OpenAI participated in the seed funding round of brain-computer interface startup Merge Labs. The round raised a total of $252 million, valuing the company at $850 million.
OpenAI is the largest investor in this funding round, and Altman also serves as a co-founder in his personal capacity.
Other investors include private equity firm Bain Capital and Gabe Newell, co-founder of Steam, the world's largest game publishing platform. It's worth noting that Newell himself also founded a neuroscience company, Starfish Neuroscience.
This figure is unusual in the brain-computer interface field. According to industry data analytics firm PitchBook, among known brain science companies, only Elon Musk's Neuralink has raised more funding in a single funding round.
To understand the logic behind this huge investment, it is necessary to first understand the composition of Merge Labs' founding team.
01
The founding team of Merge Labs is a unique combination of Silicon Valley business acumen and academic research strength.
In the business aspect, the co-founders, in addition to Altman, include Alex Blania and Sandro Herbig.
Blania is the CEO of Tools for Humanity, the company that developed the controversial iris scanning project Worldcoin. Herbich is the company's head of product and engineering.
On the scientific front, co-founder Mikhail Shapiro is a professor of chemical and medical engineering at Caltech, with significant achievements in non-invasive neuroimaging technology. His research focuses on using ultrasound to interact with the brain, rather than traditional craniotomy.
The last two co-founders, Tyson Aflalo and Sumner Norman, are the co-founders of Forest Neurotech, a Los Angeles-based nonprofit research organization founded in 2023.
Norman earned his Ph.D. in Neural Engineering from Caltech, where he worked closely with Shapiro. Afrallo previously served as the executive director of the T&C Chen Brain-Computer Interface Center at Caltech and has extensive experience in clinical research on implantable brain-computer interfaces.
Forest Neurotech can be considered, in some ways, the precursor to Merge Labs.
This nonprofit organization has been researching ultrasound-based brain-computer interface technology for the past few years, and Merge Labs is the commercial entity spun off from Forest.
While such a transformation from a non-profit research organization to a commercial company is not uncommon in the biotechnology field, Merge Labds is truly the only company to have secured such a large investment in its seed round.
Merge Labs' technology approach is fundamentally different from that of mainstream brain-computer interface companies.
The human brain has tens of billions of neurons that communicate with each other via electrical signals every moment. The mainstream brain-computer interface uses electrodes to directly record these electrical signals, but this requires placing the electrodes inside the brain, which means that craniotomy is necessary.
The company has made it clear that they do not intend to develop devices that require electrodes to be implanted in the brain.
Instead, they want to send and receive information in a non-electrode manner, by connecting molecules and neurons, using methods with deep-penetrating capabilities such as ultrasound.
This may sound abstract, but the core idea is actually not complicated.
Merge Labs aims to use gene therapy to alter the properties of certain brain cells, enabling them to respond to ultrasound waves. They then use an ultrasound device implanted in the skull but without penetrating brain tissue to detect and modulate the activity of these cells.
In a recent speech, Shapiro explained that introducing genes into cells is much easier than implanting electrodes into brain tissue. The goal is to develop ways to interact with neurons in the brain and other parts of the body in a less invasive manner than current technologies.
Another advantage of this method is that it can cover a larger area of the brain. Traditional electrodes can only record neural activity in their location, while ultrasound can theoretically penetrate the entire brain, provided that the cells are modified to respond to ultrasound.
02
OpenAI is not just an investor.
Under the agreement, OpenAI will collaborate with Merge Labs to develop fundamental scientific models and other cutting-edge tools to accelerate research progress. In other words, future Merge Labs users will have ChatGPT pre-installed with them.
In its announcement, OpenAI stated that brain-computer interfaces are an important new field, representing a new way to interact with AI.
Ultraman's idea of human-machine integration was not a spur-of-the-moment idea. As early as 2017, he published an article predicting that humans would merge with machines sometime between 2025 and 2075.
Ultraman's view is that human-machine integration is the best solution for human survival.
In August 2025, Ultraman expressed this vision in a public speech, hoping that ChatGPT would respond whenever he thought of something.
However, he added that perhaps read-only functionality would suffice, which seems like a reasonable approach. After all, allowing AI to read brain signals and allowing AI to write information into the brain are two completely different concepts, the latter involving far more complex ethical and security issues.
From a technical feasibility perspective, installing ChatGPT in the brain is not a pipe dream.
If brain-computer interfaces can achieve high-bandwidth bidirectional communication, then theoretically, AI could indeed directly receive the brain's thought signals and transmit the responses directly back to the brain.
This means you don't need to speak or type; you just need to think about the question, and the AI can understand and answer it. Furthermore, if the interface supports writing, the AI can even present the answer directly in your mind, not just through sight or hearing.
03
When discussing brain-computer interfaces, one cannot ignore Elon Musk's Neuralink.
The company completed a $650 million Series E funding round in 2025, reaching a valuation of $9 billion.
As of 2026, Neuralink has implanted brain-computer interface devices in more than 20 patients. These patients can type, browse the web, and play games using their thoughts. The first participant, Noland Arbaugh, now uses the device for about 10 hours a day and is able to control a computer, play chess, browse the internet, and send text messages.
But Neuralink takes an invasive approach. Its device requires surgical implantation, specifically by using a surgical robot to remove a small piece of skull and then inserting ultra-fine electrode wires into the brain to read neural signals.
The advantage of this method is that it can record neural activity very accurately, but the disadvantage is that it requires craniotomy and long-term electrode implantation may trigger immune responses or tissue damage.
Neuralink has also encountered technical issues, such as some electrode wires detaching from brain tissue, but the company later restored functionality through software adjustments.
Merge Labs' non-invasive or minimally invasive solutions mean lower risk, easier regulatory approval, and a wider potential user base.
After all, only a minority of people are willing to undergo craniotomy, but if a small device is implanted in the skull without touching the brain tissue, the acceptance rate will be much higher.
This difference determines the market positioning of the two companies: Neuralink mainly focuses on the medical rehabilitation field in the short term, while Merge Labs is clearly targeting a broader consumer market.
A press release from Merge Labs stated that this project may take decades, not years, to complete. Their fully implantable brain-computer interface microsystem will require a considerable period to ensure safety and effectiveness before entering clinical trials, estimated to be at least two more years away.
This article is from the WeChat public account "Alphabet AI" , author: Miao Zheng, and published with authorization from 36Kr.
