In April 2026, the California-based business Sabi emerged from stealth with a product that sounded like science fiction: a winter hat that contains between 70,000 and 100,000 small EEG sensors that are intended to translate silently imagined words into letters on a screen. Khosla Ventures, Accel, Initialized, and Kevin Weil, the former chief product officer of OpenAI, are investors in the business.

The advantages of this technique would actually go well beyond just achieving typing speed parity. This kind of technology would be revolutionary for people with disabilities or injuries that make communication challenging. Another feature that sets Sabi apart from modern rivals like Neuralink is that the sensors would be implanted in a baseball cap or beanie rather than in users' brains.

Electroencephalography (EEG), a non-invasive technique that analyzes electrical signals from the brain through the scalp, is how the beanie operates. Sabi's method does not require surgery, in contrast to implant-based systems like those being developed by businesses like Neuralink.

The objective is to identify "internal speech," or the words individuals think but don't express aloud, and translate it into text on a linked device. If it works, people might be able to interact or operate gadgets without having to talk or type, reports Digital Trends.

Why It Matters?

Although brain-computer interfaces are not new, the majority of current devices can be classified as either invasive implants or large external gear. Due to issues with usability, expense, or medical hazards, neither strategy has gained widespread acceptance.

Sabi's beanie adopts a distinct approach by emphasizing accessibility and comfort. The gadget is made to function right out of the box without the need for daily calibration, which has been a significant drawback for a lot of BCI systems.

Related Brain Implants Allow Paralyzed People Type with Thoughts

Early goals call for typing speeds of about 30 words per minute, which might increase as users get more accustomed to the system.

This could lead to new applications, such as hands-free computing for regular users and accessible aids for individuals with disabilities.

The Challenges

The technology has a lot of obstacles to overcome despite its potential. Individual differences in brain signals are significant, and even the identical thought can result in slightly different neural patterns each time.

Sabi is creating a large-scale AI model that has been trained on hundreds of hours of volunteer brain data in order to address this. Finding patterns that match internal speech across several users is the goal.

Privacy Concerns

Privacy is one of the main issues with this technology. Deeply personal neural data can disclose intentions, thoughts, and cognitive habits.

Sabi claims that in order to ensure safe handling, it is addressing this by encrypting data and collaborating with neurosecurity specialists. However, as these devices get closer to commercialization, the larger discussion over brain data ownership and ethical use is probably going to expand.

What's Next?

By late 2026, the business hopes to launch the beanie's first consumer version along with a cap version.