Samsung will make next-generation Neuralink brain chip

Expanding its footprint across Elon Musk’s hardware ecosystem, Samsung Foundry has secured a landmark contract to manufacture Neuralink’s next-generation brain-computer interface (BCI) chip. According to a report from The Korea Economic Daily, the South Korean tech giant has officially initiated research and development on Neuralink’s fourth-generation implantable hardware.

The deal marks the first direct partnership between Samsung and Neuralink, shifting the neurotechnology company’s supply chain away from its historic manufacturing partner, TSMC.

Technical Specifications: The 4nm “O1” Node

The upcoming semiconductor architecture has been given the internal corporate codename “O1.” Rather than targeting bleeding-edge 2-nanometer lines, the production roadmap leans heavily into proven, ultra-stable infrastructure:

Process Node: The chip will be fabricated using Samsung’s advanced 4-nanometer (nm) lithography process.
Why 4nm Matters: While newer nodes offer denser logic, the 4nm line offers superior yields and thermal stabilization. For an implantable, medical-grade device, minimizing heat dissipation and guaranteeing zero-defect reliability takes strict precedence over raw processing speed.
Diversification Drive: The shift comes as TSMC’s advanced 3nm and AI packaging facilities remain almost entirely choked by massive, continuous orders from Nvidia and Apple, forcing hardware firms to diversify their foundry dependencies to secure stable production windows.

The Feature Leap: Bidirectional Neural Pathways

The transition to the fourth-generation chip marks a major milestone in Neuralink’s technological timeline:

                  NEURALINK HARDWARE EVOLUTION
┌────────────────────────────────────────┐    ┌────────────────────────────────────────┐
│     Earlier Eras (Gen 1 - Gen 3)       │    │        Next-Gen Era (Gen 4 / O1)       │
│              Via TSMC                  │    │              Via Samsung               │
│────────────────────────────────────────│    │────────────────────────────────────────│
│ • One-way transmission                 │    │ • True bidirectional communication     │
│ • Reads electrical spikes from brain   │    │ • Reads spikes AND feeds data back in  │
│ • Decodes intent to drive external UI  │    │ • Simulates neurons to trigger body    │
└────────────────────────────────────────┘    └────────────────────────────────────────┘

While earlier iterations functioned primarily as unidirectional readers—decoding a patient’s neural electrical spikes to translate thought into digital mouse clicks—the “O1” architecture is built for true bidirectional communication. The chip can ingest external data arrays and feed artificial stimulation directly back into the brain tissue. This capability is vital for Neuralink’s long-term medical pipeline, which includes projects aimed at restoring vision by directly stimulating the visual cortex or bypassing damaged spinal tissue to restore motor function.

Expected Production and Clinical Timelines

Development on the system integration began in late 2025, and pilot wafer fabrication officially launched in May 2026.

The current engineering roadmap places the delivery of initial operational test samples in the first half of 2027. If the pilot chips successfully pass strict laboratory testing and bio-compatibility validations, Samsung is scheduled to trigger full-scale mass production in the second half of 2027.

The deal further cements Samsung’s position as Elon Musk’s primary hardware manufacturing partner outside of the aerospace sector, building on existing multi-year agreements to supply specialized automated driving and computing chips (the AI6 and AI6.5 platforms) for Tesla’s electric vehicles and humanoid robotics divisions.

Search for an article