When Merge Labs emerged from stealth in January 2026 with $252 million in seed funding and an $850 million valuation, it immediately became the second most heavily capitalised brain-computer interface company in the United States, behind only Neuralink. Two months later, the scientific questions surrounding its core technology remain largely unanswered.
The company was co-founded by Sam Altman, Caltech bioengineering professor Mikhail Shapiro, Forest Neurotech co-founders Tyson Aflalo and Sumner Norman, and Alex Blania and Sandro Herbig, both previously at Tools for Humanity. OpenAI wrote the largest cheque in the round, with Bain Capital, Interface Fund, Fifty Years, and video game developer Gabe Newell also participating.
Merge’s pitch departs from the electrode-based paradigm that has dominated BCI development for the past two decades. Rather than implanting hardware into brain tissue, the company proposes to use focused ultrasound combined with engineered molecules to read neural activity at high bandwidth without surgery. Shapiro’s laboratory at Caltech has published work on using gene therapy to make individual cells visible to ultrasound imaging, offering a potential pathway to detecting neural signals without physical contact with neurons.
The approach, if it works, would sidestep the principal limitations of implanted BCIs: tissue scarring that degrades signal quality over time, infection risk from chronic implants, and the requirement for neurosurgery that restricts the addressable patient population. Merge has said it intends to develop medical applications first, then consumer products.
The scale of the ambition, however, has drawn pointed questions from researchers in the field. A February 2026 report in Nature News examined whether the underlying physics of ultrasound can support the spatial and temporal resolution needed for a practical BCI. Current functional ultrasound imaging, while capable of measuring blood-flow changes associated with neural activity, operates at a fundamentally different scale from the single-neuron or small-population recordings that existing electrode-based BCIs achieve. The gap between detecting haemodynamic signals across millimetre-scale regions and decoding the kind of fine-grained neural activity needed for cursor control or speech decoding is substantial.
The molecular component of Merge’s approach — using genetically engineered reporters to enhance ultrasound sensitivity to neural firing — is at an early research stage. Shapiro’s Caltech lab has demonstrated acoustic reporter genes in cell cultures and animal models, but translating this to a functioning human BCI would require safe and effective gene delivery to targeted brain regions, stable long-term expression of reporter genes, and ultrasound hardware capable of reading the resulting signals at sufficient speed and resolution. Each of these steps represents an open engineering and regulatory challenge.
None of this means the approach is unworkable. But the distance between the current science and a product that competes with implanted BCIs on bandwidth and latency is considerable, and the company has disclosed no clinical timeline, no product specifications, and no peer-reviewed data from its own laboratories. What it has disclosed is capital: $252 million is enough to sustain a large research operation for several years.
The funding itself is notable for what it signals about the BCI investment landscape. Neuralink, Synchron, Paradromics, and Precision Neuroscience have all raised significant rounds in recent years, but each of those companies had working prototypes or clinical data at the time of their major fundraises. Merge’s round was raised on the strength of its founding team, its scientific thesis, and the backing of OpenAI, which framed the investment as part of its broader interest in human-AI interaction.
For the BCI field, the question is whether Merge’s approach represents a genuine alternative to electrodes or an expensive exploration of physics that may not converge on a usable device within the timescales that venture capital demands. Shapiro’s research is credible and his lab is productive, but the path from acoustic reporter genes in a mouse model to a non-invasive human BCI with competitive bandwidth involves scientific leaps that no laboratory has yet demonstrated.
The $252 million gives Merge the resources to attempt those leaps. Whether the physics cooperates is a different matter.