They Raised a $1 Million for a Device That Gives You Lucid Dreams. Could It Really?

Eric Wollberg’s interest in facilitating lucid dreams emerged while living in Jerusalem, reading lots of theology. “Abraham, Muhammad, Buddha, all those prophets received their prophetic wisdom in their dreams,” the startup and tech investment alum told me recently. During some periods of his life, Wollberg too regularly experienced the sorts of dreams where he knew he was awake. He wondered whether there was a way to use emerging technology to have them on demand.

Last February a hint came from, of all people, Grimes. The electronic pop artist retweeted a software engineer named Wesley Berry III, who was playing around with turning brain waves into art and running his computer with his thoughts. Soon Wollberg was at Berry’s house in San Francisco, pitching him on applying one of these electrode-filled brain activity–monitoring headsets to lucid dreams.

Within just four months, Wollberg and Berry’s new company, Prophetic, raised more than $1 million in funding for a consumer device—the “Halo”—from venture capital heavy hitters and acquired advisers who’d worked in neurotech at Apple. They also promptly forged a research collaboration with a widely respected neuroscience institute and a company that built hardware for Elon Musk’s brain-computer interface company Neuralink, a detail that has been featured in articles about this new “A.I. startup.” Already they’ve invited customers to pay to reserve one of the first devices, which they intend to start shipping in 2025. At an event in New York City on Friday, potential investors will see a “prototype” of the Halo and learn about the company’s plan to use generative A.I., similar to the tech behind ChatGPT, on brain data to induce and stabilize lucidity in the dreamer.

But their own dream remains incomplete: The team accomplished all of this without ever inducing a single lucid dream, I learned after repeatedly pushing for specifics during a video call. Indeed, scientist collaborators from the Donders Institute for Brain, Cognition, and Behavior in the Netherlands are in the early stages of what Wollberg called “the best funded study ever in lucid dreaming.” But the idea that you can prompt a lucid dream with ultrasound transducers—their brain stimulation technology of choice—is still purely hypothetical, Wollberg acknowledged.

This is something the language they use in marketing materials makes easy to miss. For example: “The combination of ultrasound and machine learning models (created using EEG & fMRI data) allows us to detect when dreamers are in REM to induce and stabilize lucid dreams,” the Prophetic website states.

To focus on any inconvenient details is to miss the point, Wollberg told me, because the real breakthrough already happened: having the insight to apply a particular type of A.I.—with its ability to spot patterns in enormous data sets—and ultrasound transducers to lucid dreams.

“We’re very confident that we will be very successful,” he said. He likened it to the oft-cited Steve Jobs Xerox plant breakthrough: In the late ’70s, Jobs may have noticed a graphical user interface being underutilized by Xerox and suggested applying it to the Macintosh. “It’s all there,” Wollberg  said.

Viewed one way, their endeavor hints at one rarely discussed upside to A.I.: invoking the tech can lure investors to fund scientific research in off-the-beaten-track topics like consciousness and dreaming. Viewed another way, it’s an example of how buzzy terms like artificial intelligence may artificially inflate optimism that the human brain can be programmed as easily as a computer. Regardless, it would be cool if it (safely) worked.

Lucid dreams are generally thought to emerge from standard dreams in rapid eye movement, or REM, sleep. The sleeper becomes conscious that they are dreaming but does not wake up. Many lucid dreamers are able to direct the dream, taking flight, having a picnic with a relative who died, playing onstage with a famous musician, or anything else they desire. “Lucid dreaming denotes a rare state of sleep comprising cognitive features of both waking and dreaming,” German scientists wrote in a 2012 study on the topic.  Two years later, a different group of German scientists, joined by a researcher from Harvard Medical School, wrote, “Scientifically, lucid dreams present the unique opportunity to watch the brain change conscious states, from primary to secondary consciousness.”

