OpenAI And Broadcom Launch First Custom Chip To Expand AI Infrastructure
OpenAI has unveiled its first custom-designed AI inference chip, Jalapeño, developed in partnership with Broadcom. The launch highlights a significant step in OpenAI’s broader effort to build more of the technology stack that powers its AI services.
OpenAI said Jalapeño is designed specifically for AI inference, the stage where models process user requests and generate responses. Unlike training chips that build AI models, inference accelerators focus on delivering responses efficiently and at scale. The company described Jalapeño as the first processor in a multi-generation compute platform being developed with Broadcom and systems partner Celestica.
According to OpenAI, the chip has been architected around large language model workloads and is intended to support current and future AI models, including those beyond OpenAI’s own ecosystem. The move reflects a broader trend among AI companies seeking greater control over infrastructure as computing requirements continue to rise.
Infrastructure Is Becoming Strategic
For AI companies,...
Copyright of this story solely belongs to ciol.com. To see the full text click HERE
