your mind might come later.
“I don’t see a cyberpunk reality,” says Tom Oxley, the CEO of Synchron, which is already testing its BCI in humans. “I see a medical industry which is more akin to, like, LASIK surgery. It’ll be elective. It’ll be safe. It’ll be invisible. And it’s going to help you engage with the digital world better.”
The history of how brain chips got here followed a familiar pattern in science: It went slowly for a long time until it suddenly went quickly. The first time anyone plugged an array of electrodes into a mammal’s brain was at Stanford in the late 1960s. The electrodes were made the same way people still make computer chips, by chemically etching away a bit of silicon and adding traces of metal — in this case, gold. The brains belonged to cats, and the researchers were able to use the implants to detect the activity of a single nerve cell, a good proof of concept.
Until the early 2000s, implantable chips were mostly a research tool, a way to understand how those networks of neurons interacted and which parts of the brain did what. But then came America’s wars in the Middle East. Better trauma medicine was allowing soldiers to survive serious injuries, but more and more of them were coming home missing arms and legs. Looking for a technological solution, the US government stepped in. DARPA, the Defense Advanced Research Projects Agency, spun up blue-sky programs to develop advanced prosthetics — and innovative ways to control them. Armed with government funding, researchers began to hone more effective electrodes, developing bulky, hardwired BCIs that let monkeys control robot arms and that eventually let humans (with lots of practice and lots of hardware) control computers.
But the brain is a challenging neighborhood. For one thing, electrodes tend to be stiff, and brains are gooey. So the implants could actually injure the brains they were trying to read, like sticking a fork into bread dough. And to make matters worse, the dough fights back: A living brain will gum up an invasive electrode with cells called glia. So developing a long-lasting, implantable BCI that could link human thought to a computer was slow going at best. “We were a little before our time,” says Amy Kruse, a neuroscientist who ran early programs in noninvasive BCIs at DARPA. “It was right in terms of timing and in terms of the science, but not in terms of the adoption.”
At about the same time DARPA was getting into the game, a German engineer named Florian Solzbacher arrived at the University of Utah. There, he reconnected with an old mentor, Richard Normann, the inventor of a brain implant known as the “Utah array.” Imagine a tiny silicon chip the size of an eighth-teaspoon, bristling with 100 tiny teeth — millimeter-long electrode shanks coated with plastic and tipped with iridium or platinum. The array, which is able to record brain activity in everything from felines to fish, is approved by the FDA for use in humans, but only for 30 days at a time. But even with that cautious limitation, it has been a workhorse of neuroscience, a Hubble Space Telescope for looking into the brain. And it has served as a test bed for a variety of mind-controlled technologies, from cyborg arms to computer games.
Solzbacher decided to launch his own company to manufacture the implants. The first person he called was Marcus Gerhardt, an old friend from boarding school who had become something of a serial entrepreneur. Gerhardt reminded his pal that back in their high-school days, Solzbacher had dreamed of creating implant-controlled prosthetic limbs, inspired in part by his love for “The Six Million Dollar Man” — cheeseball 1970s television about a secret agent with superpowered robot limbs. “I had honestly forgotten that,” Solzbacher tells me. In 2007 he and Gerhardt launched Blackrock Neurotech, a startup dedicated to finding a way to make brain chips a routine part of medical care — to help patients with severe paralysis communicate via computers, say, or regain some independence of motion. To date, only about three dozen people have ever had a BCI embedded in their brains — and almost all of them were made by Blackrock.
The desire for a real-world implant was certainly there: Surveys of people with severe paralysis show, again and again, that they are prepared to put up with the technology’s limitations, including wires coming out of their heads and difficult training regimens. Many people who have had an electrode array implanted in their heads say that once they learn to operate it, it’s almost as unconscious as using a limb they were born with. But the process of getting there can be brutal, requiring repeated trips to a lab for training sessions and working up a sweat just by thinking at the machine over and over. It’s physically and mentally exhausting, and it doesn’t always work.
