OpenAI Invests in Merge Labs’ Brain Interface Startup

OpenAI has taken a stake in Merge Labs, a brain-computer interface startup co-founded by OpenAI CEO Sam Altman, deepening the alignment between advanced AI models and next-generation human–machine interaction. Merge emerged from stealth with a $250 million seed round at an $850 million valuation, according to people familiar with the deal, with OpenAI writing the largest single check.

Positioning itself as a research lab, Merge says it aims to bridge biological and artificial intelligence through noninvasive technology that communicates with neurons using molecules and deep-reaching modalities like ultrasound, not electrodes. The pitch is ambitious: restore lost function, stabilize mental health, and eventually give people faster, more natural ways to work with AI.

Why OpenAI Is Betting On Brain Interfaces

OpenAI framed BCIs as a “human-centered” gateway to AI, arguing in a company blog post that interfaces capable of interpreting intent could make models more useful under the messy conditions of real life, where signals are noisy and context changes by the second. As part of the deal, OpenAI plans to collaborate with Merge on scientific foundation models and tools that could accelerate bioengineering and device R&D.

The strategic logic is circular but compelling: if Merge makes it easier to think with, not just type at, AI systems, OpenAI’s own products become stickier. In turn, OpenAI’s models can help Merge design, simulate, and refine interfaces faster than traditional lab cycles allow.

Inside Merge Labs’ Approach to Noninvasive Brain Tech

Merge’s bet on molecular and ultrasound-based methods targets a long-standing trade-off in neurotech. Invasive implants can deliver high signal-to-noise ratios and fine spatial resolution but require surgery. Noninvasive tools are safer and scalable but often suffer from lower bandwidth and precision. Merge is trying to bend that curve.

Co-founder Mikhail Shapiro, a Caltech researcher known for advances in ultrasound neuromodulation and molecular imaging, signals a focus on bioengineered “transducers” that could both read and write neural activity more deeply than surface methods. If successful, such tech could move beyond today’s lab demos toward everyday cognitive assistance—think silent intent capture for writing, design, or communication without a keyboard.

It’s early days. Demonstrating safe, repeatable “write” capabilities that influence brain states without surgery remains one of the hardest problems in neuroscience. Metrics that truly matter here include bandwidth, latency, spatial specificity, and long-term safety; Merge’s claims will be judged on those metrics as preclinical and, eventually, human data arrive.

Who Is Backing and Building Merge Labs’ BCI Ambitions

Alongside OpenAI, investors include Bain Capital, Interface Fund, and Fifty Years, plus video game veteran Gabe Newell. Fifty Years co-founder Seth Bannon described the company as part of a centuries-long effort to extend human capability through tools—now aimed at the brain itself.

Merge’s founding roster blends AI, neurotech, and bioengineering. In addition to Altman and Shapiro, co-founders include Alex Blania and Sandro Herbig of Tools for Humanity, and Tyson Aflalo and Sumner Norman of Forest Neurotech. The companies involved say they will continue operating independently, with founders maintaining roles at their existing organizations.

Competitive Landscape and Regulation in the Fast-Evolving BCI Field

The move places Altman in more direct competition with Elon Musk’s Neuralink, which uses an implanted array to let patients with paralysis control devices. Neuralink last raised a $650 million Series E at a $9 billion valuation and continues to pursue high-bandwidth readouts via electrode threads.

Other players are testing alternative paths. Synchron threads a stent-like array through blood vessels to reach motor cortex without open-brain surgery. Precision Neuroscience seeks thin-film interfaces atop the cortical surface. Each approach navigates a different balance of surgical risk, data quality, and regulatory scrutiny by agencies like the FDA.

Merge’s noninvasive stance could open a wider consumer funnel if performance approaches invasive benchmarks, but it must still clear rigorous safety and efficacy milestones. Expect early proof points in specific use cases—such as assistive communication or attention modulation—before any general-purpose “superhuman” enhancements are credible.

Strategy and Governance Questions Around OpenAI and Merge

The investment also raises familiar governance questions. OpenAI, which primarily backs startups through the OpenAI Startup Fund, has made or partnered on efforts linked to Altman before, including Rain AI, Harvey, and energy ventures such as Helion Energy and Oklo. Supporters argue these ties tighten execution across AI models, hardware, and infrastructure; critics point to circular incentives if companies Altman helps build also expand OpenAI’s footprint.

OpenAI has hinted at a broader hardware play via collaboration with Jony Ive’s design team, exploring post-screen interfaces. Merge could become the neural layer in that stack, translating intent directly into actions for AI agents, especially as models grow better at operating devices and services on a user’s behalf.

What to Watch Next as Merge and OpenAI Pursue Noninvasive BCIs

Key milestones include peer-reviewed preclinical results, clarity on safety profiles for molecular and ultrasound modalities, and a regulatory path to human studies. On the AI side, look for joint work on foundation models tuned for low-latency intent decoding, personalization, and reliability under limited signals.

If Merge proves a noninvasive interface with usable bandwidth and robust safety, it could shift BCI from a hospital-bound specialty to a mass-market input layer for AI—an outcome that would justify OpenAI’s bet and reshape how people create, learn, and communicate with machines.