According to the Korea Economic Daily, Samsung Electronics’ foundry division has initiated the R&D and trial production of Neuralink’s fourth-generation chip — the brain-computer interface company founded by Elon Musk. Project internal codename: “O1”, manufactured using Samsung’s 4nm process. This marks the first time Samsung has received a chip foundry order from Neuralink — the first three generations were all mass-produced by TSMC. Samsung launched the project in late 2025, and the first batch of test chips completed tape-out in May this year, with delivery planned for the first half of 2027. If testing goes smoothly, mass production could begin as early as the end of 2027.
The core upgrade of the fourth-generation Neuralink chip lies in enabling bidirectional communication: the first three generations could only read brain signals and convert them into device commands, while the fourth generation can also write data from external devices into the brain — for example, helping patients regain vision by simulating neural signals. Samsung’s winning of this order is closely related to its existing partnership with Musk’s Tesla, which has previously commissioned Samsung to manufacture its autonomous driving chips AI6 and AI6.5 using a 2nm process. TSMC’s exclusion is also linked to its capacity crunch: TSMC’s 3nm monthly production target of 175,000 wafers still cannot fully meet market demand, creating an opportunity for Samsung to take on overflow orders. Samsung’s foundry business targets profitability in 2028, and this order is expected to support that goal.