DEVOURED
Jalapeño: OpenAI's new Chip
OpenAI and Broadcom have designed a new custom inference chip called Jalapeño to optimize gigawatt-scale data center performance.
What: OpenAI and Broadcom developed the Jalapeño accelerator in nine months using AI-assisted design tools, specifically focusing on energy efficiency for LLM inference.
Why it matters: This move suggests OpenAI is mirroring strategies used by Google and Amazon to verticalize their silicon supply chain, reducing reliance on Nvidia for high-throughput inference tasks.
Decoder
- Inference: The process of running a trained machine learning model on new data to make predictions or generate content.
Original article
OpenAI and Broadcom unveiled Jalapeño, the first accelerator in a planned family of LLM inference chips optimized for performance per watt and rapid deployment. The companies said the processor was designed in nine months with AI-assisted development and is intended for gigawatt-scale data center deployments.
Explore all 57 articles in today's digest
Fresh dev news, AI-digested daily. Never miss a trend at devoured.fyi