On February 7, 2026, OpenAI — together with CEO Sam Altman personally — announced a major investment in Merge Labs, a newly formed startup developing next-generation brain-computer interfaces (BCI).
They are joined by an unexpected co-investor: Gabe Newell (the legendary co-founder of Valve), adding significant hardware and human-interface expertise to the round.
Merge Labs positions itself as a direct competitor to Elon Musk’s Neuralink — but with a radically different technical philosophy. Instead of invasive implants, the company is betting on minimally invasive or eventually fully non-invasive methods that could read from much larger brain volumes with fewer risks and far broader accessibility.
Neuralink’s Invasive Approach vs. Merge Labs’ Vision
Neuralink currently relies on a highly invasive procedure:
- Surgeons drill a small hole in the skull;
- A coin-sized device with ultra-thin electrode threads is inserted;
- Threads extend several millimeters into brain tissue to record and stimulate neurons.
This gives excellent signal quality — but only from a limited spatial region, requires brain surgery, and carries long-term risks (scarring, inflammation, immune response, device degradation).
Merge Labs is pursuing the opposite path: technologies that avoid opening the skull entirely or require only minimal intervention, while ideally covering much wider areas of the brain.
Merge Labs’ Core Technology: Ultrasound + Molecular Reporters
According to an exclusive report published by Core Memory on February 6, 2026, Merge Labs’ initial research focuses on functional ultrasound imaging combined with engineered molecular reporters — specialized proteins designed to make neural activity much easier to detect from outside the skull.
The high-level plan:
- Deliver specially engineered proteins (molecular reporters) into the brain;
- These proteins integrate with neurons and dramatically amplify the natural mechanical/hemodynamic signals produced during neural firing.
- Advanced ultrasound systems then detect those amplified signals through the skull — potentially with much greater spatial coverage than electrode arrays can achieve.
Molecular reporters act like “biological loudspeakers”: when a neuron fires, the protein changes in a way that produces a stronger, more distinct signature that ultrasound can pick up with high precision and from deeper/wider regions.
The biggest unanswered question right now: how will these proteins be delivered to the brain?
Merge Labs has not disclosed the method, but the most plausible and widely discussed hypothesis is gene therapy — using safe viral vectors (such as AAV) to deliver genetic instructions that cause neurons to produce the reporter proteins themselves. If successful, this would be a one-time (or very infrequent) procedure, after which the brain would continuously express the signal-enhancing molecules.
Why OpenAI and Sam Altman Are Investing
OpenAI has repeatedly emphasized that building safe, aligned superintelligence will require deep understanding of human cognition, values, preferences, and real-time decision-making.
Brain-computer interfaces could become one of the most powerful tools for:
- Reading human intent with far greater fidelity;
- Creating bidirectional communication channels (human ↔ AI);
- Enabling new forms of human–AI collaboration or augmentation.
Sam Altman has long been personally interested in BCI. His participation (beyond OpenAI’s institutional investment) signals strong conviction that safer, more scalable interfaces could eventually leapfrog today’s invasive approaches — both in technical capability and in societal acceptance.
Gabe Newell’s involvement adds another dimension. Valve has quietly explored BCI for gaming, VR/AR interaction, and human augmentation since at least 2019. Newell likely sees non-invasive BCI as the logical successor to headsets — and wants to ensure Valve is not locked out if Neuralink (or anyone else) dominates the space.
Very Early Days — But Extremely High Stakes
Merge Labs remains in stealth/pre-clinical mode. No human data, no prototype device, and no public timeline for trials have been shared. Many scientific, engineering, and regulatory challenges remain — especially around safe, effective, and scalable delivery of molecular reporters to large brain volumes.
But the backers are extraordinary:
- World-leading AI research and scaling expertise (OpenAI / Altman);
- Deep experience building human-facing hardware and interfaces (Newell / Valve);
- A fundamentally different — and potentially far more scalable — technical approach.
If Merge Labs can deliver even a fraction of its ambition — high-resolution, large-scale brain reading without drilling holes — it could reshape the entire BCI landscape and accelerate the arrival of practical brain–machine symbiosis.
For now the company is mostly silent beyond the funding announcement. But one thing is already clear: the race to connect minds and machines just became significantly more interesting — and significantly less invasive.
Further reading / watching:
→ Core Memory exclusive: OpenAI and Sam Altman back Merge Labs BCI (Feb 2026)
→ Recent interview with one of the co-founders (YouTube, early February 2026)
Also read:
- Anthropic Unveils Claude's New Constitution: A Blueprint for AI Alignment
- NVIDIA's H200 Conundrum: US Approvals Meet Chinese Blocks in the AI Chip Wars
- Utah's AI Prescription Revolution: Doctronic Ushers in Autonomous Medicine
Thank you!
