OpenAI is advancing its strategy to lessen its dependence on Nvidia by developing its first-generation in-house AI chips.
According to sources cited by Reuters, the ChatGPT maker is finalizing the design for its custom silicon and expects to send it for fabrication at Taiwan Semiconductor Manufacturing Co. in the coming months. The process of finalizing a chip design and preparing it for manufacturing, known as “taping out,” marks a significant step in OpenAI’s effort to secure greater control over its hardware supply chain.
