By ByteBot
Sam Altman’s latest venture, Merge Labs, emerged from stealth on January 15, 2026 with a $252 million seed round at an $850 million valuation—with OpenAI writing the largest check. The brain-computer interface startup aims to connect human thoughts directly to AI using molecules and ultrasound instead of Neuralink’s surgical implants. But there’s a catch: the company admits the technology won’t arrive for “decades rather than years.” That’s a long wait for an $850 million bet, especially when the CEO is also running OpenAI—and OpenAI is the biggest investor.
This marks the first major non-invasive BCI approach attempting high-bandwidth neural interfaces without cutting into the brain. OpenAI’s investment signals BCIs as critical infrastructure for natural AI interaction—think thought-to-ChatGPT queries instead of typing. However, the governance optics are terrible: Altman co-founding a company while OpenAI invests heavily raises questions about conflicts of interest.
Molecules Replace Electrodes: The Technology
Unlike Neuralink’s approach—64 ultrathin wires with 1,024 electrodes surgically implanted into brain tissue—Merge Labs uses molecular reporters that fuse with neurons and ultrasound to detect neural activity. No scalpel required.
The technology builds on Caltech researcher Mikhail Shapiro’s work using gas vesicles, hollow protein nanostructures from buoyant microbes, as genetically encoded imaging agents for ultrasound. His published research in Nature Biotechnology (2023) demonstrated that ultrasound readouts from the brain can predict intended movements before they happen. Consequently, Shapiro’s lab showed the information content ultrasound can extract is sufficient for brain-machine interface functionality—proving the concept works in research settings.
Merge Labs’ approach combines biology (molecular reporters), devices (ultrasound sensors), and AI (signal processing) in a non-invasive form factor. The molecular reporters enhance neural signals, while ultrasound—which penetrates deeper than EEG without surgery—captures the data.
The challenge? Non-invasive BCIs historically deliver much lower signal resolution than invasive approaches. Signals must transmit through the skull, which reduces clarity and limits bandwidth. Merge Labs claims its molecular approach can achieve high bandwidth without brain surgery, but that’s entirely unproven at scale in humans.
The Conflict: OpenAI Invests in CEO’s Own Company
Here’s where it gets messy. Sam Altman co-founded Merge Labs while serving as OpenAI CEO, and OpenAI became the largest investor in the $252 million round. This isn’t an isolated incident—it’s a pattern.
Altman co-founded Tools for Humanity (the company behind Worldcoin, now World) while running OpenAI. Moreover, that project has faced regulatory scrutiny in Bavaria, Brazil, Kenya, and Spain for data privacy violations. He stepped down as Oklo chairman in April 2025 to “avoid conflict of interest,” yet co-founded Merge Labs just months later. Now OpenAI is directly funding his ventures.
Industry watchers are calling the Merge Labs deal a governance red flag. When a CEO invests company funds in their own side projects, it raises questions about board oversight and where Altman’s focus truly lies. Furthermore, OpenAI is a $100+ billion organization—this level of conflict demands scrutiny. The timing is particularly awkward: Elon Musk’s lawsuit against OpenAI and Altman goes to trial on April 27, 2026, centered on claims that OpenAI abandoned its nonprofit mission.
Invasive vs Non-Invasive: The Trade-offs Matter
Neuralink announced mass production of surgical BCI devices in 2026 with automated procedures. The technical advantages of invasive BCIs are clear: higher spatial and temporal resolution, better signal-to-noise ratio, and more robust performance against electrical noise and movement artifacts. Electrodes placed directly in target cortical areas capture cleaner signals.
Non-invasive approaches like EEG and fNIRS are safer and easier to deploy, but they suffer from “limited signal-to-noise ratio and information transfer rate” because signals blur together as they pass through the skull. In fact, current non-invasive BCIs work for basic tasks—meditation tracking, simple gaming controls—but can’t match invasive bandwidth.
The BCI market is growing from $2.94 billion in 2025 to a projected $12.87 billion by 2034, with medical applications dominating at 46.41% of revenue. Invasive BCIs own the medical space (paralysis, ALS, neurological disorders) where high bandwidth justifies surgical risk.
Merge Labs is betting on a different market: consumer applications where healthy people won’t accept brain surgery. Gaming (hence Gabe Newell’s investment), AI interaction, workplace productivity—use cases where “good enough” bandwidth could enable mainstream adoption. If the molecular/ultrasound approach works.
“Decades Not Years”: Timeline Reality Check
Merge Labs explicitly stated the technology will take “decades rather than years” to develop. Read that again. This is an $850 million valuation for research that won’t yield consumer products until the 2040s at earliest.
OpenAI’s official rationale: “BCIs will create a natural, human-centered way for anyone to seamlessly interact with AI.” Translation: skip the keyboard, think your ChatGPT queries directly into existence. That vision makes sense for OpenAI’s long-term AI interaction strategy. But if deployment is truly decades away, why invest $252 million now?
Two possible answers. First, fundamental BCI research today could pay off massively in 20-30 years when AI is ubiquitous and everyone wants direct neural interfaces. Second, the Gabe Newell investment hints at nearer-term gaming applications—VR/AR control via thought could arrive sooner than full-bandwidth AI interaction.
Either way, honesty about timelines is refreshingly rare in tech. Most startups overpromise and underdeliver. Merge Labs saying “this will take decades” up front is either brutal transparency or a hedge against investor expectations. Time will tell which.
What This Means for Developers
Merge Labs represents the first serious attempt at a non-invasive, high-bandwidth brain-computer interface. The molecular reporter + ultrasound approach is genuinely novel, backed by published Caltech research. Therefore, OpenAI’s investment signals BCIs as important infrastructure for AI’s future—when everyone uses LLMs daily, direct thought-based queries could beat typing.
However, the conflicts of interest are glaring. Altman co-founding companies while running OpenAI, with OpenAI investing heavily in his ventures, looks bad. The “decades not years” timeline means most people reading this won’t see consumer products in their lifetime—at least not anytime soon.
Non-invasive BCIs make sense for consumer adoption. Nobody’s getting elective brain surgery for better ChatGPT queries. If Merge Labs proves molecular reporters achieve sufficient bandwidth without cutting into skulls, it could enable mainstream BCI adoption for gaming, productivity, and AI interaction.
Big if. Unproven technology. Decades-long timeline. Questionable governance. Worth watching, but don’t hold your breath.