The bigger problem was that the money and the science weren’t ready. “The technology wasn’t there,” Solzbacher says. “The financial market wasn’t there.” So just as it had in the 1990s, the government came to the rescue. In 2013, an Obama initiative called Brain Research through Advancing Innovative Neurotechnologies — yes, BRAIN — began handing out millions in federal funds to improve BCIs. Batteries got smaller and able to run cooler. New coatings for electrodes, and smaller electrode shanks, reduced the brain’s messy backlash to implants. Researchers started developing chips that could communicate wirelessly; Synchron’s device is wireless, Blackrock has one in development, and a startup called Paradromics has one in animal trials.
But the government money wasn’t enough. That’s where venture capital came in, attracted by Musk. Not, to be sure, by his company’s tech. Neuralink doesn’t have a chip in human trials yet, and it has hemorrhaged scientific leadership, including its president and cofounder Max Hodak, who left earlier this year and put his money into Synchron. What attracted big-money investors was Musk’s business cachet — his imprimatur as a technological Midas. When The Wall Street Journal revealed the existence of Neuralink in 2017, it jump-started the industry.
“We have to thank Elon Musk and Neuralink for putting neurotech on the map,” says Enke Bashllari, a neuroscientist who is a managing director of Arkitekt Ventures, which has invested in Paradromics. “Blackrock worked on the Utah array for more than a decade, but it was put on the investor map by Neuralink.”
According to the research firm PitchBook, US investment in neurotech — including BCIs of all kinds — bottomed out at $1.3 million in 2014. In 2017 it spiked to $171 million, and last year it hit $378 million. And the number of deals has been rising every year. By some estimates, “neurotech” could be a $3 billion market by mid-decade and nearly double that by decade’s end.
“When we started in 2015 and I was pitching venture capitalists on brain computer interfaces, no one knew what a brain computer interface was,” says Matt Angle, the CEO of Paradromics. “People assumed it was 25 years away. Now when I pitch they say, ‘Oh, yeah, we definitely have a BCI thesis.'”
That thesis, in keeping with Silicon Valley’s driving obsession, almost always involves scalability. And each company competing to be the first and biggest has a different plan for how to get there. Blackrock, for instance, has decades of experience making BCIs with the kind of safety data that regulators will demand. But implanting the Utah-array-like device requires opening the skull and working on the brain. Right now, only about 150 surgeons in the United States know how to perform that kind of “functional neurosurgery.” So Blackrock is working to train more doctors to do craniotomies. “It has to scale,” says Dr. Brian Lee, the director of functional and stereotactic neurosurgery at the University of Southern California. “It can’t be, one guy knows how to do it, and everyone else can’t.” (Neuralink, in a typical Muskian flourish, expects its arrays to be implanted — in some futuristic process yet to be determined — by a sophisticated robot rather than human surgeons.)
Synchron has opted for a less complicated approach. Its device is mounted on a stent, an expanding mesh cylinder that gets threaded through a vein and up into the brain. A vascular surgeon — a more common specialty than neurosurgeon — can do it. The process is easier than brain surgery, is arguably safer, is definitely clever, and has already been used in several human trials. But the convenience comes with a compromise. Synchron’s “stentrode” can’t gather as much data as a Blackrock array. “We have the problem of a low-fidelity system,” Oxley says, “but it’s scalable.”
Now, thanks once again to the government, the field of BCIs is about to move into hyperdrive. This year, the FDA released its final guidance on what will be required from implants designed to treat paralysis and mobility issues. Despite the sterility of the language, it’s hard to overstate the significance of those 44 pages. The new rules came from more than eight years of intensive work by the FDA to figure out what is possible and how to make it safe and effective. “It’s been a collaborative process with them,” Oxley says, “working together to figure out the design, benchtop, animal, and then clinical testing requirements to address all of the potential safety issues.”
Instead of burning money or time on clinical trials that might wind up getting slapped down, companies know exactly what they need to do to receive FDA approval. Synchron has already received a green light from the agency for studying its implant in people with no time limits; Blackrock hopes to go to the FDA with its application later this year. Those are necessary steps toward final approvals.
