
With graphene chips and AI support, the Barcelona company wants to revolutionize the medical treatment of neurological diseases.
AI-generated summary
Inbrain emerged from the EU Graphene Flagship Initiative. The company uses graphene for brain chips to process more precise neural signals.
Barcelona. Carolina Aguilar pulls up the skullcap of the white Styrofoam head in her office - as naturally as if it were an office utensil. The boss and co-founder of the Spanish start-up Inbrain points to the place above the ear where her flexible, wafer-thin chips are to be implanted in people. Behind it, a pulse generator is connected to the chip via a cable and is the size of a USB stick. Its battery can be charged wirelessly from the outside through the skullcap.
What sounds strange is actually one of the great hopes of medicine. The chips read neural signals from the brain and convert them into actions using computers. These “Brain Computer Interfaces” (BCI) are already helping paralyzed patients in experiments, for example, to operate a computer mouse.
“The technology has the potential to significantly improve the lives of millions of people around the world suffering from a variety of medical conditions,” said an analysis by investment bank Morgan Stanley. The bankers estimate the market potential for the USA alone at 400 billion US dollars.
Accordingly, there is a global race for leadership in key technology - and with Inbrain, Europe is at the forefront. According to the data platform Dealroom, the Spaniards are among the ten largest BCI start-ups in the world, alongside tech giants such as Elon Musk's Neuralink.
The US tech group Microsoft and the German pharmaceutical company Merck have selected Inbrain as a partner. The start-up also owes its good position to an EU research initiative on the new material graphene. “Without them, Inbrain wouldn’t exist,” says Aguilar.
Inbrain is the only company in the world to use graphene instead of a precious metal for its brain chips. Graphene is a carbon that was first created in a laboratory in 2004. It is extremely flexible, but at the same time harder than steel and very electrically conductive. This offers numerous advantages for brain chips.
“The graphene chips can stay in the body longer, hold many more electrodes in a small space than metals and therefore work much more precisely,” says Aguilar. More electrodes read more signals from the nervous system and provide a detailed picture of the complex processes in the body.
“Most importantly,” Aguilar says, leaning forward in her chair to emphasize the sentence. "Above all, we are developing a system that not only reads the electrical signals from the nerves, but also sends signals itself. This means we can treat problems in the body in real time."
In an epilepsy patient, for example, the inbrain chip is supposed to detect when a seizure is imminent and send targeted electrical shocks to the nervous system in order to prevent it. “We call this therapeutic BCI instead of assistive BCI,” says Aguilar, leaning back.
Most providers, such as the US company Medtronic, where Aguilar previously worked for 13 years, would read the signals from the nervous system and forward them to a PC, which would then execute commands.
The start-up Inbrain, founded in 2019, uses the cloud infrastructure and large language models from Microsoft. They should learn to understand the signals of each patient and automatically send impulses for treatment under medical supervision. That would be an AI agent for the brain.
“Inbrain’s basic idea of using graphene as a transistor is fantastic,” says Thomas Stieglitz, professor of biomedical microengineering at the University of Freiburg. Graphene can also measure slow waves of electrical activity in nerve pathways, such as those that occur in strokes and epilepsy. “All other electrodes cannot do this, or do it much worse,” says the expert.
However, it is questionable whether as many electrodes as Inbrain is installing are necessary. While competitors use a handful to a few dozen, the Spanish install up to 1024 on their graphene chips. “This increases the complexity for surgeons and doctors and may not be necessary,” says Stieglitz.
Inbrain tested its first use in humans five years after the company was founded. In 2024, the Manchester Center for Clinical Neurosciences used inbrain chips in brain tumor surgery. They helped surgeons distinguish healthy tissue from tumor with micrometer precision and not remove more tissue than necessary.
The application as a surgical assistant has the status of a “breakthrough device” in the USA. Accelerated approval procedures are planned for this purpose. If all goes well, Inbrain's first products should come onto the market at the end of 2027.
The start-up emerged from several Catalan research centers and the University of Manchester. They had researched the new material as part of the EU Graphene Flagship Initiative, which was endowed with a billion euros.
The institutes have assigned three patents to Inbrain, and the start-up has now registered over 40 more itself. “The EU has shown foresight when it comes to graphene and has thereby secured us a head start,” says Aguilar.
However, other regions are now pushing the issue more strongly. China has declared BCI a strategic industry and aims to build two or three world-class manufacturers by 2030. According to Aguilar, bureaucratic hurdles such as the responsibilities of different ministries do not exist there any more than in the USA. “We have received a lot of support in Europe, but now we are not making progress fast enough,” says the 50-year-old.
Nevertheless, Aguilar is fully committed to the EU. In order not to be dependent on suppliers from the USA or China, Inbrain produces all parts of its hardware itself, apart from the wafers for the semiconductors. Aguilar proudly shows the 1,000 square meter clean room in Barcelona that is currently being set up.
Inbrain has raised €135 million in investor funds, primarily from non-governmental, institutional investors in Europe. A financing round is currently underway in which Aguilar wants to raise a further 100 million euros.
Inbrain does not yet have unicorn status, as a billion-dollar valuation is called in industry jargon. Merck has already given the Spaniards a fabric unicorn for their five-year anniversary. This now sits in the corner of an Inbrain meeting room in the new offices on Avenida Diagonal, one of Barcelona's main arteries.
Merck has been working with Inbrain since 2019 and says it has invested a high double-digit million amount. “Inbrain is an unusual partner and a real stroke of luck for us,” says Robert Spoelgen, Head of Bioelectronics at Merck. The Spaniards are extremely fast, poached top people from the industry early on and have a clear vision.
Inbrain is not developing brain chips for Merck, but rather systems for stimulating the vagus nerve, which controls many bodily functions. In addition to the operating room assistant and the brain chip, it is the Spaniards' third business area.
As a first step, Merck wants to use the solution to treat the lung disease COPD and high blood pressure, which cannot be controlled with medication. Spoelgen appreciates the advantages of graphene. “The many electrodes work so precisely that they can, for example, only specifically address a specific lung function,” he says.
AI outlook — possibilities, not facts
Market launch of the first Inbrain products by the end of 2027.
Possible · Within years

AI pioneer Yoshua Bengio warns of the dangers of autonomous AI agents that pursue their own goals and can overcome security barriers. He reflects on his role in the development of neural networks and the resulting risks.

There is a threat of a new semiconductor shortage in Germany. Manufacturers quote delivery times of 52 weeks or more, and components are becoming significantly more expensive. Experts warn that the auto industry and mechanical engineering could come to a standstill without supplies.
OpenAI has published solutions to 377 difficult mathematical problems, including algebra and number theory, according to a media report.

According to a BSI and TÜV study, only eleven percent of German companies are prepared for AI-supported cyber attacks. The most common threats are phishing emails and automated attack scripts.

The book industry is increasingly defending itself against AI providers who use copyrighted works to train their models without permission. While publishers like Penguin Random House are suing, the case of author Thélyson Orélien is causing a stir.
The French AI company Mistral has presented its new flagship model Large 4. With improved capabilities in programming and cybersecurity, the model aims to close the technological gap with leading competitors from the US and China.