That study found that it was possible to induce lucid dreams using electric currents. Wollberg and Berry plan to build a headset that utilizes a different kind of “noninvasive neurostimulation”—a transcranial focused ultrasound. This emerging technology activates highly specific regions of the brain, through the skull, using high-frequency sound waves. Researchers have been exploring whether it’s possible to apply something similar toward fighting brain tumors and inducing a meditative state. Wollberg and Berry hope that Martin Dresler, a cognitive neuroscientist trained in philosophy, and his team will be able to identify the precise brain coordinates activated during lucid dreaming and the frequency and length of the ultrasonic pulses needed to trigger them. Then they can figure out how to reverse-engineer this brain state.

Once they have the brains of lots of lucid dreamers mapped, Prophetic will train A.I. on that neural activity data, “in the same way that you train ChatGPT on words and sentences and stories and then it outputs the activation patterns,” Wollberg said.

Asked what would happen if the researchers fail to find that lucid dreams can be induced by these high-frequency sound waves, Wollberg, who studied economics, sounded utterly unphased.

“Lucid dreams are a naturally occurring brain state,” he reminded me with the confidence of a man accustomed to navigating the highly financed fringes. (Before starting this company, he was the head of community at Praxis, a Peter Thiel– and Winklevoss twins–backed company aiming to cultivate people to “live in an autonomous charter-state built on a decentralized crypto economy” in the Mediterranean, as Curbed put it. Sam Bankman-Fried’s now-defunct hedge fund, Alameda Research, also invested. Before that, Wollberg worked for a venture capital firm focused on psychedelics. )

Is any of this plausible? I reached out to two scientists specializing in lucid dreaming. Both spoke highly of Dresler and were pleased to hear that lucid dream research is advancing. Both were skeptical of what Prophetic—now offering early product access for $100—is promising within two years.

Yes, lucid dreams can be induced, Tadas Stumbrys, the assistant director of research at the Alef Trust, a nonprofit organization focused on the study of consciousness, confirmed. (Stumbrys has conducted lucid dream studies himself.) But “as of now, we do not yet have a method of how to induce lucid dreams reliably and consistently outside of sleep laboratory settings,” he wrote to me. And in those studies, the success rate hovered at around 50 percent—with induction possible only among those who regularly had lucid dreams.

None of these lucid dream studies examined the particular focused ultrasound method that Prophetic is building its product around. Stumbrys called it “an interesting novel approach, which may have some potential in facilitating lucid dreaming,” but there is zero empirical research, so far, that it can induce lucid dreams. He also had concerns about safety. While neurostimulation has been shown to be “relatively safe” when used for short periods in a lab, “we do not know what might be the cumulative side effects due to the prolonged use in the uncontrolled settings such as the home environment.”

Benjamin Baird, a psychology professor at the University of Texas at Austin who also has published papers on lucid dreams, offered a similar blend of optimism and caution. Activating the frontal region of the brain to prompt or stabilize lucid dreams has “some prior plausibility,” but “no one has yet produced solid empirical evidence of any technique or device that can do this,” he said.

And if they can’t, they will lose what Wollberg told a podcast gives his product an advantage over augmented reality headsets “The user is generating the content for themselves” he said. You don’t need to hire AR or VR designers when you’re selling the customers access to their own imaginations.

Still, their headset seems more promising than some of the other lucid-dreaming products out there. (And yes, there are many.) It also seems to be easier to access than the Dormio, a dream-capturing glove from MIT Media Lab, which requires building it yourself.

There is also no denying that A.I. will accelerate medical research. And in the grand scheme of the multibillion-dollar A.I. hype bubble, the $1 million in funding the company obtained is “nothing” Gabriel De La Rosa Cols, a principal with the “A.I.-powered” PR platform Intelligent Relations told me. De La Rosa Cols, who is familiar with the venture funding landscape, didn’t find it nearly as strange as I did that investors were willing to fund something based on an unproven hypothesis. Venture capitalists “are supposed to invest in things that are not yet reality.”

They also like to circulate tantalizing ideas like “Your brain is software; program it.” But the thinking among people who have spent their lives studying the brain tends to be quite different. The brain is not software. Nor is it a computer or anything else that you can say—with certainty—can be engineered to give you the results you want on the timeline you want. That’s true no matter how powerful the technology you’re using or the hype you build around your product.