A big investment
The trail to making a custom AI chip requires substantial resources. Industry experts told Reuters that designing a single version of such a processor could cost as much as $500 million, with additional expenses for developing supporting software and hardware potentially doubling that quantity.
The present OpenAI chip project, led by former Google chip designer Richard Ho, involves a team of 40 engineers working with Broadcom on the processor design, based on Reuters. The Taiwanese company TSMC, which also produces Nvidia’s chips, will manufacture OpenAI’s chips using its 3-nanometer process technology. The chips will reportedly incorporate high-bandwidth memory and networking features much like those present in Nvidia’s processors.
Initially, OpenAI’s first chip will focus totally on running AI models (often called “inference”) relatively than training them, with limited deployment across the corporate. The timeline suggests mass production could begin at TSMC in 2026, though the primary tape-out and manufacturing run faces technical risks that might require additional fixes and will delay the project for months.
OpenAI’s move into AI hardware comes as major tech corporations spend record amounts on AI infrastructure. Microsoft plans to speculate $80 billion in 2025, while Meta put aside $60 billion for the following 12 months, Reuters notes. Last month, OpenAI (working with SoftBank, Oracle, and MGX) announced a brand new $500 billion “Stargate” infrastructure project aimed toward constructing latest AI data centers within the US.