The FDA rules also pave the way for insurance coverage for BCIs. “This is a $30,000-to-$50,000 device,” Oxley says. “Patients are not going to pay for it. So you have to figure out a pathway to reimbursement from Medicare or private insurance companies.” And once insurance starts picking up the tab, the market will follow. Sure, the devices may be expensive. But if they’re just a smidge less expensive than the labor-intensive, round-the-clock care that many patients with severe paralysis require, they can be a financial success. An inability to speak because of ALS might affect only 150,000 people, says Angle, the Paradromics CEO. “But due to the fact that it is an unmet need, the cost the insurers can bear is considerably higher,” he says. “You can still have a $20 billion business.”
Christian Angermayer, who’s on the board of Blackrock and has invested in Synchron, also foresees a huge market, despite the relatively small number of patients. “If you add all severely disabled people together, you’re in the millions,” he says. “And because they’re so costly for the healthcare system, it’s an amazing business.”
Investors and scientists believe that BCIs can do far more than provide autonomy to people with severe disabilities. Once they start plugging the devices into people’s heads and gathering data on their effectiveness at easing severe disability, they expect to offer them to otherwise healthy consumers looking for things like memory enhancement — the science-fiction stuff.
This is where venture capitalists start using the language they throw around in other fields. A BCI can be a “platform,” they hope, that will host a bunch of software tools, sort of like apps for the mind. “What I believe is going to happen is that there will be two or three companies who bring the chip into the brain,” says Angermayer. “Then you’ll have app stores and thousands of companies building on this sort of platform technology. For me, Blackrock is Apple.”
Blackrock, in Silicon Valley parlance, has the biggest installed base — the number of brain implants actually in use. But the company, which is backed by Peter Thiel, had a $10 million investment round. Neuralink, by contrast, has $373 million from investors like Sam Altman and Founders Fund. Paradromics and Synchron have received about $50 million each.
Dr. Alex Morgan, a partner at Khosla Ventures, is banking on Synchron because he sees its stent technology as less invasive and more scalable than Blackrock’s array. “Once you’re there in the brain, you’re collecting information,” he says. “I’ll be very proud if Synchron is a product just used by paralyzed people, and that’s all it ever is. But if you said, what makes it an extraordinarily exciting venture investment? It’s all the other stuff.”
Investors are excited by that “other stuff” because it unlocks markets well beyond people with paralysis. That could include things like mental health — maybe sleep trackers, say, or the ability to quell anxiety. It could also include the wider world of pharmaceuticals. Big Pharma has traditionally steered clear of investing in medical devices, which differ from drugs in their approaches to R&D, acquisitions, and regulatory approval. But implantable chips could provide a wealth of valuable data on how existing drugs affect the brain’s complex chemistry and how to create new drugs that would be more effective at treating psychiatric illnesses. At least one pharmaceutical company has already started its own in-house neurotech venture fund, and others are likely to follow.
This will all involve complicated ethical considerations and data-use agreements. Collecting people’s brain activity, after all, is a lot more invasive than just following which ads they click on. But the future of mental health, it appears, could be more electronic than chemical.
“Neuroscientists, and those turning neuroscience into venture-backed businesses, are just starting to understand the value of that data,” says Kruse, the former DARPA researcher who led a $20 million round of seed funding into Paradromics as a partner at Prime Movers Lab, a VC firm that supports scientific startups. “It’s full stack, right? Full-stack neuroscience.”
Whoever winds up being first to market, the technology will keep improving. New kinds of BCIs are already working their way through the lab: ones that are nearly microscopic, or injectable, or laser-powered. The horizon of technical improvement will probably take 30 years. But it won’t take anywhere near that long for brain implants to create a lucrative market — and to change the world for thousands of people with severe disabilities. Gerhardt, the Blackrock cofounder, recalls attending a neuroscience conference that featured a panel of people with paralysis as well as someone who had lost an arm. All of them were asking for implanted BCIs as soon as possible — wired, wireless, hard to learn, anything. “They all saw that it wasn’t perfect,” Gerhardt says. “But all of them said they wanted this device now.” It seems as if they’re finally going to get it.
Adam Rogers is a senior correspondent at Insider.
Read next
Listen to The Refresh, Insider’s real-time news show
Listen to The Refresh, Insider’s real-time news show