OpenAI is accelerating the deployment of AMD chips under a six-gigawatt agreement signed in October, even though its models still run almost entirely on Nvidia GPUs. The first gigawatt is expected to come online in the second half of 2026, followed by 457 megawatts from TensorWave across several U.S. data centers in the first half of 2027. Full deployment is expected by early 2028, when OpenAI also plans to scale production of its proprietary Jalapeño chip, developed with Broadcom. This diversification comes as Jensen Huang has resigned from Nvidia, although his potential return appears unlikely because the board is already searching for a successor.
Source: Read the original article


